AB 222: Data centers: power usage effectiveness: cost shifts.
- Session Year: 2025-2026
- House: Assembly
- Latest Version Date: 2025-07-07
Current Status:
In Progress
(2025-08-29: In committee: Held under submission.)
Introduced
In Committee
Awaiting First Chamber Vote
Passed First Chamber
In Committee
Awaiting Second Chamber Vote
Passed Second Chamber
Enacted
Existing law requires the State Energy Resources Conservation and Development Commission (Energy Commission) to biennially adopt an integrated energy policy report, as specified, and to make the reports accessible to state, local, and federal entities and to the general public.
This bill would require the Energy Commission to include energy consumption trends for data centers in its integrated energy policy reports.This bill would require the Energy Commission to establish a process for the owner of a data center, as defined, to submit the power usage effectiveness ratio, as defined, for the data center to the Energy Commission on a biannual basis, and require the owner of a data center to submit this information for the data center in the manner and timeframe specified by the Energy Commission. The bill would require the Energy Commission, as part of the 2027 edition of the integrated energy policy report, to include an assessment of electrical load trends for data centers, as provided.
Existing law vests the Public Utilities Commission (PUC) with regulatory authority over public utilities, including electrical corporations. Existing law authorizes the PUC to fix the rates and charges for every public utility, and requires that those rates and charges be just and reasonable.
This bill would require the PUC to determine whether those costs and expenses in an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center are just and reasonable. The bill would require the PUC to minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.This bill would require the PUC to assess the extent to which electrical corporation costs associated with new loads from data centers result in cost shifts to other electrical corporation customers, as provided, and require the PUC, on or before January 1, 2027, to submit the assessment to the relevant policy committees of the Legislature and to publicly post a copy of the assessment on the PUCs internet website.
Existing constitutional provisions require that a statute that limits the right of access to the meetings of public bodies or the writings of public officials and agencies be adopted with findings demonstrating the interest protected by the limitation and the need for protecting that interest.
This bill would make legislative findings to that effect.
Existing law, on or before January 1, 2026, and before each time thereafter that a generative artificial intelligence system or service, as defined, or a substantial modification to a generative artificial intelligence system or service, released on or after January 1, 2022, is made available to Californians for use, regardless of whether the terms of that use include compensation, requires a developer of the system or service to post on the developers internet website documentation regarding the data used to train the generative artificial intelligence system or service.
This bill would require a developer, before using a covered model commercially or before making a covered model available for use by a third party, to estimate the total energy used to develop the covered model and the percentage of the total energy used to develop the covered model that was generated in California. The bill would also require a developer, on or before February 1, 2027, and annually thereafter, to estimate the total energy used by the developer to operate the covered model during the previous calendar year and the percentage of the estimated total energy that was generated in California. The bill would require the developer to publish on its internet website that energy usage data, as provided.
Existing law requires the State Energy Resources Conservation and Development Commission (Energy Commission) to biennially adopt an integrated energy policy report, as specified, and to make the reports accessible to state, local, and federal entities and to the general public. Existing law requires the Energy Commission to prescribe, by regulation, energy and water conservation design standards for new residential and new nonresidential buildings, as provided.
This bill would require the Energy Commission to include energy consumption trends for data centers in its integrated energy policy reports. The bill would require the Energy Commission, on or before January 1, 2027, to adopt efficiency standards for data centers, as provided.
Existing law vests the Public Utilities Commission (PUC) with regulatory authority over public utilities, including electrical corporations. Existing law authorizes the PUC to fix the rates and charges for every public utility, and requires that those rates and charges be just and reasonable.
This bill would require the PUC to determine whether those costs and expenses in an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center are just and reasonable. The bill would require the PUC to minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.
Existing law, on or before January 1, 2026, and before each time thereafter that a generative artificial intelligence system or service, as defined, or a substantial modification to a generative artificial intelligence system or service, released on or after January 1, 2022, is made available to Californians for use, regardless of whether the terms of that use include compensation, requires a developer of the system or service to post on the developers internet website documentation regarding the data used to train the generative artificial intelligence system or service.
This bill would require a developer, before using a covered model commercially or before making a covered model available for use by a third party, to calculate estimate the total energy used to develop the covered model and the percentage of the total energy used to develop the covered model that was generated in California. The bill would also require a developer, on or before February 1, 2027, and annually thereafter, to estimate the total energy used by the developer to operate the covered model during the previous calendar year and the percentage of the estimated total energy that was generated in California. The bill would require the developer to publish on its internet website that energy usage data, as provided.
Existing law requires the State Energy Resources Conservation and Development Commission (Energy Commission) to biennially adopt an integrated energy policy report, as specified, and to make the reports accessible to state, local, and federal entities and to the general public. Existing law requires the Energy Commission to prescribe, by regulation, energy and water conservation design standards for new residential and new nonresidential buildings, as provided.
This bill would require the Energy Commission to include energy consumption trends for data centers in its integrated energy policy reports. The bill would require the Energy Commission, on or before January 1, 2027, to adopt efficiency standards for data centers, as provided.
Existing law vests the Public Utilities Commission (PUC) with regulatory authority over public utilities, including electrical corporations. Existing law authorizes the PUC to fix the rates and charges for every public utility, and requires that those rates and charges be just and reasonable.
This bill would require the PUC to determine whether those costs and expenses in an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center are just and reasonable. The bill would require the PUC to minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.
Existing law, on or before January 1, 2026, and before each time thereafter that a generative artificial intelligence system or service, as defined, or a substantial modification to a generative artificial intelligence system or service, released on or after January 1, 2022, is made available to Californians for use, regardless of whether the terms of that use include compensation, requires a developer of the system or service to post on the developers internet website documentation regarding the data used to train the generative artificial intelligence system or service.
This bill would require a data center that provides computing resources to a developer for the purpose of developing a covered model, as defined, to estimate developer, before using a covered model commercially or before making a covered model available for use by a third party, to calculate the total energy used to develop the covered model and to report that information to the developer, as specified. the percentage of the total energy used to develop the covered model that was generated in California. The bill would require a developer, before using the computing resources of a data center to develop a covered model, to inform the data center of the developers intent to develop a covered model. The bill would require a developer, before using a covered model or covered system commercially or making a covered model or covered system available for use by a third party, to request its energy usage data from all data centers involved in developing the covered model and to estimate and the developer to publish on its internet website the total energy used to develop the covered model, as specified. that energy usage data, as provided.
Existing law requires the State Energy Resources Conservation and Development Commission (Energy Commission) to biennially adopt an integrated energy policy report, as specified, and to make the reports accessible to state, local, and federal entities and to the general public. Existing law requires the Energy Commission to prescribe, by regulation, energy and water conservation design standards for new residential and new nonresidential buildings, as provided.
This bill would require the Energy Commission to require operators of data centers to annually report energy consumption and performance data, as specified. The bill would require the Energy Commission to include energy consumption trends for data centers in its integrated energy policy reports. The bill would require the Energy Commission Commission, on or before January 1, 2027, to adopt energy efficiency performance standards, as described, standards for data centers. The bill would require the Energy Commission to adopt regulations to implement these requirements on or before January 1, 2027. centers, as provided.
Existing law vests the Public Utilities Commission (PUC) with regulatory authority over public utilities, including electrical corporations. Existing law authorizes the PUC to fix the rates and charges for every public utility, and requires that those rates and charges be just and reasonable.
This bill would require the PUC to determine whether those costs and expenses in an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center are just and reasonable. The bill would require the PUC to minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.
Existing law, on or before January 1, 2026, and before each time thereafter that a generative artificial intelligence system or service, as defined, or a substantial modification to a generative artificial intelligence system or service, released on or after January 1, 2022, is made available to Californians for use, regardless of whether the terms of that use include compensation, requires a developer of the system or service to post on the developers internet website documentation regarding the data used to train the generative artificial intelligence system or service.
This bill would require a data center that provides computing resources to a developer for the purpose of developing a covered model, as defined, to estimate the total energy used to develop the covered model and to report that information to the developer, as specified. The bill would require a developer, before using the computing resources of a data center to develop a covered model, to inform the data center of the developers intent to develop a covered model. The bill would require a developer, before using a covered model or covered system commercially or making a covered model or covered system available for use by a third party, to request its energy usage data from all data centers involved in developing the covered model and to estimate and publish on its internet website the total energy used to develop the covered model, as specified.
Existing law requires the State Energy Resources Conservation and Development Commission (Energy Commission) to biennially adopt an integrated energy policy report, as specified, and to make the reports accessible to state, local, and federal entities and to the general public.
This bill would require the Energy Commission to require operators of data centers to annually report energy consumption and performance data, as specified. The bill would require the Energy Commission to include energy consumption trends for data centers in its integrated energy policy reports. The bill would require the Energy Commission to adopt energy efficiency performance standards, as described, for data centers. The bill would require the Energy Commission to adopt regulations to implement these requirements on or before January 1, 2027.
Existing law vests the Public Utilities Commission (PUC) with regulatory authority over public utilities, including electrical corporations. Existing law authorizes the PUC to fix the rates and charges for every public utility, and requires that those rates and charges be just and reasonable.
This bill would require the PUC to determine whether those costs and expenses in an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center are just and reasonable. The bill would require the PUC to minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.
(a)For purposes of this section, both of the following definitions apply:
(1)Data center means a room, or a portion of a room, in a building used primarily to house information technology equipment that serves a total information technology equipment load greater than 10 kilowatts and 20 watts per square foot of conditioned floor area.
(2)Power usage effectiveness means a ratio of the total energy consumption of a data center to the energy specifically used by the information technology equipment housed in that data center.
(b)(1)The commission shall establish a process for the owner of a data center to submit the power usage effectiveness ratio for the data center to the commission on a biannual basis.
(2)The owner of a data center shall submit the power usage effectiveness ratio for the data center in a manner and timeframe specified by the commission.
(c)As part of the 2027 edition of the integrated energy policy report required pursuant to Section 25302, the commission shall include an assessment of electrical load trends for data centers. This assessment shall include all of the following:
(1)A projection of future load trends from data centers.
(2)Identification of potential net peak load demands.
(3)Recommendations for mitigating data center electricity consumption impacts on the electrical grid, including any recommended energy efficiency and demand response measures.
(d)The commission may report data center energy consumption in an aggregate basis, but shall not disclose data center energy consumption information in a manner that would result in the disclosure of personally identifiable information or energy consumption data for a specific utility customer.
SEC. 2.Section 913.18 is added to the Public Utilities Code, to read:913.18.(a)The commission shall assess the extent to which electrical corporation costs associated with new loads from data centers result in cost shifts to other electrical corporation customers. This assessment shall include all of the following:
(1)An analysis of potential electrical corporation costs associated with procurement to meet growing load demands from data centers increased energy consumption.
(2)An analysis of potential electrical corporation costs associated with the installation of new transmission and distribution assets to serve new data centers or expansions of existing data centers.
(3)To the extent that the commission finds that electrical corporation costs to serve new loads from data centers will result in substantial cost shifts to other electrical corporation customers, the commissions assessment shall also identify opportunities to prevent or mitigate these costs.
(b)On or before January 1, 2027, the commission shall submit the assessment completed pursuant to this section to the relevant policy committees of the Legislature and publicly post a copy of the assessment on the commissions internet website.
(c)Pursuant to Section 10231.5 of the Government Code, this section is repealed on January 1, 2031.
SEC. 3.The Legislature finds and declares that Section 1 of this act, which adds Section 25302.9 to the Public Resources Code, imposes a limitation on the publics right of access to the meetings of public bodies or the writings of public officials and agencies within the meaning of Section 3 of Article I of the California Constitution. Pursuant to that constitutional provision, the Legislature makes the following findings to demonstrate the interest protected by this limitation and the need for protecting that interest:
In order to protect the confidential and proprietary information of an entity subject to Section 1 of this act, it is necessary that this act limit the publics right of access to that information.
All matter omitted in this version of the bill appears in the bill as amended in the Assembly, May 23, 2025. (JR11)For purposes of this chapter, all of the following definitions apply:
(a)Artificial intelligence or AI means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.
(b)Computer room means a room within a building whose primary function is to house electronic equipment and that has a design information technology equipment power density exceeding 20 watts per square foot or 215 watts per square meters of conditioned floor area.
(c)Covered model means an AI model developed using a quantity of computing power exceeding 10^25 integer or floating-point operations.
(d)Data center means a building the primary function of which is to house one or more computer rooms with an energy demand of at least 5 megawatts.
(e)Developer means a person, partnership, state or local government agency, or corporation that designs, codes, trains, or otherwise produces a covered model.
(f)Information Technology Equipment includes computers, data storage, servers, and network or communication equipment located in a computer room.
25345.1.(a)Before using a covered model commercially, or before making a covered model available for use by a third party, a developer shall do both of the following:
(1)Estimate both of the following:
(A)The total energy, in megawatthours, used to develop the covered model.
(B)The percentage of total energy used to develop the covered model that was generated in California.
(2)Publish the energy usage data estimated pursuant to paragraph (1) on the developers internet website.
(b)On or before February 1, 2027, and annually thereafter, a developer shall do both of the following:
(1)Estimate both of the following:
(A)The total energy, in megawatthours, used by the developer to operate the covered model during the previous calendar year.
(B)The percentage of the total energy estimated pursuant to subparagraph (A) that was generated in California.
(2)Publish the energy usage data estimated pursuant to paragraph (1) on the developers internet website.
(c)Energy usage data published pursuant to this section shall be all of the following:
(1)Publicly available for as long as the developer continues to use the covered model or continues to make the covered model available for use by a third party.
(2)Provided at no cost to a user of the developers internet website.
(3)Accessible and comprehensible to a natural person with ordinary cognitive abilities.
SEC. 2.Section 25402 of the Public Resources Code is amended to read:25402.The commission shall, after one or more public hearings, do all of the following to reduce the wasteful, uneconomic, inefficient, or unnecessary consumption of energy, including the energy associated with the use of water, and to manage energy loads to help maintain electrical grid reliability:
(a)(1)Prescribe, by regulation, lighting, insulation, climate control system, and other building design and construction standards that increase efficiency in the use of energy and water for new residential and new nonresidential buildings. The commission shall periodically update the standards and adopt any revision that, in its judgment, it deems necessary. Six months after the commission certifies an energy conservation manual pursuant to subdivision (c) of Section 25402.1, a city, county, city and county, or state agency shall not issue a permit for a building unless the building satisfies the standards prescribed by the commission pursuant to this subdivision or subdivision (b) that are in effect on the date an application for a building permit is filed. Water efficiency standards adopted pursuant to this subdivision shall be demonstrated by the commission to be necessary to save energy.
(2)Before adopting a water efficiency standard for residential buildings, the Department of Housing and Community Development and the commission shall issue a joint finding whether the standard (A) is equivalent or superior in performance, in safety, and for the protection of life, health, and general welfare to standards in Title 24 of the California Code of Regulations and (B) does not unreasonably or unnecessarily impact the ability of Californians to purchase or rent affordable housing, as determined by taking account of the overall benefit derived from water efficiency standards. This subdivision does not in any way reduce the authority of the Department of Housing and Community Development to adopt standards and regulations pursuant to the State Housing Law (Part 1.5 (commencing with Section 17910) of Division 13 of the Health and Safety Code).
(3)Water efficiency standards and water conservation design standards adopted pursuant to this subdivision and subdivision (b) shall be consistent with the legislative findings of this division to ensure and maintain a reliable supply of electrical energy and be equivalent to or superior to the performance, safety, and protection of life, health, and general welfare standards contained in Title 24 of the California Code of Regulations. The commission shall consult with the members of the coordinating council as established in Section 18926 of the Health and Safety Code in the development of these standards.
(b)(1)Prescribe, by regulation, energy and water conservation design standards for new residential and new nonresidential buildings. The standards shall be performance standards and shall be promulgated in terms of energy consumption per gross square foot of floorspace, but may also include devices, systems, and techniques required to conserve energy and water. The commission shall periodically review the standards and adopt any revision that, in its judgment, it deems necessary. A building that satisfies the standards prescribed pursuant to this subdivision need not comply with the standards prescribed pursuant to subdivision (a). Water conservation design standards adopted pursuant to this subdivision shall be demonstrated by the commission to be necessary to save energy. Before adopting a water conservation design standard for residential buildings, the Department of Housing and Community Development and the commission shall issue a joint finding whether the standard (A) is equivalent or superior in performance, in safety, and for the protection of life, health, and general welfare to standards in Title 24 of the California Code of Regulations, and (B) does not unreasonably or unnecessarily impact the ability of Californians to purchase or rent affordable housing, as determined by taking account of the overall benefit derived from the water conservation design standards. This subdivision does not in any way reduce the authority of the Department of Housing and Community Development to adopt standards and regulations pursuant to the State Housing Law (Part 1.5 (commencing with Section 17910) of Division 13 of the Health and Safety Code).
(2)To increase public participation and improve the efficacy of the standards adopted pursuant to subdivision (a) and this subdivision, the commission shall, before publication of the notice of proposed action required by Section 18935 of the Health and Safety Code, involve parties who would be subject to the proposed regulations in public meetings regarding the proposed regulations. All potential affected parties shall be provided advance notice of these meetings and given an opportunity to provide written or oral comments. During these public meetings, the commission shall receive and take into consideration input from all parties concerning the parties design recommendations, cost considerations, and other factors that would affect consumers and California businesses of the proposed standard. The commission shall take into consideration before the start of the notice of proposed action any input provided during these public meetings.
(3)The standards adopted or revised pursuant to subdivision (a) and this subdivision shall be cost effective when taken in their entirety and when amortized over the economic life of the structure compared with historic practice. When determining cost-effectiveness, the commission shall consider the value of the water or energy saved, the impact on product efficacy for the consumer, and the life-cycle cost of complying with the standard. The commission shall consider other relevant factors, as required by Sections 18930 and 18935 of the Health and Safety Code, including, but not limited to, the impact on housing costs, the total statewide costs and benefits of the standard over its lifetime, the economic impact on California businesses, and any alternative approaches and their associated costs.
(4)(A)For purposes of this paragraph, data center has the same meaning as defined in Section 25345.
(B)On or before January 1, 2027, the commission shall adopt efficiency standards for data centers that do all of the following:(i)Be technologically feasible and cost effective.(ii)Align with Californias greenhouse gas emission reduction targets.(iii)Require new data centers and substantial alterations to existing data centers to incorporate load-management and load-shifting capabilities, including the ability to participate in demand response programs.(C)(B)Before initially operating as a data center and following any change that materially affects the correctness of information previously provided to the commission, a data center shall report all of the following information to the commission:
(i)The name of the data center.
(ii)Any physical addresses associated with the operation of the data center.
(iii)The internet website of the data center if the data center has an internet website.
(iv)A telephone number and email address for a point of contact for the commission to communicate with the data center.
(D)(C)The commission shall include energy consumption trends for data centers in its integrated energy policy report prepared pursuant to Section 25302.
(c)(1)(A)(i)Prescribe, by regulation, standards for minimum levels of operating efficiency, based on a reasonable use pattern, and may prescribe other cost-effective measures, including incentive programs, fleet averaging, energy and water consumption labeling not preempted by federal labeling law, and consumer education programs, to promote the use of energy- and water-efficient appliances whose use, as determined by the commission, requires a significant amount of energy or water on a statewide basis. The minimum levels of operating efficiency shall be based on feasible and attainable efficiencies or feasible improved efficiencies that will reduce the energy or water consumption growth rates. The standards for minimum levels of operating efficiency shall become effective no sooner than one year after the date of adoption or revision, unless the commission adopts a finding of good cause in which case the standards may become effective sooner than one year after the date of adoption or revision. A new appliance manufactured on or after the effective date of the standards or other cost-effective measures shall not be sold or offered for sale in the state, unless it is certified by the manufacturer of the appliance to be in compliance with the standards or other cost-effective measures. The standards or other cost-effective measures shall be drawn so that they do not result in any added total costs for consumers over the designed life of the appliances concerned. Sold or offered for sale in the state means any sale of or offer to sell an appliance for end use in the state, regardless of the sellers physical location, and includes, without limitation, internet, telephone, and mail order transactions. For purposes of this section, the Uniform Commercial CodeSales (Division 2 (commencing with Section 2101) of the Commercial Code) does not define sold or offered for sale or determine where sales or offers for sale occur.
(ii)In making a finding of good cause for purposes of a standard becoming effective sooner than one year after the date of adoption or revision pursuant to clause (i), the commission shall consider, among other things, the following factors:
(I)The availability of products on the market that meet the proposed standard.
(II)The impact of an earlier effective date on manufacturers.
(III)The health and safety benefits of an earlier effective date.
(IV)The impact on innovation resulting from a one-year delay between the date of adoption or revision and the effective date of the standard.
(V)The concerns raised by comments, provided to the commission pursuant to subparagraph (B), for an earlier effective date.
(B)To increase public participation and improve the efficacy of the standards adopted pursuant to this subdivision, the commission shall, before publication of the notice of proposed action required by Section 18935 of the Health and Safety Code, involve parties who would be subject to the proposed regulations in public meetings regarding the proposed regulations. All potential affected parties shall be provided advance notice of these meetings and given an opportunity to provide written or oral comments. During these public meetings, the commission shall receive and take into consideration input from all parties concerning the parties design recommendations, cost considerations, and other factors that would affect consumers and California businesses of the proposed standard. The commission shall take into consideration before the start of the notice of proposed action any input provided during these public meetings.
(C)The standards adopted or revised pursuant to this subdivision shall not result in any added total costs for consumers over the designed life of the appliances concerned. When determining cost-effectiveness, the commission shall consider the value of the water or energy saved, the impact on product efficacy for the consumer, and the life-cycle cost to the consumer of complying with the standard. The commission shall consider other relevant factors, as required by Sections 11346.5 and 11357 of the Government Code, including, but not limited to, the impact on housing costs, the total statewide costs and benefits of the standard over its lifetime, the economic impact on California businesses, and any alternative approaches and their associated costs.
(2)A new appliance, except for a plumbing fitting, regulated under paragraph (1), that is manufactured on or after July 1, 1984, shall not be sold, or offered for sale, in the state, unless the date of the manufacture is permanently displayed in an accessible place on that appliance.
(3)During the period of five years after the commission has adopted a standard for a particular appliance under paragraph (1), an increase or decrease in the minimum level of operating efficiency required by the standard for that appliance shall not become effective, unless the commission adopts other cost-effective measures for that appliance.
(4)Neither the commission nor any other state agency shall take any action to decrease any standard adopted under this subdivision on or before June 30, 1985, prescribing minimum levels of operating efficiency or other energy conservation measures for any appliance, unless the commission finds by a four-fifths vote that a decrease is of benefit to ratepayers, and that there is significant evidence of changed circumstances. At least one year before the adoption or amendment of a standard for an appliance, the commission shall notify the Legislature of its intent, and the justification to adopt or amend a standard for the appliance. Notwithstanding paragraph (3) and this paragraph, the commission may do any of the following:
(A)Increase the minimum level of operating efficiency in an existing standard up to the level of the National Voluntary Consensus Standards 90, adopted by the American Society of Heating, Refrigerating and Air-Conditioning Engineers or, for appliances not covered by that standard, up to the level established in a similar nationwide consensus standard.
(B)Change the measure or rating of efficiency of any standard, if the minimum level of operating efficiency remains substantially the same.
(C)Adjust the minimum level of operating efficiency in an existing standard to reflect changes in test procedures that the standards require manufacturers to use in certifying compliance, if the minimum level of operating efficiency remains substantially the same.
(D)Readopt a standard preempted, enjoined, or otherwise found legally defective by an administrative agency or a lower court, if final legal action determines that the standard is valid and if the standard that is readopted is not more stringent than the standard that was found to be defective or preempted.
(E)Adopt or amend any existing or new standard at any level of operating efficiency, if the Governor has declared an energy emergency as described in Section 8558 of the Government Code.
(d)Recommend minimum standards of efficiency for the operation of a new facility at a particular site that are technically and economically feasible. A site and related facility shall not be certified pursuant to Chapter 6 (commencing with Section 25500), unless the applicant certifies that standards recommended by the commission have been considered, which certification shall include a statement specifying the extent to which conformance with the recommended standards will be achieved.
(e)(1)Adopt, by regulation, and periodically update, standards for appliances to facilitate the deployment of flexible demand technologies. These regulations may include labeling provisions to promote the use of appliances with flexible demand capabilities. The flexible demand appliance standards shall be based on feasible and attainable efficiencies or feasible improvements that will enable appliance operations to be scheduled, shifted, or curtailed to enhance grid reliability, reduce emissions of greenhouse gases associated with electricity generation, or both enhancement and reduction. The standards shall become effective no sooner than one year after the date of their adoption or updating.
(2)In adopting the flexible demand appliance standards, the commission shall consider the National Institute of Standards and Technologys reliability and cybersecurity protocols, or other cybersecurity protocols that are equally or more protective, and shall adopt, at a minimum, the North American Electric Reliability Corporations Critical Infrastructure Protection standards.
(3)The flexible demand appliance standards shall be cost effective. When determining cost-effectiveness, solely for purposes of this subdivision, the commission may consider, as appropriate, the cost of flexible demand appliances compared to nonflexible demand appliances, the value of increased or decreased emissions of greenhouse gases associated with the timing of an appliances use, the life-cycle cost to the consumer from using a product that complies with the standard, and the life-cycle costs and benefits to consumers, including the ability to conserve energy and better align consumer and electric system demand. The commission shall consider other relevant factors, as required by Sections 11346.5 and 11357 of the Government Code, including, but not limited to, the impact on housing costs, the total statewide costs and benefits of the standard over its lifetime, the economic impact on California businesses, and alternative approaches and their associated costs.
(4)The commission shall consult with the Public Utilities Commission, load-serving entities, and local publicly owned electric utilities to better align the flexible demand appliance standards with demand response programs administered by the state and load-serving entities and local publicly owned electric utilities and to incentivize the deployment of flexible demand appliances.
(5)The flexible demand appliance standards shall prioritize all of the following:
(A)Appliances that can more conveniently have their electrical demand controlled by load-management technology and third-party load-management programs.
(B)Appliances with load-management technology options that are readily available.
(C)Appliances that have a user-friendly interface and follow a straightforward setup and connection process, such as remote setup by means of an internet website or application.
(D)Appliances with load-management technology options that follow simple standards for third-party direct operation of the appliances.
(E)Appliances that are interoperable or open source.
(6)On or before January 1, 2021, and as necessary thereafter, the commission shall include as part of each integrated energy policy report adopted pursuant to Chapter 4 (commencing with Section 25300) a description of any actions it has taken pursuant to this subdivision and the flexible demand appliance standards cost to consumers.
(7)For purposes of this subdivision, all of the following definitions apply:
(A)Flexible demand means the capability to schedule, shift, or curtail the electrical demand of a load-serving entitys customer or a local publicly owned electric utilitys customer through direct action by the customer or through action by a third party, the load-serving entity, or a grid balancing authority, with the customers consent.
(B)Load-serving entity has the same meaning as defined in Section 380 of the Public Utilities Code.
(C)Local publicly owned electric utility has the same meaning as defined in Section 224.3 of the Public Utilities Code.
SEC. 3.Section 451.4 is added to the Public Utilities Code, to read:451.4.(a)For purposes of this section, data center has the same meaning as defined in Section 25345 of the Public Resources Code.
(b)In an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center, the commission shall determine whether those costs and expenses are just and reasonable in accordance with Section 451.
(c)In determining whether costs and expenses are just and reasonable pursuant to subdivision (b), the commission shall minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.
For purposes of this chapter, all of the following definitions apply:
(a)Artificial intelligence or AI means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.
(b)Computer room means a room within a building whose primary function is to house electronic equipment and that has a design information technology equipment power density exceeding 20 watts per square foot or 215 watts per square meters of conditioned floor area.
(c)Covered model means an AI model developed using a quantity of computing power exceeding 10^25 integer or floating-point operations.
(d)Data center means a building the primary function of which is to house one or more computer rooms with an energy demand of at least 5 megawatts.
(e)Developer means a person, partnership, state or local government agency, or corporation that designs, codes, trains, or otherwise produces a covered model.
(f)Information Technology Equipment includes computers, data storage, servers, and network or communication equipment located in a computer room.
25345.1.(a)Before using a covered model commercially, or before making a covered model available for use by a third party, a developer shall do both of the following:
(a)Calculate(1)Estimate both of the following:
(1)(A)The total energy, in megawatthours, used to develop the covered model.
(2)(B)The percentage of total energy used to develop the covered model that was generated in California.
(b)(1)(2)Publish the energy usage data estimated pursuant to subdivision (a) paragraph (1) on the developers internet website.
(b)On or before February 1, 2027, and annually thereafter, a developer shall do both of the following:
(1)Estimate both of the following:
(A)The total energy, in megawatthours, used by the developer to operate the covered model during the previous calendar year.
(B)The percentage of the total energy estimated pursuant to subparagraph (A) that was generated in California.
(2)Publish the energy usage data estimated pursuant to paragraph (1) on the developers internet website.
(2)(c)Energy usage data published pursuant to paragraph (1) this section shall be all of the following:
(A)(1)Publicly available for as long as the developer continues to use the covered model or continues to make the covered model available for use by a third party.
(B)(2)Provided at no cost to a user of the developers internet website.
(C)(3)Accessible and comprehensible to a natural person with ordinary cognitive abilities.
SEC. 2.Section 25402 of the Public Resources Code is amended to read:25402.The commission shall, after one or more public hearings, do all of the following to reduce the wasteful, uneconomic, inefficient, or unnecessary consumption of energy, including the energy associated with the use of water, and to manage energy loads to help maintain electrical grid reliability:
(a)(1)Prescribe, by regulation, lighting, insulation, climate control system, and other building design and construction standards that increase efficiency in the use of energy and water for new residential and new nonresidential buildings. The commission shall periodically update the standards and adopt any revision that, in its judgment, it deems necessary. Six months after the commission certifies an energy conservation manual pursuant to subdivision (c) of Section 25402.1, a city, county, city and county, or state agency shall not issue a permit for a building unless the building satisfies the standards prescribed by the commission pursuant to this subdivision or subdivision (b) that are in effect on the date an application for a building permit is filed. Water efficiency standards adopted pursuant to this subdivision shall be demonstrated by the commission to be necessary to save energy.
(2)Before adopting a water efficiency standard for residential buildings, the Department of Housing and Community Development and the commission shall issue a joint finding whether the standard (A) is equivalent or superior in performance, in safety, and for the protection of life, health, and general welfare to standards in Title 24 of the California Code of Regulations and (B) does not unreasonably or unnecessarily impact the ability of Californians to purchase or rent affordable housing, as determined by taking account of the overall benefit derived from water efficiency standards. This subdivision does not in any way reduce the authority of the Department of Housing and Community Development to adopt standards and regulations pursuant to the State Housing Law (Part 1.5 (commencing with Section 17910) of Division 13 of the Health and Safety Code).
(3)Water efficiency standards and water conservation design standards adopted pursuant to this subdivision and subdivision (b) shall be consistent with the legislative findings of this division to ensure and maintain a reliable supply of electrical energy and be equivalent to or superior to the performance, safety, and protection of life, health, and general welfare standards contained in Title 24 of the California Code of Regulations. The commission shall consult with the members of the coordinating council as established in Section 18926 of the Health and Safety Code in the development of these standards.
(b)(1)Prescribe, by regulation, energy and water conservation design standards for new residential and new nonresidential buildings. The standards shall be performance standards and shall be promulgated in terms of energy consumption per gross square foot of floorspace, but may also include devices, systems, and techniques required to conserve energy and water. The commission shall periodically review the standards and adopt any revision that, in its judgment, it deems necessary. A building that satisfies the standards prescribed pursuant to this subdivision need not comply with the standards prescribed pursuant to subdivision (a). Water conservation design standards adopted pursuant to this subdivision shall be demonstrated by the commission to be necessary to save energy. Before adopting a water conservation design standard for residential buildings, the Department of Housing and Community Development and the commission shall issue a joint finding whether the standard (A) is equivalent or superior in performance, in safety, and for the protection of life, health, and general welfare to standards in Title 24 of the California Code of Regulations, and (B) does not unreasonably or unnecessarily impact the ability of Californians to purchase or rent affordable housing, as determined by taking account of the overall benefit derived from the water conservation design standards. This subdivision does not in any way reduce the authority of the Department of Housing and Community Development to adopt standards and regulations pursuant to the State Housing Law (Part 1.5 (commencing with Section 17910) of Division 13 of the Health and Safety Code).
(2)To increase public participation and improve the efficacy of the standards adopted pursuant to subdivision (a) and this subdivision, the commission shall, before publication of the notice of proposed action required by Section 18935 of the Health and Safety Code, involve parties who would be subject to the proposed regulations in public meetings regarding the proposed regulations. All potential affected parties shall be provided advance notice of these meetings and given an opportunity to provide written or oral comments. During these public meetings, the commission shall receive and take into consideration input from all parties concerning the parties design recommendations, cost considerations, and other factors that would affect consumers and California businesses of the proposed standard. The commission shall take into consideration before the start of the notice of proposed action any input provided during these public meetings.
(3)The standards adopted or revised pursuant to subdivision (a) and this subdivision shall be cost effective when taken in their entirety and when amortized over the economic life of the structure compared with historic practice. When determining cost-effectiveness, the commission shall consider the value of the water or energy saved, the impact on product efficacy for the consumer, and the life-cycle cost of complying with the standard. The commission shall consider other relevant factors, as required by Sections 18930 and 18935 of the Health and Safety Code, including, but not limited to, the impact on housing costs, the total statewide costs and benefits of the standard over its lifetime, the economic impact on California businesses, and any alternative approaches and their associated costs.
(4)(A)For purposes of this paragraph, data center has the same meaning as defined in Section 25345.
(B)On or before January 1, 2027, the commission shall adopt efficiency standards for data centers that do all of the following:
(i)Be technologically feasible and cost effective.
(ii)Align with Californias greenhouse gas emission reduction targets.
(iii)Require new data centers and substantial alterations to existing data centers to incorporate load-management and load-shifting capabilities, including the ability to participate in demand response programs.
(C)Before initially operating as a data center and following any change that materially affects the correctness of information previously provided to the commission, a data center shall report all of the following information to the commission:
(i)The name of the data center.
(ii)Any physical addresses associated with the operation of the data center.
(iii)The internet website of the data center if the data center has a an internet website.
(iv)A telephone number and email address for a point of contact for the commission to communicate with the data center.
(D)The commission shall include energy consumption trends for data centers in its integrated energy policy report prepared pursuant to Section 25302.
(c)(1)(A)(i)Prescribe, by regulation, standards for minimum levels of operating efficiency, based on a reasonable use pattern, and may prescribe other cost-effective measures, including incentive programs, fleet averaging, energy and water consumption labeling not preempted by federal labeling law, and consumer education programs, to promote the use of energy- and water-efficient appliances whose use, as determined by the commission, requires a significant amount of energy or water on a statewide basis. The minimum levels of operating efficiency shall be based on feasible and attainable efficiencies or feasible improved efficiencies that will reduce the energy or water consumption growth rates. The standards for minimum levels of operating efficiency shall become effective no sooner than one year after the date of adoption or revision, unless the commission adopts a finding of good cause in which case the standards may become effective sooner than one year after the date of adoption or revision. A new appliance manufactured on or after the effective date of the standards or other cost-effective measures shall not be sold or offered for sale in the state, unless it is certified by the manufacturer of the appliance to be in compliance with the standards or other cost-effective measures. The standards or other cost-effective measures shall be drawn so that they do not result in any added total costs for consumers over the designed life of the appliances concerned. Sold or offered for sale in the state means any sale of or offer to sell an appliance for end use in the state, regardless of the sellers physical location, and includes, without limitation, internet, telephone, and mail order transactions. For purposes of this section, the Uniform Commercial CodeSales (Division 2 (commencing with Section 2101) of the Commercial Code) does not define sold or offered for sale or determine where sales or offers for sale occur.
(ii)In making a finding of good cause for purposes of a standard becoming effective sooner than one year after the date of adoption or revision pursuant to clause (i), the commission shall consider, among other things, the following factors:
(I)The availability of products on the market that meet the proposed standard.
(II)The impact of an earlier effective date on manufacturers.
(III)The health and safety benefits of an earlier effective date.
(IV)The impact on innovation resulting from a one-year delay between the date of adoption or revision and the effective date of the standard.
(V)The concerns raised by comments, provided to the commission pursuant to subparagraph (B), for an earlier effective date.
(B)To increase public participation and improve the efficacy of the standards adopted pursuant to this subdivision, the commission shall, before publication of the notice of proposed action required by Section 18935 of the Health and Safety Code, involve parties who would be subject to the proposed regulations in public meetings regarding the proposed regulations. All potential affected parties shall be provided advance notice of these meetings and given an opportunity to provide written or oral comments. During these public meetings, the commission shall receive and take into consideration input from all parties concerning the parties design recommendations, cost considerations, and other factors that would affect consumers and California businesses of the proposed standard. The commission shall take into consideration before the start of the notice of proposed action any input provided during these public meetings.
(C)The standards adopted or revised pursuant to this subdivision shall not result in any added total costs for consumers over the designed life of the appliances concerned. When determining cost-effectiveness, the commission shall consider the value of the water or energy saved, the impact on product efficacy for the consumer, and the life-cycle cost to the consumer of complying with the standard. The commission shall consider other relevant factors, as required by Sections 11346.5 and 11357 of the Government Code, including, but not limited to, the impact on housing costs, the total statewide costs and benefits of the standard over its lifetime, the economic impact on California businesses, and any alternative approaches and their associated costs.
(2)A new appliance, except for a plumbing fitting, regulated under paragraph (1), that is manufactured on or after July 1, 1984, shall not be sold, or offered for sale, in the state, unless the date of the manufacture is permanently displayed in an accessible place on that appliance.
(3)During the period of five years after the commission has adopted a standard for a particular appliance under paragraph (1), an increase or decrease in the minimum level of operating efficiency required by the standard for that appliance shall not become effective, unless the commission adopts other cost-effective measures for that appliance.
(4)Neither the commission nor any other state agency shall take any action to decrease any standard adopted under this subdivision on or before June 30, 1985, prescribing minimum levels of operating efficiency or other energy conservation measures for any appliance, unless the commission finds by a four-fifths vote that a decrease is of benefit to ratepayers, and that there is significant evidence of changed circumstances. At least one year before the adoption or amendment of a standard for an appliance, the commission shall notify the Legislature of its intent, and the justification to adopt or amend a standard for the appliance. Notwithstanding paragraph (3) and this paragraph, the commission may do any of the following:
(A)Increase the minimum level of operating efficiency in an existing standard up to the level of the National Voluntary Consensus Standards 90, adopted by the American Society of Heating, Refrigerating and Air-Conditioning Engineers or, for appliances not covered by that standard, up to the level established in a similar nationwide consensus standard.
(B)Change the measure or rating of efficiency of any standard, if the minimum level of operating efficiency remains substantially the same.
(C)Adjust the minimum level of operating efficiency in an existing standard to reflect changes in test procedures that the standards require manufacturers to use in certifying compliance, if the minimum level of operating efficiency remains substantially the same.
(D)Readopt a standard preempted, enjoined, or otherwise found legally defective by an administrative agency or a lower court, if final legal action determines that the standard is valid and if the standard that is readopted is not more stringent than the standard that was found to be defective or preempted.
(E)Adopt or amend any existing or new standard at any level of operating efficiency, if the Governor has declared an energy emergency as described in Section 8558 of the Government Code.
(d)Recommend minimum standards of efficiency for the operation of a new facility at a particular site that are technically and economically feasible. A site and related facility shall not be certified pursuant to Chapter 6 (commencing with Section 25500), unless the applicant certifies that standards recommended by the commission have been considered, which certification shall include a statement specifying the extent to which conformance with the recommended standards will be achieved.
(e)(1)Adopt, by regulation, and periodically update, standards for appliances to facilitate the deployment of flexible demand technologies. These regulations may include labeling provisions to promote the use of appliances with flexible demand capabilities. The flexible demand appliance standards shall be based on feasible and attainable efficiencies or feasible improvements that will enable appliance operations to be scheduled, shifted, or curtailed to enhance grid reliability, reduce emissions of greenhouse gases associated with electricity generation, or both enhancement and reduction. The standards shall become effective no sooner than one year after the date of their adoption or updating.
(2)In adopting the flexible demand appliance standards, the commission shall consider the National Institute of Standards and Technologys reliability and cybersecurity protocols, or other cybersecurity protocols that are equally or more protective, and shall adopt, at a minimum, the North American Electric Reliability Corporations Critical Infrastructure Protection standards.
(3)The flexible demand appliance standards shall be cost effective. When determining cost-effectiveness, solely for purposes of this subdivision, the commission may consider, as appropriate, the cost of flexible demand appliances compared to nonflexible demand appliances, the value of increased or decreased emissions of greenhouse gases associated with the timing of an appliances use, the life-cycle cost to the consumer from using a product that complies with the standard, and the life-cycle costs and benefits to consumers, including the ability to conserve energy and better align consumer and electric system demand. The commission shall consider other relevant factors, as required by Sections 11346.5 and 11357 of the Government Code, including, but not limited to, the impact on housing costs, the total statewide costs and benefits of the standard over its lifetime, the economic impact on California businesses, and alternative approaches and their associated costs.
(4)The commission shall consult with the Public Utilities Commission, load-serving entities, and local publicly owned electric utilities to better align the flexible demand appliance standards with demand response programs administered by the state and load-serving entities and local publicly owned electric utilities and to incentivize the deployment of flexible demand appliances.
(5)The flexible demand appliance standards shall prioritize all of the following:
(A)Appliances that can more conveniently have their electrical demand controlled by load-management technology and third-party load-management programs.
(B)Appliances with load-management technology options that are readily available.
(C)Appliances that have a user-friendly interface and follow a straightforward setup and connection process, such as remote setup by means of an internet website or application.
(D)Appliances with load-management technology options that follow simple standards for third-party direct operation of the appliances.
(E)Appliances that are interoperable or open source.
(6)On or before January 1, 2021, and as necessary thereafter, the commission shall include as part of each integrated energy policy report adopted pursuant to Chapter 4 (commencing with Section 25300) a description of any actions it has taken pursuant to this subdivision and the flexible demand appliance standards cost to consumers.
(7)For purposes of this subdivision, all of the following definitions apply:
(A)Flexible demand means the capability to schedule, shift, or curtail the electrical demand of a load-serving entitys customer or a local publicly owned electric utilitys customer through direct action by the customer or through action by a third party, the load-serving entity, or a grid balancing authority, with the customers consent.
(B)Load-serving entity has the same meaning as defined in Section 380 of the Public Utilities Code.
(C)Local publicly owned electric utility has the same meaning as defined in Section 224.3 of the Public Utilities Code.
SEC. 3.Section 451.4 is added to the Public Utilities Code, to read:451.4.(a)For purposes of this section, data center has the same meaning as defined in Section 25345 of the Public Resources Code.
(b)In an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center, the commission shall determine whether those costs and expenses are just and reasonable in accordance with Section 451.
(c)In determining whether costs and expenses are just and reasonable pursuant to subdivision (b), the commission shall minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.
For purposes of this chapter, all of the following definitions apply:
(a)Artificial intelligence or AI means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.
(b)Computer room means a room within a building whose primary function is to house electronic equipment and that has a design information technology equipment power density exceeding 20 watts per square foot or 215 watts per square meters of conditioned floor area.
(b)(c)Covered model means an AI model developed using a quantity of computing power exceeding 10^25 integer or floating-point operations.
(c)Covered system means an AI system that incorporates a covered model. (d)(1)Data center means a facility that houses computing infrastructure, including graphics and central processing units, servers, storage devices, networking equipment, and associated power and cooling systems, for the primary purpose of processing, storing, or distributing electronic data.(2)Data center does not include a facility with an annual energy consumption of 10 megawatthours or less.(d)Data center means a building the primary function of which is to house one or more computer rooms with an energy demand of at least 5 megawatts.
(e)Developer means a person, partnership, state or local government agency, or corporation that designs, codes, trains, or otherwise produces a covered model.
(f)Information Technology Equipment includes computers, data storage, servers, and network or communication equipment located in a computer room.
25345.1.A data center that provides computing resources to a developer for the purpose of developing a covered model shall do both of the following:(a)Estimate the total energy, in megawatthours, used to develop the covered model.(b)Report the estimated total energy used to develop the covered model to the developer under both of the following circumstances:(1)Upon the termination of a contract facilitating the transfer of computing resources from the data center to the developer.(2)Upon receiving a request from the developer pursuant to paragraph (1) of subdivision (b) of Section 25345.2.25345.2.(a)Before using the computing resources of a data center to develop a covered model, a developer shall inform the data center of the developers intent to develop a covered model.(b)25345.1.Before using a covered model or covered system commercially, or before making a covered model or covered system available for use by a third party, a developer shall do both of the following:
(1)Request the developers energy usage data from all data centers involved in developing the covered model.(2)Estimate and publish on the developers internet website the total energy, in megawatthours, used to develop the covered model.(a)Calculate both of the following:
(1)The total energy, in megawatthours, used to develop the covered model.
(2)The percentage of total energy used to develop the covered model that was generated in California.
(b)(1)Publish the energy usage data estimated pursuant to subdivision (a) on the developers internet website.
(c)(2)Energy usage data published pursuant to subdivision (b) paragraph (1) shall be all of the following:
(1)(A)Publicly available for as long as the developer continues to use the covered model or covered system, or continues to make the covered model or covered system available for use by a third party.
(2)(B)Provided at no cost to a user of the developers internet website.
(3)(C)Accessible and comprehensible to a natural person with ordinary cognitive abilities.
25345.3.(a)(1)The commission shall require operators of data centers to annually report energy consumption and performance data to the commission.(2)Reports submitted to the commission pursuant to this subdivision shall include all of the following:(A)Total annual energy consumption, including electricity usage from grid and nongrid sources, peak factors, and load profiles.(B)Energy usage effectiveness or equivalent energy efficiency metrics.(C)The percentage of energy supplied by renewable or carbon-free resources.(D)Annual energy consumption resulting from the development of covered models.(E)Any other information deemed necessary by the commission to assess grid impacts. (3)Data centers reporting pursuant to this section shall submit data in a form and manner prescribed by the commission.(b)The commission shall include energy consumption trends for data centers in its integrated energy policy report prepared pursuant to Section 25302.(c)On or before January 1, 2027, the commission shall adopt regulations to implement this section.25345.4.(a)The commission shall adopt energy efficiency performance standards for data centers.(b)The standards adopted pursuant to this section shall do all of the following:(1)Prioritize cost effectiveness, technological feasibility, and alignment with Californias greenhouse gas emission reduction targets.(2)Consider best practices for energy usage effectiveness, energy management systems, and the use of renewable energy resources.(3)Require new data centers and substantial alterations to existing data centers to incorporate load-management and load-shifting capabilities, including the ability to participate in demand response programs.(c)(1)On or before January 1, 2027, the commission shall adopt regulations to implement this section.(2)In adopting regulations pursuant to this subdivision, the commission shall consult with data center operators, utilities, technical experts, and other industry stakeholders.SEC. 2.Section 25402 of the Public Resources Code is amended to read:25402.The commission shall, after one or more public hearings, do all of the following to reduce the wasteful, uneconomic, inefficient, or unnecessary consumption of energy, including the energy associated with the use of water, and to manage energy loads to help maintain electrical grid reliability:
(a)(1)Prescribe, by regulation, lighting, insulation, climate control system, and other building design and construction standards that increase efficiency in the use of energy and water for new residential and new nonresidential buildings. The commission shall periodically update the standards and adopt any revision that, in its judgment, it deems necessary. Six months after the commission certifies an energy conservation manual pursuant to subdivision (c) of Section 25402.1, a city, county, city and county, or state agency shall not issue a permit for a building unless the building satisfies the standards prescribed by the commission pursuant to this subdivision or subdivision (b) that are in effect on the date an application for a building permit is filed. Water efficiency standards adopted pursuant to this subdivision shall be demonstrated by the commission to be necessary to save energy.
(2)Before adopting a water efficiency standard for residential buildings, the Department of Housing and Community Development and the commission shall issue a joint finding whether the standard (A) is equivalent or superior in performance, in safety, and for the protection of life, health, and general welfare to standards in Title 24 of the California Code of Regulations and (B) does not unreasonably or unnecessarily impact the ability of Californians to purchase or rent affordable housing, as determined by taking account of the overall benefit derived from water efficiency standards. This subdivision does not in any way reduce the authority of the Department of Housing and Community Development to adopt standards and regulations pursuant to the State Housing Law (Part 1.5 (commencing with Section 17910) of Division 13 of the Health and Safety Code).
(3)Water efficiency standards and water conservation design standards adopted pursuant to this subdivision and subdivision (b) shall be consistent with the legislative findings of this division to ensure and maintain a reliable supply of electrical energy and be equivalent to or superior to the performance, safety, and protection of life, health, and general welfare standards contained in Title 24 of the California Code of Regulations. The commission shall consult with the members of the coordinating council as established in Section 18926 of the Health and Safety Code in the development of these standards.
(b)(1)Prescribe, by regulation, energy and water conservation design standards for new residential and new nonresidential buildings. The standards shall be performance standards and shall be promulgated in terms of energy consumption per gross square foot of floorspace, but may also include devices, systems, and techniques required to conserve energy and water. The commission shall periodically review the standards and adopt any revision that, in its judgment, it deems necessary. A building that satisfies the standards prescribed pursuant to this subdivision need not comply with the standards prescribed pursuant to subdivision (a). Water conservation design standards adopted pursuant to this subdivision shall be demonstrated by the commission to be necessary to save energy. Before adopting a water conservation design standard for residential buildings, the Department of Housing and Community Development and the commission shall issue a joint finding whether the standard (A) is equivalent or superior in performance, in safety, and for the protection of life, health, and general welfare to standards in Title 24 of the California Code of Regulations, and (B) does not unreasonably or unnecessarily impact the ability of Californians to purchase or rent affordable housing, as determined by taking account of the overall benefit derived from the water conservation design standards. This subdivision does not in any way reduce the authority of the Department of Housing and Community Development to adopt standards and regulations pursuant to the State Housing Law (Part 1.5 (commencing with Section 17910) of Division 13 of the Health and Safety Code).
(2)To increase public participation and improve the efficacy of the standards adopted pursuant to subdivision (a) and this subdivision, the commission shall, before publication of the notice of proposed action required by Section 18935 of the Health and Safety Code, involve parties who would be subject to the proposed regulations in public meetings regarding the proposed regulations. All potential affected parties shall be provided advance notice of these meetings and given an opportunity to provide written or oral comments. During these public meetings, the commission shall receive and take into consideration input from all parties concerning the parties design recommendations, cost considerations, and other factors that would affect consumers and California businesses of the proposed standard. The commission shall take into consideration before the start of the notice of proposed action any input provided during these public meetings.
(3)The standards adopted or revised pursuant to subdivision (a) and this subdivision shall be cost effective when taken in their entirety and when amortized over the economic life of the structure compared with historic practice. When determining cost-effectiveness, the commission shall consider the value of the water or energy saved, the impact on product efficacy for the consumer, and the life-cycle cost of complying with the standard. The commission shall consider other relevant factors, as required by Sections 18930 and 18935 of the Health and Safety Code, including, but not limited to, the impact on housing costs, the total statewide costs and benefits of the standard over its lifetime, the economic impact on California businesses, and any alternative approaches and their associated costs.
(4)(A)For purposes of this paragraph, data center has the same meaning as defined in Section 25345.
(B)On or before January 1, 2027, the commission shall adopt efficiency standards for data centers that do all of the following:
(i)Be technologically feasible and cost effective.
(ii)Align with Californias greenhouse gas emission reduction targets.
(iii)Require new data centers and substantial alterations to existing data centers to incorporate load-management and load-shifting capabilities, including the ability to participate in demand response programs.
(C)Before initially operating as a data center and following any change that materially affects the correctness of information previously provided to the commission, a data center shall report all of the following information to the commission:
(i)The name of the data center.
(ii)Any physical addresses associated with the operation of the data center.
(iii)The internet website of the data center if the data center has a internet website.
(iv)A telephone number and email address for a point of contact for the commission to communicate with the data center.
(D)The commission shall include energy consumption trends for data centers in its integrated energy policy report prepared pursuant to Section 25302.
(c)(1)(A)(i)Prescribe, by regulation, standards for minimum levels of operating efficiency, based on a reasonable use pattern, and may prescribe other cost-effective measures, including incentive programs, fleet averaging, energy and water consumption labeling not preempted by federal labeling law, and consumer education programs, to promote the use of energy- and water-efficient appliances whose use, as determined by the commission, requires a significant amount of energy or water on a statewide basis. The minimum levels of operating efficiency shall be based on feasible and attainable efficiencies or feasible improved efficiencies that will reduce the energy or water consumption growth rates. The standards for minimum levels of operating efficiency shall become effective no sooner than one year after the date of adoption or revision, unless the commission adopts a finding of good cause in which case the standards may become effective sooner than one year after the date of adoption or revision. A new appliance manufactured on or after the effective date of the standards or other cost-effective measures shall not be sold or offered for sale in the state, unless it is certified by the manufacturer of the appliance to be in compliance with the standards or other cost-effective measures. The standards or other cost-effective measures shall be drawn so that they do not result in any added total costs for consumers over the designed life of the appliances concerned. Sold or offered for sale in the state means any sale of or offer to sell an appliance for end use in the state, regardless of the sellers physical location, and includes, without limitation, internet, telephone, and mail order transactions. For purposes of this section, the Uniform Commercial CodeSales (Division 2 (commencing with Section 2101) of the Commercial Code) does not define sold or offered for sale or determine where sales or offers for sale occur.
(ii)In making a finding of good cause for purposes of a standard becoming effective sooner than one year after the date of adoption or revision pursuant to clause (i), the commission shall consider, among other things, the following factors:
(I)The availability of products on the market that meet the proposed standard.
(II)The impact of an earlier effective date on manufacturers.
(III)The health and safety benefits of an earlier effective date.
(IV)The impact on innovation resulting from a one-year delay between the date of adoption or revision and the effective date of the standard.
(V)The concerns raised by comments, provided to the commission pursuant to subparagraph (B), for an earlier effective date.
(B)To increase public participation and improve the efficacy of the standards adopted pursuant to this subdivision, the commission shall, before publication of the notice of proposed action required by Section 18935 of the Health and Safety Code, involve parties who would be subject to the proposed regulations in public meetings regarding the proposed regulations. All potential affected parties shall be provided advance notice of these meetings and given an opportunity to provide written or oral comments. During these public meetings, the commission shall receive and take into consideration input from all parties concerning the parties design recommendations, cost considerations, and other factors that would affect consumers and California businesses of the proposed standard. The commission shall take into consideration before the start of the notice of proposed action any input provided during these public meetings.
(C)The standards adopted or revised pursuant to this subdivision shall not result in any added total costs for consumers over the designed life of the appliances concerned. When determining cost-effectiveness, the commission shall consider the value of the water or energy saved, the impact on product efficacy for the consumer, and the life-cycle cost to the consumer of complying with the standard. The commission shall consider other relevant factors, as required by Sections 11346.5 and 11357 of the Government Code, including, but not limited to, the impact on housing costs, the total statewide costs and benefits of the standard over its lifetime, the economic impact on California businesses, and any alternative approaches and their associated costs.
(2)A new appliance, except for a plumbing fitting, regulated under paragraph (1), that is manufactured on or after July 1, 1984, shall not be sold, or offered for sale, in the state, unless the date of the manufacture is permanently displayed in an accessible place on that appliance.
(3)During the period of five years after the commission has adopted a standard for a particular appliance under paragraph (1), an increase or decrease in the minimum level of operating efficiency required by the standard for that appliance shall not become effective, unless the commission adopts other cost-effective measures for that appliance.
(4)Neither the commission nor any other state agency shall take any action to decrease any standard adopted under this subdivision on or before June 30, 1985, prescribing minimum levels of operating efficiency or other energy conservation measures for any appliance, unless the commission finds by a four-fifths vote that a decrease is of benefit to ratepayers, and that there is significant evidence of changed circumstances. At least one year before the adoption or amendment of a standard for an appliance, the commission shall notify the Legislature of its intent, and the justification to adopt or amend a standard for the appliance. Notwithstanding paragraph (3) and this paragraph, the commission may do any of the following:
(A)Increase the minimum level of operating efficiency in an existing standard up to the level of the National Voluntary Consensus Standards 90, adopted by the American Society of Heating, Refrigerating and Air-Conditioning Engineers or, for appliances not covered by that standard, up to the level established in a similar nationwide consensus standard.
(B)Change the measure or rating of efficiency of any standard, if the minimum level of operating efficiency remains substantially the same.
(C)Adjust the minimum level of operating efficiency in an existing standard to reflect changes in test procedures that the standards require manufacturers to use in certifying compliance, if the minimum level of operating efficiency remains substantially the same.
(D)Readopt a standard preempted, enjoined, or otherwise found legally defective by an administrative agency or a lower court, if final legal action determines that the standard is valid and if the standard that is readopted is not more stringent than the standard that was found to be defective or preempted.
(E)Adopt or amend any existing or new standard at any level of operating efficiency, if the Governor has declared an energy emergency as described in Section 8558 of the Government Code.
(d)Recommend minimum standards of efficiency for the operation of a new facility at a particular site that are technically and economically feasible. A site and related facility shall not be certified pursuant to Chapter 6 (commencing with Section 25500), unless the applicant certifies that standards recommended by the commission have been considered, which certification shall include a statement specifying the extent to which conformance with the recommended standards will be achieved.
(e)(1)Adopt, by regulation, and periodically update, standards for appliances to facilitate the deployment of flexible demand technologies. These regulations may include labeling provisions to promote the use of appliances with flexible demand capabilities. The flexible demand appliance standards shall be based on feasible and attainable efficiencies or feasible improvements that will enable appliance operations to be scheduled, shifted, or curtailed to enhance grid reliability, reduce emissions of greenhouse gases associated with electricity generation, or both enhancement and reduction. The standards shall become effective no sooner than one year after the date of their adoption or updating.
(2)In adopting the flexible demand appliance standards, the commission shall consider the National Institute of Standards and Technologys reliability and cybersecurity protocols, or other cybersecurity protocols that are equally or more protective, and shall adopt, at a minimum, the North American Electric Reliability Corporations Critical Infrastructure Protection standards.
(3)The flexible demand appliance standards shall be cost effective. When determining cost-effectiveness, solely for purposes of this subdivision, the commission may consider, as appropriate, the cost of flexible demand appliances compared to nonflexible demand appliances, the value of increased or decreased emissions of greenhouse gases associated with the timing of an appliances use, the life-cycle cost to the consumer from using a product that complies with the standard, and the life-cycle costs and benefits to consumers, including the ability to conserve energy and better align consumer and electric system demand. The commission shall consider other relevant factors, as required by Sections 11346.5 and 11357 of the Government Code, including, but not limited to, the impact on housing costs, the total statewide costs and benefits of the standard over its lifetime, the economic impact on California businesses, and alternative approaches and their associated costs.
(4)The commission shall consult with the Public Utilities Commission, load-serving entities, and local publicly owned electric utilities to better align the flexible demand appliance standards with demand response programs administered by the state and load-serving entities and local publicly owned electric utilities and to incentivize the deployment of flexible demand appliances.
(5)The flexible demand appliance standards shall prioritize all of the following:
(A)Appliances that can more conveniently have their electrical demand controlled by load-management technology and third-party load-management programs.
(B)Appliances with load-management technology options that are readily available.
(C)Appliances that have a user-friendly interface and follow a straightforward setup and connection process, such as remote setup by means of an internet website or application.
(D)Appliances with load-management technology options that follow simple standards for third-party direct operation of the appliances.
(E)Appliances that are interoperable or open source.
(6)On or before January 1, 2021, and as necessary thereafter, the commission shall include as part of each integrated energy policy report adopted pursuant to Chapter 4 (commencing with Section 25300) a description of any actions it has taken pursuant to this subdivision and the flexible demand appliance standards cost to consumers.
(7)For purposes of this subdivision, all of the following definitions apply:
(A)Flexible demand means the capability to schedule, shift, or curtail the electrical demand of a load-serving entitys customer or a local publicly owned electric utilitys customer through direct action by the customer or through action by a third party, the load-serving entity, or a grid balancing authority, with the customers consent.
(B)Load-serving entity has the same meaning as defined in Section 380 of the Public Utilities Code.
(C)Local publicly owned electric utility has the same meaning as defined in Section 224.3 of the Public Utilities Code.
SEC. 2.Section 451.10 is added to the Public Utilities Code, to read:451.10.SEC. 3.Section 451.4 is added to the Public Utilities Code, to read:451.4.(a)For purposes of this section, data center has the same meaning as defined in Section 25345 of the Public Resources Code.
(b)In an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center, the commission shall determine whether those costs and expenses are just and reasonable in accordance with Section 451.
(c)In determining whether costs and expenses are just and reasonable pursuant to subdivision (b), the commission shall minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.
For purposes of this chapter, the following definitions apply:
(a)Artificial intelligence or AI means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.
(b)Covered model means an AI model developed using a quantity of computing power exceeding 10^25 integer or floating-point operations.
(c)Covered system means an AI system that incorporates a covered model.
(d)(1)Data center means a facility that houses computing infrastructure, including graphics and central processing units, servers, storage devices, networking equipment, and associated power and cooling systems, for the primary purpose of processing, storing, or distributing electronic data.
(2)Data center does not include a facility with an annual energy consumption of 10 megawatthours or less.
(e)Developer means a person, partnership, state or local government agency, or corporation that designs, codes, trains, or otherwise produces a covered model.
25345.1.A data center that provides computing resources to a developer for the purpose of developing a covered model shall do both of the following:
(a)Estimate the total energy, in megawatthours, used to develop the covered model.
(b)Report the estimated total energy used to develop the covered model to the developer under both of the following circumstances:
(1)Upon the termination of a contract facilitating the transfer of computing resources from the data center to the developer.
(2)Upon receiving a request from the developer pursuant to paragraph (1) of subdivision (b) of Section 25345.2.
25345.2.(a)Before using the computing resources of a data center to develop a covered model, a developer shall inform the data center of the developers intent to develop a covered model.
(b)Before using a covered model or covered system commercially, or before making a covered model or covered system available for use by a third party, a developer shall do both of the following:
(1)Request the developers energy usage data from all data centers involved in developing the covered model.
(2)Estimate and publish on the developers internet website the total energy, in megawatthours, used to develop the covered model.
(c)Energy data published pursuant to subdivision (b) shall be all of the following:
(1)Publicly available for as long as the developer continues to use the covered model or covered system, or continues to make the covered model or covered system available for use by a third party.
(2)Provided at no cost to a user of the developers internet website.
(3)Accessible and comprehensible to a natural person with ordinary cognitive abilities.
25345.3.(a)(1)The commission shall require operators of data centers to annually report energy consumption and performance data to the commission.
(2)Reports submitted to the commission pursuant to this subdivision shall include all of the following:
(A)Total annual energy consumption, including electricity usage from grid and nongrid sources, peak factors, and load profiles.
(B)Energy usage effectiveness or equivalent energy efficiency metrics.
(C)The percentage of energy supplied by renewable or carbon-free resources.
(D)Annual energy consumption resulting from the development of covered models.
(E)Any other information deemed necessary by the commission to assess grid impacts.
(3)Data centers reporting pursuant to this section shall submit data in a form and manner prescribed by the commission.
(b)The commission shall include energy consumption trends for data centers in its integrated energy policy report prepared pursuant to Section 25302.
(c)On or before January 1, 2027, the commission shall adopt regulations to implement this section.
25345.4.(a)The commission shall adopt energy efficiency performance standards for data centers.
(b)The standards adopted pursuant to this section shall do all of the following:
(1)Prioritize cost effectiveness, technological feasibility, and alignment with Californias greenhouse gas emission reduction targets.
(2)Consider best practices for energy usage effectiveness, energy management systems, and the use of renewable energy resources.
(3)Require new data centers and substantial alterations to existing data centers to incorporate load-management and load-shifting capabilities, including the ability to participate in demand response programs.
(c)(1)On or before January 1, 2027, the commission shall adopt regulations to implement this section.
(2)In adopting regulations pursuant to this subdivision, the commission shall consult with data center operators, utilities, technical experts, and other industry stakeholders.
SEC. 2.Section 451.10 is added to the Public Utilities Code, to read:451.10.(a)For purposes of this section, data center has the same meaning as defined in Section 25345 of the Public Resources Code.
(b)In an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center, the commission shall determine whether those costs and expenses are just and reasonable in accordance with Section 451.
(c)In determining whether costs and expenses are just and reasonable pursuant to subdivision (b), the commission shall minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.
|
Date
|
Status
|
|---|---|
| Aug 29, 2025 | In committee: Held under submission. |
| Aug 18, 2025 | In committee: Referred to suspense file. |
| Jul 16, 2025 | From committee: Do pass and re-refer to Com. on APPR. (Ayes 11. Noes 2.) (July 15). Re-referred to Com. on APPR. |
| Jul 7, 2025 | Read second time and amended. Re-referred to Com. on JUD. |
| Jul 3, 2025 | From committee: Amend, and do pass as amended and re-refer to Com. on JUD. (Ayes 13. Noes 3.) (July 1). |
| Jun 11, 2025 | Referred to Coms. on E., U & C. and JUD. |
| Jun 3, 2025 | In Senate. Read first time. To Com. on RLS. for assignment. |
| Jun 2, 2025 | Read third time. Passed. Ordered to the Senate. (Ayes 57. Noes 18. Page 1897.) |
| May 27, 2025 | Read second time. Ordered to third reading. |
| May 23, 2025 | Read second time and amended. Ordered returned to second reading. |
| May 23, 2025 | From committee: Amend, and do pass as amended. (Ayes 11. Noes 3.) (May 23). |
| May 23, 2025 | Assembly Rule 63 suspended. (Ayes 51. Noes 16. Page 1644.) |
| May 14, 2025 | In committee: Set, first hearing. Referred to suspense file. |
| May 6, 2025 | Re-referred to Com. on APPR. |
| May 5, 2025 | Read second time and amended. |
| May 1, 2025 | From committee: Amend, and do pass as amended and re-refer to Com. on APPR. (Ayes 10. Noes 1.) (May 1). |
| Apr 8, 2025 | Re-referred to Com. on P. & C.P. |
| Apr 7, 2025 | Read second time and amended. |
| Apr 3, 2025 | From committee: Amend, and do pass as amended and re-refer to Com. on P. & C.P. (Ayes 13. Noes 5.) (April 2). |
| Mar 26, 2025 | In committee: Set, first hearing. Hearing canceled at the request of author. |
| Feb 18, 2025 | Referred to Coms. on U. & E. and P. & C.P. |
| Jan 9, 2025 | From printer. May be heard in committee February 8. |
| Jan 8, 2025 | Read first time. To print. |
|
Date
|
Description
|
Location
|
Votes
|
|
|---|---|---|---|---|
| Aug 18, 2025 | Placed on suspense file | Senate Standing Committee on Appropriations |
Passed |
|
| Jul 15, 2025 | Do pass, but first be re-referred to the Committee on [Appropriations] | Senate Standing Committee on Judiciary |
Passed |
|
| Jul 1, 2025 | Do pass as amended, but first amend, and re-refer to the Committee on [Judiciary] | Senate Standing Committee on Energy, Utilities and Communications |
Passed |
|
| Jun 2, 2025 | AB 222 Bauer-Kahan Assembly Third Reading | Assembly Floor |
Passed |
|
| May 23, 2025 | Do pass as amended. | Assembly Standing Committee on Appropriations |
Passed |
|
| May 1, 2025 | Do pass as amended and be re-referred to the Committee on [Appropriations] | Assembly Standing Committee on Privacy and Consumer Protection |
Passed |
|
| Apr 2, 2025 | Do pass as amended and be re-referred to the Committee on [Privacy and Consumer Protection] | Assembly Standing Committee on Utilities and Energy |
Passed |
This list shows positions taken on this bill by organizations on the date displayed, which is linked to the original source.
Positions by individuals are not included.
|
Organization
|
Position
|
Date
|
|---|---|---|
| 350 Bay Area Action | ** Support | |
| 350 Conejo/San Fernando Valley | ** Support | |
| 350 Humboldt: Grass Roots Climate Action | ** Support | |
| 350 South Bay LA | ** Support | |
| 350 Southland Legislative Alliance | ** Support | |
| 350 Ventura County Climate Hub | ** Support | |
| An individual | ** Support | |
| BanSup | ** Support | |
| Bay Area Council | ** Against | |
| CA Blockchain Advocacy Coalition | ** Against | |
| CalAsian Chamber of Commerce | ** Against | |
| California African American Chamber Of Commerce | ** Against | |
| California Chamber Of Commerce | ** Against | |
| California Environmental Voters | ** Support | |
| California Hispanic Chambers Of Commerce | ** Against | |
| California Pacific Asian Chamber of Commerce | ** Against | |
| California River Watch | ** Support | |
| California State Association Of Electrical Workers | ** Against | |
| Central Valley Business Federation | ** Against | |
| Chamber San Mateo County | ** Against | |
| City of Santa Barbara | ** Support | |
| cleanearth4kids.org | ** Support | |
| Climate Action California | ** Support | |
| Coalition of California Utility Employees | ** Against | |
| Compute Exchange INC. | ** Support | |
| Data Center Coalition | ** Against | |
| Indivisible Alta Pasadena | ** Support | |
| Indivisible Green Team | ** Support | |
| Information Technology Industry Council | ** Against | |
| League Of California Cities | ** Support | |
| Los Angeles Business Federation | ** Against | |
| Los Angeles County Business Federation | ** Against | |
| Orange County Business Council | ** Against | |
| San Diego Regional Chamber Of Commerce | ** Against | |
| San Jose Chamber of Commerce | ** Against | |
| San Mateo County Economic Development Association | ** Against | |
| Santa Barbara South Coast Chamber of Commerce | ** Against | |
| Santa Cruz Climate Action Network | ** Support | |
| Sierra Club | ** Support | |
| Silicon Valley Leadership Group | ** Against | |
| Solano County Democratic Central Committee | ** Support | |
| Sustainable Rossmoor | ** Support | |
| Techca | ** Against | |
| Technet | ** Against | |
| The Center for Ai and Digital Policy | ** Support | |
| Transparency Coalition.ai | ** Support | |
| Valley Industry & Commerce Association | ** Against | |
| Vica | ** Against |
SENATE COMMITTEE ON
Senator Anna Caballero, Chair
2025 - 2026 Regular Session
AB 222 (Bauer-Kahan) - Data centers: power usage effectiveness: cost shifts
Version: | Policy Vote: |
Urgency: | Mandate: |
Hearing Date: | Consultant: |
Bill Summary:
This bill would require the California Public Utilities Commission (CPUC) to establish a process for the owner of a data center, as defined, to submit the power usage effectiveness ratio, as defined, for the data center to the CPUC, as provided. The bill would also require the CPUC to assess the extent to which electrical corporation costs associated with new loads from data centers result in cost shifts to other electrical corporation customers, as specified.
Fiscal Impact:
- The CPUC estimates ongoing costs of about $938,000 annually (ratepayer funds) until the sunset date of January 1, 2031 to assess and submit a legislative report by January 1, 2027, on the extent to which electrical corporation costs associated with new loads from data centers could result in cost shifts to other customers.
- The California Energy Commission (CEC) estimates ongoing costs of about $176,000 annually (Energy Resources Programs Account [ERPA] or other fund) until the sunset date of January 1, 2031 to handle the data management, data collection, and support data analyses necessary.
Background:
This bill would limit the access to public records by providing that data center energy consumption can be reported in the aggregate, but not in a manner that would result in the disclosure of personally identifiable information or energy consumption data for a specific utility customer.
Access to information concerning the conduct of the people’s business is a fundamental and necessary right of every person in this state. (Gov. Code § 7921.000.) In 2004, the right of public access was enshrined in the California Constitution with the passage of Proposition 59 (Nov. 3, 2004, statewide general election), which amended the California Constitution to specifically protect the right of the public to access and obtain government records: “The people have the right of access to information concerning the conduct of the people’s business, and therefore . . . the writings of public officials and agencies shall be open to public scrutiny.” (Cal. Const., art. I, sec. 3 (b)(1).) In 2014, voters approved Proposition 42 (Jun. 3, 2014, statewide direct primary election) to further increase public access to government records by requiring local agencies to comply with the CPRA and the Ralph M. Brown Act , and with any subsequent statutory enactment amending either act, as provided. (Cal. Const., art. I, sec. 3 (b)(7).)
Under the CPRA, public records are open to inspection by the public at all times during the office hours of the agency, unless they are exempt from disclosure. (Gov. Code § 7922.525.) A public record is defined as any writing containing information relating to the conduct of the public’s business that is prepared, owned, used, or retained by any public agency regardless of physical form or characteristics. (Gov. Code § 7920.530.) There are several general categories of documents or information that are permissively exempt from disclosure under the CPRA essentially due to the character of the information. The exempt information can be withheld by the public agency with custody of the information, but it also may be disclosed if it is shown that the public’s interest in disclosure outweighs the public’s interest in non-disclosure of the information. (CBS, Inc. v. Block (1986) 42 Cal.3d 646, at 652.). Additionally, some records are prohibited from disclosure or are specifically stated to not be public records. (see Gov. Code § 7924.110(a).)
California generally recognizes that public access to information concerning the conduct of the people’s business is a fundamental and necessary right. At the same time, the state recognizes that this right must be balanced against the right to privacy. The general right of access to public records may, therefore, be limited when records include personal information. The bill states that this limitation on access to public records is needed to protect the confidential and proprietary information of an entity subject to the bill.
Proposed Law:
This bill would:
- Require the CPUC to include, as a part of the 2027 edition of the IEPR, an assessment of electrical load trends for data centers. The assessment shall include all of the following:
- A projection of future load trends from data centers.
- Identification of potential net peak load demands.
- Recommendations for mitigating data center electricity consumption impacts on the electrical grid, including any recommended energy efficiency and demand response measures.
- Authorize the CPUC to report data center energy consumption in an aggregate basis, but prohibits disclosure of data center energy consumption information in a manner that would result in the disclosure of personally identifiable information or energy consumption data for a specific utility customer.
- Require the CPUC to establish a process for the owner of a data center to submit the power usage effectiveness ratio for the data center to the commission on a biannual basis. The owner of a data center shall submit the power usage effectiveness ratio for the data center in a manner and timeframe specified by the CPUC.
- Define, for purposes of 1) through 3), above, the following definitions:
- “Data center” means a room, or a portion of a room, in a building used primarily to house information technology equipment that serves a total information technology equipment load greater than 10 kilowatts and 20 watts per square foot of conditioned floor area.
- “Power usage effectiveness” means a ratio of the total energy consumption of a data center to the energy specifically used by the information technology equipment housed in that data center.
- Require the CPUC to assess the extent to which electrical corporation costs associated with new loads from data centers result in cost shifts to other electrical corporation customers. This assessment shall include all of the following:
- An analysis of potential electrical corporation costs associated with procurement to meet growing load demands from data centers’ increased energy consumption.
- An analysis of potential electrical corporation costs associated with the installation of new transmission and distribution assets to serve new data centers or expansions of existing data centers.
- To the extent that the commission finds that electrical corporation costs to serve new loads from data centers will result in substantial cost shifts to other electrical corporation customers, the commission’s assessment shall also identify opportunities to prevent or mitigate these costs.
- Require the CPUC, on or before January 1, 2027, to submit the assessment completed pursuant to this section to the relevant policy committees of the Legislature and publicly post a copy of the assessment on the CPUC’s website.
- Repeal these provisions on January 1, 2031.
- Make Legislative findings regarding limiting access to public records.
Related Legislation:
SB 57 (Padilla, 2025) would establish a tariff for customers with at least 50 MW of load interconnecting with transmission, identifies utility costs included in this tariff, requires the tariff to ensure just and reasonable rates, and prohibits cost shifts to customers who are not participating in the tariff.
SB 58 (Padilla, 2025) would provide certain tax incentives for data center equipment if those data centers using the equipment meet certain job creation, economic investment, and renewable energy requirements.
Prior Legislation: SB 1298 (Cortese, 2024, would have increased the amount of thermal generation a data center could use as backup power from 100 MW to 150 MW without triggering the CEC’s power plant siting process. The bill would have also created conditions for data centers to use this exemption. The bill died in the Assembly.
-- END --
Senator Thomas Umberg, Chair
2025-2026 Regular Session
Bill No: AB 222 (Author:Bauer-Kahan)
Version: July 7, 2025
Hearing Date: July 15, 2025
Fiscal: Yes
Urgency: No
AMConsultant:
SUBJECT
Data centers: power usage effectiveness: cost shifts
DIGEST
This bill requires the California Public Utilities Commission (CPUC) to establish a process for the owner of a data center, as defined, to submit the power usage effectiveness ratio, as defined, for the data center to the CPUC, as provided. The bill also requires the CPUC to assess the extent to which electrical corporation costs associated with new loads from data centers result in cost shifts to other electrical corporation customers, as specified.
EXECUTIVE SUMMARY
The author notes that the rapid development of the AI industry[1] is leading to increased construction of data centers in California,[2] which require large amounts of electricity. Additionally, the author points out that the construction of energy-intensive data centers often requires an expansion of grid infrastructure to accommodate increased energy demand. When this occurs, utility companies can seek approval from the CPUC to pass on the costs of electricity production and infrastructure upgrades to ratepayers. As noted in the Senate Committee on Energy, Utilities and Communications analysis of this bill “[f]ive of the anticipated facilities serving OpenAI could collectively use more electricity than three million households.”[3] This bill is intended to address this issue by requiring the CPUC to assess the extent to which electrical corporation costs associated with new loads from data centers result in cost shifts to other electrical corporation customers. Additionally, the bill requires the CPUC to establish a process for the owner of a data center to submit the power usage effectiveness ratio for the data center to the CPUC, and requires the CPUC to include an assessment of electrical load trends for data centers as part of the 2027 edition of the integrated energy policy report. The only provision of this bill in this Committee’s jurisdiction is the limitation on access to public records, and as such the analysis will only focus on this piece.
The bill is author sponsored. It is supported by various environmental organizations and the League of California Cities. It is opposed by various business organizations and advocates for the technology industry. The bill passed the Senate Committee on Energy, Utilities and Communications on a vote of 13 to 3.
PROPOSED CHANGES TO THE LAW
Existing law:
- Authorizes the CPUC to supervise and regulate every public utility in the state and permits the CPUC to do anything that is necessary and convenient to exercise its power and jurisdiction. (Pub. Util. Code § 701.)
- Authorizes the CPUC to set rates for public utilities and specifies that every cost charged by utilities to customers must be just and reasonable. (Pub. Util. Code § 451.)
- Defines an electrical corporation as every corporation or person owning, controlling, operating, or managing any electric plant for compensation within this state, except where electricity is generated on or distributed by the producer through private property solely for its own use or the use of its tenants and not for sale or transmission to others. (Pub. Util. Code § 218.)
- Defines “electrical or gas consumption data” as a customer’s electrical or natural gas usage that is made available as part of an advanced metering infrastructure, and includes incremental and monthly meter-specific electricity data, to the extent produced by that infrastructure, and the name, account number, and address of the customer. Existing law prohibits electric and gas utilities from disclosing customers’ electrical or gas consumption data without the customer’s consent unless state law, federal law, or the CPUC orders the disclosure of the data. Utilities may also share customer consumption data with certain third parties for operational needs or implementation of certain utility programs if specific data protection requirements are met. (Pub. Util. Code § 8380.)
- Requires the California Energy Commission (CEC) to adopt an integrated energy policy report (IEPR) every two years, with an update published every year. Existing law specifies the contents of the IEPR and requires the CEC to report on major energy trends in the IEPR, including assessments of statewide electricity, natural gas and transportation fuel demands. (Pub. Res. Code § 25302.)
- Provides, pursuant to the California Constitution, that the people have the right of access to information concerning the conduct of the people’s business, and, therefore, the meetings of public bodies and the writings of public officials and agencies are required to be open to public scrutiny. (Cal. const. art. I, § 3(b)(1).)
- Requires a statute to be broadly construed if it furthers the people’s right of access, and narrowly construed if it limits the right of access. (Cal. const. art. I, § 3(b)(1).)
- Requires a statute that limits the public’s right of access to be adopted with findings demonstrating the interest protected by the limitation and the need for protecting that interest. (Cal. const. art. I, § 3(b)(1).)
- Governs the disclosure of information collected and maintained by public agencies pursuant to the California Public Records Act (CPRA). (Gov. Code §§ 7920.000 et seq.)
- States that the Legislature, mindful of the individual right to privacy, finds and declares that access to information concerning the conduct of the people’s business is a fundamental and necessary right of every person in this state. (Gov. Code § 7921.000.)
- Defines “public records” as any writing containing information relating to the conduct of the public’s business prepared, owned, used, or retained by any state or local agency regardless of physical form or characteristics. (Gov. Code § 7920.530.)
- Provides that all public records are accessible to the public upon request, unless the record requested is exempt from public disclosure. (Gov. Code § 7922.530.)
This bill:
- Requires the CPUC to include, as a part of the 2027 edition of the IEPR, an assessment of electrical load trends for data centers. The assessment shall include all of the following:
- A projection of future load trends from data centers.
- Identification of potential net peak load demands.
- Recommendations for mitigating data center electricity consumption impacts on the electrical grid, including any recommended energy efficiency and demand response measures.
- Authorizes the CPUC to report data center energy consumption in an aggregate basis, but prohibits disclosure of data center energy consumption information in a manner that would result in the disclosure of personally identifiable information or energy consumption data for a specific utility customer.
- Requires the CPUC to establish a process for the owner of a data center to submit the power usage effectiveness ratio for the data center to the commission on a biannual basis. The owner of a data center shall submit the power usage effectiveness ratio for the data center in a manner and timeframe specified by the CPUC.
- Defines, for purposes of 1) through 3), above, the following definitions:
- “Data center” means a room, or a portion of a room, in a building used primarily to house information technology equipment that serves a total information technology equipment load greater than 10 kilowatts and 20 watts per square foot of conditioned floor area.
- “Power usage effectiveness” means a ratio of the total energy consumption of a data center to the energy specifically used by the information technology equipment housed in that data center.
- Requires the CPUC to assess the extent to which electrical corporation costs associated with new loads from data centers result in cost shifts to other electrical corporation customers. This assessment shall include all of the following:
- An analysis of potential electrical corporation costs associated with procurement to meet growing load demands from data centers’ increased energy consumption.
- An analysis of potential electrical corporation costs associated with the installation of new transmission and distribution assets to serve new data centers or expansions of existing data centers.
- To the extent that the commission finds that electrical corporation costs to serve new loads from data centers will result in substantial cost shifts to other electrical corporation customers, the commission’s assessment shall also identify opportunities to prevent or mitigate these costs.
- Requires the CPUC, on or before January 1, 2027, to submit the assessment completed pursuant to this section to the relevant policy committees of the Legislature and publicly post a copy of the assessment on the CPUC’s website.
- Repeals these provisions on January 1, 2031.
- Makes Legislative findings regarding limiting access to public records.
COMMENTS
- Stated need for the bill
The author writes:
Across California, energy-intensive data centers are being built to support the rapid expansion of the artificial intelligence (AI) industry. These data centers increase energy demand and frequently require expansions to the electrical grid; together, these factors threaten to increase energy costs for Californians. AB 222 increases transparency around data center energy use, and requires the Public Utilities Commission to assess cost shifting due to data center development. California’s energy costs are already among the highest in the country, and ratepayers should not be forced to bear the additional costs of AI development.
- Access to public records is a statutory and constitutional right
This bill limits the access to public records by providing that data center energy consumption can be reported in the aggregate, but not in a manner that would result in the disclosure of personally identifiable information or energy consumption data for a specific utility customer.
Access to information concerning the conduct of the people’s business is a fundamental and necessary right of every person in this state. (Gov. Code § 7921.000.) In 2004, the right of public access was enshrined in the California Constitution with the passage of Proposition 59 (Nov. 3, 2004, statewide general election),[4] which amended the California Constitution to specifically protect the right of the public to access and obtain government records: “The people have the right of access to information concerning the conduct of the people’s business, and therefore . . . the writings of public officials and agencies shall be open to public scrutiny.” (Cal. Const., art. I, sec. 3 (b)(1).) In 2014, voters approved Proposition 42 (Jun. 3, 2014, statewide direct primary election)[5] to further increase public access to government records by requiring local agencies to comply with the CPRA and the Ralph M. Brown Act[6], and with any subsequent statutory enactment amending either act, as provided. (Cal. Const., art. I, sec. 3 (b)(7).)
Under the CPRA, public records are open to inspection by the public at all times during the office hours of the agency, unless they are exempt from disclosure. (Gov. Code § 7922.525.) A public record is defined as any writing containing information relating to the conduct of the public’s business that is prepared, owned, used, or retained by any public agency regardless of physical form or characteristics. (Gov. Code § 7920.530.) There are several general categories of documents or information that are permissively exempt from disclosure under the CPRA essentially due to the character of the information. The exempt information can be withheld by the public agency with custody of the information, but it also may be disclosed if it is shown that the public’s interest in disclosure outweighs the public’s interest in non-disclosure of the information. (CBS, Inc. v. Block (1986) 42 Cal.3d 646, at 652.). Additionally, some records are prohibited from disclosure or are specifically stated to not be public records. (see Gov. Code § 7924.110(a).)
California generally recognizes that public access to information concerning the conduct of the people’s business is a fundamental and necessary right.[7] At the same time, the state recognizes that this right must be balanced against the right to privacy.[8] The general right of access to public records may, therefore, be limited when records include personal information. The bill states that this limitation on access to public records is needed to protect the confidential and proprietary information of an entity subject to the bill.
- Statements in support
The League of California Cities writes in support, stating:
[…] AB 222 would require developers of AI models to estimate the total energy used to develop the model and the percentage of the total energy used to develop the model that was generated in California. The bill would require this information to be posted on their website, so it is publicly available. The bill would also require a data center to report energy consumption trends, its physical address, website, and data center point of contact to the Commission, so the Commission can include relevant information in their integrated energy policy report.
Data centers are an important type of development to attract and retain in California. Some cities may be involved with the siting of new or expanded data centers in their communities, where other cities may allow data centers to occupy vacant buildings where these data centers can run their server systems. While these are critically important technology investments for California to maintain for economic purposes, Cal Cities supports the goal of AB 222 which would provide transparency measures on energy consumption. AB 222 would provide for greater understanding of the estimated energy consumption of data centers to harden the grid to support their demand and prevent grid outages or load management impacts. […]
- Statements in opposition
A coalition of various business organizations, including the California Chamber of Commerce and TechNet, write in opposition, stating:
[…] The recent amendments accepted by the author from the Senate Committee on Energy, Utilities and Communications moves the bill in a positive direction and we appreciate the ongoing engagement with the office. The committee amendments are helpful in acknowledging how reporting certain information raises privacy and security concerns, and the protections outlined in the amendments are helpful.
However, it is unclear why data centers would be the only industry subject to these additional reporting requirements when they are but one large end user of electric utilities and part of a larger portfolio driving increased demand. It is also unclear how existing reporting and benchmarking requirements through the Energy Commission don't already accomplish the purpose of understanding usage and allowing the Commission to evaluate usage trends. Additionally, utility providers already collect forecasting data from large end users as part of their grid planning exercises, and that extends beyond data centers. While Power Usage Effectiveness (PUE) is a metric that raises fewer concerns surrounding trade secrets, security, and privacy, there are also limitations. As noted in a December 2024 report by Virginia’s Joint Legislative Audit & Review Commission, PUE “does not indicate a data center’s overall energy efficiency; it measures only the efficiency of cooling and other building systems that support facility operations” and “[r]equiring a specific and narrow requirement, like meeting a specific PUE ratio, could have unintended consequences.”1 By isolating only one end user of energy AB 222 would fail to provide a full picture of broader energy needs in the state, including other drivers with significant energy demands, and therefore, it is unclear what benefit such reporting would provide to California. As detailed further in the following comments, there is a diverse set of purposes across multiple industries driving increased electricity demand. Addressing the important issues of load growth and ensured responsible, effective grid planning would be more effectively addressed through active partnerships with utilities, grid operators, and regulators and communication, collaboration, and transparency among all stakeholders to ensure grid investments are data-driven, appropriately scaled, and protect all customers from unnecessary costs. […]
SUPPORT
350 Bay Area Action
350 Conejo/San Fernando Valley
350 Humboldt: Grass Roots Climate Action
350 South Bay LA
350 Southland Legislative Alliance
350 Ventura County Climate Hub
BanSup
California River Watch
City of Santa Barbara
Cleanearth4kids.org
Indivisible Alta Pasadena
Indivisible Green Team
League of California Cities
Santa Cruz Climate Action Network
Sierra Club
Solano County Democratic Central Committee
Sustainable Rossmoor
OPPOSITION
Bay Area Council
CA Blockchain Advocacy Coalition
CalAsian Chamber of Commerce
California African American Chamber of Commerce
California Chamber of Commerce
California Hispanic Chambers of Commerce
Central Valley Business Federation
Data Center Coalition
Los Angeles Business Federation
Orange County Business Council
San Diego Regional Chamber of Commerce
San Jose Chamber of Commerce
San Mateo County Economic Development Association
Santa Barbara South Coast Chamber of Commerce
Silicon Valley Leadership Group
TechCA
TechNet
Valley Industry & Commerce Association
RELATED LEGISLATION
Pending Legislation:
SB 57 (Padilla, 2025) would establish a tariff for customers with at least 50 MW of load interconnecting with transmission, identifies utility costs included in this tariff, requires the tariff to ensure just and reasonable rates, and prohibits cost shifts to customers who are not participating in the tariff. SB 57 is currently pending before the Assembly.
SB 58 (Padilla, 2025) would provide certain tax incentives for data center equipment if those data centers using the equipment meet certain job creation, economic investment, and renewable energy requirements. SB 58 is currently in the Senate Revenue and Taxation Committee.
Prior Legislation: SB 1298 (Cortese, 2024, would have increased the amount of thermal generation a data center could use as backup power from 100 MW to 150 MW without triggering the CEC’s power plant siting process. The bill would have also created conditions for data centers to use this exemption. The bill died in the Assembly.
PRIOR VOTES
Senate Energy, Utilities and Communications Committee (Ayes 13, Noes 3)
Assembly Floor (Ayes 57, Noes 18)
Assembly Appropriations Committee (Ayes 11, Noes 3)
Assembly Privacy and Consumer Protection Committee (Ayes 10, Noes 1)
Assembly Utilities and Energy Committee (Ayes 13, Noes 5)
**************
Taiba Jafari,et al, Projecting the Electricity Demand Growth of Generative AI Large Language Models in the US, Center on Global Energy Policy, (Jul. 17, 2024), available at https://www.energypolicy.columbia.edu/projecting-the-electricity-demand-growth-of-generative-ai-large-language-models-in-the-us/. ↑
Dan Swinhoe, PG&E: 3.5GW of data center capacity in California's connection pipeline over next five years, Data Center Dynamics, (Jun. 24, 2024), available at https://www.datacenterdynamics.com/en/news/pge-35gw-of-data-center-capacity-in-connection-pipeline-over-next-five-years/. ↑
Sen. Comm. on Energy, Util. and Communications analysis of AB 222 (2025-26 reg. sess.) as amended May 23, 2025 at p. 3. ↑
Prop. 59 was placed on the ballot by a unanimous vote of both houses of the Legislature. (SCA 1 (Burton, Ch. 1, Stats. 2004).) ↑
Prop. 42 was placed on the ballot by a unanimous vote of both houses of the Legislature. (SCA 3 (Leno, Ch. 123, Stats. 2013).) ↑
The Ralph M. Brown Act is the open meetings laws that applies to local agencies. (Gov. Code §§ 59450 et. seq.) ↑
Cal. Const., art. I, § 3; Gov. Code, § 7921.000. ↑
Cal. Const., art. I, § 1. ↑
SENATE COMMITTEE ON ENERGY, UTILITIES AND COMMUNICATIONS
Senator Josh Becker, Chair
2025 - 2026 Regular
Bill No: Hearing Date: 7/1/2025
Author: | |||
Version: | Amended | ||
Urgency: | Fiscal: | ||
Consultant: | |||
SUBJECT:
DIGEST: This bill requires an artificial intelligence (AI) developer to publicly post the amount of energy used to create an AI model on the developer’s website. This bill requires the California Energy Commission (CEC) to include data center energy consumption trends in the Integrated Energy Policy Report (IEPR). This bill also requires the California Public Utilities Commission (CPUC) to determine if electrical corporation costs associated with the construction or expansion of a data center are just and reasonable, as specified.
ANALYSIS:
Existing law:
- Authorizes the CPUC to supervise and regulate every public utility in the state and permits the CPUC to do anything that is necessary and convenient to exercise its power and jurisdiction. (Public Utilities Code §701)
- Authorizes the CPUC to set rates for public utilities and specifies that every cost charged by utilities to customers must be just and reasonable. (Public Utilities Code §451)
- Defines an electrical corporation as every corporation or person owning, controlling, operating, or managing any electric plant for compensation within this state, except where electricity is generated on or distributed by the producer through private property solely for its own use or the use of its tenants and not for sale or transmission to others. (Public Utilities Code §218)
- Defines “electrical or gas consumption data” as a customer’s electrical or natural gas usage that is made available as part of an advanced metering infrastructure, and includes incremental and monthly meter-specific electricity data, to the extent produced by that infrastructure, and the name, account number, and address of the customer. Existing law prohibits electric and gas utilities from disclosing customers’ electrical or gas consumption data without the customer’s consent unless state law, federal law, or the CPUC orders the disclosure of the data. Utilities may also share customer consumption data with certain third parties for operational needs or implementation of certain utility programs if specific data protection requirements are met. (Public Utilities Code §8380)
- Requires the CEC to adopt an IEPR every two years, with an update published every year. Existing law specifies the contents of the IEPR and requires the CEC to report on major energy trends in the IEPR, including assessments of statewide electricity, natural gas and transportation fuel demands. (Public Resources Code §25302)
This bill:
- Establishes various definitions for its purpose, including, the following:
- “Artificial intelligence” or “AI” means an engineered or machine-based system that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.
- “Covered model” means an AI model developed using a quantity of computing power exceeding 10^25 integer or floating-point operations.
- “Data center” means a building with a primary purpose of housing one or more computer rooms with an energy demand of at least five megawatts (MW).
- “Developer” means a person, partnership, state or local government agency, or corporation that designs, codes, trains, or otherwise produces a covered model.
- Requires an AI developer to do the following before commercially using a covered AI model:
- Estimate the energy used to develop the covered model, including estimating the percentage of energy provided by in-state generation.
- Publish the estimated energy usage data on the developer’s internet website.
- Requires an AI developer to do the following annually, starting February 1, 2027:
- Estimate the energy used to operate the covered model during the prior year, including estimating the percentage of energy provided by in-state generation.
- Publish the estimated energy usage data on the developer’s internet website.
- Requires the AI developer to provide its energy use data as required by this bill for free for as long as the developer uses the AI model or offers the AI model for use, and it requires the data to be formatted in an accessible manner.
- Requires a data center to provide the following formation to the CEC before commencing operations and at any point when this information changes:
- The data center’s name.
- The data center’s physical addresses.
- The data center’s internet website.
- A telephone number and email address for a point of contact for the CEC to communicate with the data center.
- Requires the CEC to include data center energy consumption trends in the IEPR.
- Requires the CPUC to determine if utility costs associated with data centers are just and reasonable.
- Requires the CPUC to minimize ratepayer cost shifts associated with the construction or alteration of a data center.
Background
Conspicuous Consumption: Growth of AI will accelerate data centers’ energy demands. Data centers are facilities that house computer systems, including servers that help store, process, and route data used for computing and telecommunications. A May 2024 report from the CEC on data centers’ power systems indicates that data centers comprise approximately two percent of the state’s total annual electricity demand. The CEC anticipates that data centers’ electricity consumption may double in the next 10 years without implementation of any additional energy efficiency measures. In December 2024, the United States Department of Energy released a report by the Lawrence Berkeley National Laboratory on data centers’ energy use. The report shows that data centers’ electricity use comprised approximately 4.4% of the United States’ total electricity consumption in 2024. The report also estimates that this load may grow to reach between 6.7% to 12% of national electricity consumption by 2028. The growth of data centers’ electricity consumption has been partly driven by the growth of AI. Data centers that serve AI customers are particularly large electricity consumers within the data center sector. Five of the anticipated facilities serving OpenAI could collectively use more electricity than three million households.
Bill focuses on disclosure regarding software development, but these disclosures may not provide an accurate picture for energy modeling. While some data centers wholly serve a single entity, many data centers provide services to a variety of clients, including software developers, telecommunications providers, video streaming services, and other web-based applications. This bill requires AI developers to publicly post specified information about energy consumption associated with developing an AI model on the developer’s website. However, an AI developer and the data center used by the AI developer are not necessarily the same entity or owned by the same entity. As a result, a data center facilitating AI activities may be a separate utility customer from the AI developer.
Existing law (Public Utilities Code §8380) prohibits electric and gas utilities from sharing customers’ consumption data unless the customer provides the utility with consent to share that information. Historically, customers’ metered energy consumption data and associated personal information has been protected by customer privacy restrictions placed on utilities. These privacy protections are intended to prevent third parties from using customer energy consumption to infer additional personal information or proprietary and confidential business information. Disclosure of certain energy consumption and trends in energy use for a specific customer can enable third parties to make certain assumptions about the customer’s behavior, technology use, and personal habits without the customer’s consent. To the extent that an AI developer and a data center running AI processes for that developer are separate businesses, it is not clear that an AI developer can obtain the historical utility data necessary to accurately estimate or report the energy consumption associated with the development of an AI model. Even in circumstances where a data center is willing to provide its energy consumption data, the data center may not be able to fully disaggregate the energy it consumes for the development of single company’s AI model from the data center’s total load. To the extent that AI substantially impacts overall electric load growth in California, the load growth will likely come from data center energy demands where those facilities are providing services for generative AI.
Opportunities to assess data centers’ energy consumption and identify strategies to improve energy efficiency. While much of this bill focuses on energy consumption associated with specific AI models, data center owners are already monitoring their facilities’ efficiency and seeking ways to increase data centers’ capacity to run complex tasks while limiting energy costs. The efficiency of a data center may be expressed in a number of ways; however, Power Usage Effectiveness (PUE) and Data Center Infrastructure Efficiency (DCIE) are two widely used measures to express a data center’s relative efficiency. Below are descriptions of what these measures say about a data center’s performance:
- PUE: this measure is a ratio of the total amount of power used by a data center relative to the power delivered to the center’s computing equipment. A lower PUE rating indicates that a facility is more efficient. A PUE rating of 1 indicates that all power consumed by the data center is used by the computing equipment. The National Renewable Energy Laboratory (NREL) set a goal of achieving a PUE of 1.06 or better for the data center serving its Energy Systems Integration Facility. NREL states that the data center has achieved that goal every year and has an annualized PUE rating of 1.036.
- DCIE: this metric helps identify the energy consumed by a data center’s computing equipment. Unlike PUE, a higher DCIE indicates a more efficient data center because it shows that a higher proportion of the data center’s energy consumption is used for the computing units.
A variety of factors can influence a data center’s efficiency, including, but not limited to the data center’s location, size, facility type, equipment, and clients. As a result, a variety of tools exist to improve the energy efficiency in data centers. Lawrence Berkeley National Lab hosts the Center of Expertise for Energy Efficiency in Data Centers. The Center provides resources to analyze data centers’ energy consumption trends and identify tools to improve data centers’ energy performance. In its 2024 report sponsored by the Department of Energy, the Center noted that accurate projections of data center energy consumption and opportunities to mitigate potential impacts associated with that consumption would depend upon obtaining sufficient data to model real-world trends on a regular basis. Regarding the lack of sufficient data on data centers, the report says, “The lack of direct energy data available in a sector with rapidly evolving technologies limits the analysis in this report, especially when trying to understand and estimate future energy demand scenarios.” The report makes several recommendations, including the following:
The lack of primary performance and utilization data indicates that much greater transparency is needed around data centers. Very few companies report actual data center electricity use and virtually none report it in context of IT characteristics such as compute capacities, average system configurations, and workload types. These details are often considered proprietary, but novel data sharing arrangements could address these concerns by developing a repository for companies to provide energy use data that would be anonymized and aggregated for public release through coordination with entities that collect and anonymize data for other industries.
Cost-shifts directly attributable to data centers is unclear, but the CPUC is already addressing some cost-shift concerns. This bill is one of several pieces of legislation authored this year to address concerns about the potential rate and cost-shift impacts associated with meeting data centers’ energy needs. This bill requires the CPUC to determine whether an electrical corporation’s expenses associated with the construction of a data center are just and reasonable when the electrical corporation seeks to recover those costs through rates. This bill also requires the CPUC to take steps to minimize the extent to which utility costs for data centers are shifted to ratepayers who do not directly benefit from the data center. It is unclear how the CPUC or utilities would be able to assess which other ratepayers benefit from the data center. To the extent that data centers produce cost-shifts to other ratepayers, those cost shifts would manifest in two distinct ways: (1) costs associated with utility electrical generation procurements, including the potential construction of new generation resources, to serve data centers’ electrical demand growth and (2) increased costs for transmission and distribution facilities needed to interconnect and serve data centers. Since transmission rates are largely regulated by the Federal Energy Regulatory Commission, California is limited in its options for regulating how transmission costs are recovered from other ratepayers; however, the CPUC can take certain steps to mitigate overall transmission costs and address cost-shifts associated with distribution facilities and utility procurement.
Data centers are one type of commercial and/or industrial utility customer known for large energy consumption. Certain data center operations, including generative AI, are known for particularly large energy demands; however, not all data center operations consume the same amount of electricity. Those data centers that anticipate consuming particularly large loads are more likely to interconnect with utilities at the transmission-level, which provides electrical service at higher voltages. Increasingly, the investor-owned utilities are seeing a growth in data centers requesting transmission-level service. Between 2014 and 2022, Pacific Gas & Electric (PG&E) had a total of 16 retail customers interconnected with the transmission grid. Since 2023, PG&E has received 34 applications for transmission-level service from entities with an electrical demand of at least four MW. According to PG&E’s filings, data centers comprised 67% of the 34 transmission interconnection applications that PG&E has received since 2023. In the absence of an electric rule for these interconnections, PG&E used the “exceptional case” filing process at the CPUC, which is reserved for those circumstances when adhering to existing rules are not feasible, and a party requests a solution that is not authorized under existing CPUC rules and regulations. In November 2024, PG&E filed an application at the CPUC to establish a new Electric Rule 30 (Application 24-11-007). PG&E’s filing seeks to establish rules for interconnecting non-residential retail electric customers at transmission level voltages. On June 20, 2025, the CPUC issued a proposed decision allowing PG&E to implement interim measures for transmission-level customers who prepay up to 100% of specific transmission-level upgrades. The CPUC authorized this interim relief while it considers the other issues in PG&E’s application.
More information is likely needed to determine the extent to which projected load growth from the data center sector will impact utility rates in a manner that results in cost-shifts. Some stakeholders have asserted that scarcity in electrical generation resources, particularly for those resources necessary to meet reliability standards on days when the grid is strained, will necessarily mean that increased procurement to meet data center load demands will drive up rates. However, other stakeholders contend that meeting these demands will increase incentives for large energy procurements that drive down per-megawatt electricity costs.
Need for Amendments. As currently written, this bill attempts to address the need for more information about the energy impacts associated with the growth of AI by focusing data reporting on software developers creating more advanced AI models. This framework may not provide the best information to support load planning and identifying energy efficiency and load shaping opportunities. Additionally, this bill’s provisions regarding cost-shifts may be duplicative of ongoing attempts to address cost-shifts associated with transmission interconnected facilities, including data centers with large loads. As a result, the author and the committee may wish to amend this bill to do the following:
- Delete Section 1 of the bill.
- Recast requirements in Section 2 of the bill to require data center owners to report the Power Usage Effectiveness of data centers to the CEC.
- Require the CEC to assess potential grid impacts associated with new data center load and provide recommendations for mitigating these impacts.
- Specify that the CEC must include its findings regarding data center energy trends in the 2027 full IEPR.
- Clarify that the CEC may report data center energy consumption, consistent with Public Utilities Code 8380, on an aggregate basis if all information has been removed regarding the individual identity of a customer. However, the CEC shall not disclose data center energy consumption data on a single facility-basis or in a manner that would result in the disclosure of energy usage data for a specific utility customer.
- Recast provisions in Section 3 of the bill to require the CPUC to assess the extent to which data center growth will result in cost shifts to other utility customers for the following:
- Utility procurement to meet growing load demands from increased energy consumption associated with data centers.
- Increased costs for the installation of new transmission and distribution assets to serve new data centers or expansions of existing data centers.
- To the extent that the CPUC finds that energy demands from data centers will result in cost-shifts, the CPUC’s assessment shall also identify opportunities to prevent or mitigate these costs.
Dual Referral. Should this bill be approved by this committee, it will be re-referred to the Senate Judiciary Committee.
Prior/Related Legislation
SB 57 (Padilla) of 2025, establishes a tariff for customers with at least 50 MW of load interconnecting with transmission. The bill identifies utility costs included in this tariff, requires the tariff to ensure just and reasonable rates, and prohibits cost shifts to customers who are not participating in the tariff. The bill is currently pending before the Assembly.
SB 58 (Padilla) of 2025, provides certain tax incentives for data center equipment if those data centers using the equipment meet certain job creation, economic investment, and renewable energy requirements. The bill is currently in the Senate Revenue and Taxation Committee.
SB 1298 (Cortese) of 2024, would have increased the amount of thermal generation a data center could use as backup power from 100 MW to 150 MW without triggering the CEC’s power plant siting process. The bill would have also created conditions for data centers to use this exemption. The bill died in the Assembly.
FISCAL EFFECT: Appropriation: Fiscal Com.: Local:
SUPPORT:
350 Bay Area Action
350 Humboldt: Grass Roots Climate Action
Climate Action California
League of California Cities
Santa Cruz Climate Action Network
Sierra Club
Solano County Democratic Central Committee
Sustainable Rossmoor
An Individual
OPPOSITION:
Bay Area Council
CA Blockchain Advocacy Coalition
California Chamber of Commerce
California Hispanic Chambers of Commerce
California Pacific Asian Chamber of Commerce
Central Valley Business Federation
Data Center Coalition
Information Technology Industry Council
Los Angeles Business Federation
Orange County Business Council
San Diego Regional Chamber of Commerce
San Mateo County Economic Development Association
Santa Barbara South Coast Chamber of Commerce
Silicon Valley Leadership Group
TechCA
TechNet
VICA
ARGUMENTS IN SUPPORT: According to the author:
Across California, energy-intensive data centers are being built to support the rapid expansion of the artificial intelligence (AI) industry. These data centers increase energy demand and frequently require expansions to the electrical grid; together, these factors threaten to increase energy costs for Californians. AB 222 creates accountability for data centers by increasing transparency around their energy use, adopting energy efficiency standards, and preventing grid development costs from being passed onto ratepayers. California’s energy costs are already among the highest in the country, and ratepayers should not be forced to bear the additional costs of AI development.
ARGUMENTS IN OPPOSITION: In opposition, the Silicon Valley Leadership Group states:
Mandating the disclosure of detailed proprietary and sensitive operational details indicating computer workloads, energy consumption and changes, and other elements can be used by business competitors to deduce trade secrets and inform competitive strategies. Sharing of this information may also leave critical energy infrastructure more exposed to cyber threats, industrial espionage, and potential exploitation by foreign adversaries. Further, isolating energy consumption specifically for AI model training and development is difficult to accomplish, given the multifaceted role of data centers. These facilities underpin a diverse array of services, transactions, and applications across numerous industries. Supporting the training and development of AI models is merely one component of the broader functions supported by data centers.
-- END --
ASSEMBLY THIRD READING
ATR Bill Id:AB 222
Author:(Bauer-Kahan)
As Amended Ver:May 23, 2025
Majority vote
SUMMARY
Requires publication of energy use associated with the development and use of artificial intelligence (AI), requires the California Energy Commission (CEC) to track energy usage of data centers in the state, and tasks the California Public Utilities Commission (CPUC) to minimize utility costs related to data center development.
Major Provisions
- Mandates that developers of artificial intelligence (AI) models publically report the energy used to train the model, and, annually, publish amount of energy used in the state by the model.
- Requires data centers to identify themselves to the CEC.
- Mandates the CEC include energy consumption trends for data centers in its integrated energy policy report (IEPR) by January 1, 2027.
- Requires that the California Public Utilities Commission (CPUC) minimize the financial impact of building new electrical infrastructure for data centers on ratepayers.
COMMENTS
Data centers are facilities that house large volumes of high-performance computers, information technology, storage systems, and computing infrastructure. They are crucial for maintaining internet-based communications and providing services such as cloud-based computing, training and inference of artificial intelligence algorithms, as well as mining for cryptocurrency. There are multiple parts of the data center that consume energy, with the primary consumption found in the servers and storage, as well as the cooling system.
Artificial intelligence models are trained on datasets to recognize certain patterns and make decisions without human intervention. This training can take hours or up to months. To generate more sophisticated models, the size of the datasets increase, as does the demand for computational resources (measured in floating point operations per second or FLOPS). As training compute demand grows, the need for larger data centers and the requirement for more energy grows as well.
Energy usage for training AI algorithms has increased significantly over the last 10 years, despite advances in hardware efficiency. For example, GPT -3 required 1.2 Gigawatthours (GWh) to train, while GPT-4 required over 50 GWh of energy to train (equivalent to powering approximately 3,600 homes for a year or 0.02% of the electricity California generates in a year). Training phases are energy intensive and are required to make AI models; however, once the model is made, the energy usage is not over. After training, the AI models are put to use in a process called inference. Every time the model is used it has an additional energy demand. Each query is relatively small, but can lead to significant energy consumption, especially if models will be used continually throughout the day by every Californian. AI queries, like ChatGPT, require about 10x the energy of a standard Google search. If an AI model is commercially available, over the life of the model, the inference phase is likely to demand more energy than the training phase.
The AI industry is growing rapidly, and the demand to train new AI models is accelerating, resulting in data center development and construction. California has more than 270 data centers, concentrated largely around Santa Clara, close to the headquarters of Alphabet, Apple, and Meta. Data centers are already the single largest load for the municipal utility, Silicon Valley Power. Pacific Gas and Electric (PG&E), who provides distribution service in Santa Clara County, is expected to add 3.5 GW of new load attributed to data centers in the next four years, equivalent to adding ~2-3 million new homes on to the grid. As of 2023, 4.4% of all energy used in the United States is consumed by data centers. A study from Lawrence Berkeley National Laboratory predicted that by 2028, data centers would use between 6.7 and 12% of the country's electricity.
The speed at which the data center industry is growing has led to a sudden unpredictability in the future energy market. This growth is occurring alongside other changes to electricity demand such as enhanced electrification in the building and transportation sectors, and climate change induced extreme events. There is little transparency into data center energy usage; transparency which would enable understanding or predict market dynamics. Most tech companies and data center operators don't volunteer their energy and water usage information. However, Google and Oracle have published standard efficiency metrics regarding their power usage (power usage effectiveness or PUE) for individual data centers, and Baidu has published efficiency metrics of their water usage (water-usage effectiveness or WUE) for individual data centers. Consumer groups, academics, and regulators are calling for expanded transparency and the rapid generation of new data regarding AI's direct electricity use. There are ongoing efforts to mandate transparency across the United States and Europe.
As infrastructure needs rise parallel to growth of data center demand, communities are asking who is going to pay for it. Often, the cost of new infrastructure for large industrial customers is borne by the utility and ultimately paid for by the ratepayers. However, there has never been an equivalent predicted growth in energy load attributed to a single industrial customer base in this short time frame. In addition, without some protections, data centers and data center customers may have incentive to shop rates for electricity across the country, potentially leaving expensive new infrastructure stranded.
The language in the bill maintains the CPUC shall determine whether new data center "costs and expenses are just and reasonable in accordance with Section 451." This would not change the CPUC's existing mandate to assess these new costs, but affirm it. Therefore, this language will not meaningfully impact the CPUC's assessment of data center-related costs. The bill also mandates "minimizing shifting of cost," to ratepayers for infrastructure upgrades. The Assembly Committee on Utilities and Energy views this provision as affirming the CPUC's existing authority rather than mandating a new, more stringent cost protection for ratepayers.
The growing AI boom has led to concerns about energy usage transparency and shifting infrastructure and energy usage costs on average ratepayers. AB 222 is an effort to address these concerns.
According to the Author
"Across California, energy-intensive data centers are being built to support the rapid expansion of the artificial intelligence (AI) industry. These data centers increase energy demand and frequently require expansions to the electrical grid; together, these factors threaten to increase energy costs for Californians. AB 222 creates accountability for data centers by increasing transparency around their energy use, adopting energy efficiency standards, and preventing grid development costs from being passed onto ratepayers. California's energy costs are already among the highest in the country, and ratepayers should not be forced to bear the additional costs of AI development."
Arguments in Support
The bill is supported by several environmental advocacy organizations and by both the Center for AI and Digital Policy and the Transparency Coalition. These AI policy groups write that requiring AI developers to disclose energy consumption data and establishing efficiency standards for data centers are best practices in international AI governance. They believe the bill will enhance California's leadership in ethical AI and also prevent undue financial burdens on ratepayers.
It is the author and Assembly Committee on Appropriations' view that the CEC already has authority to establish data center efficiency standards, and therefore, these specific provisions have been removed from the bill language.
Arguments in Opposition
A coalition of industry representatives and associations oppose the bill. They argue, "data centers should not be singled out for different and disparate treatment…The proposed energy reporting requirements could force businesses to disclose sensitive trade secrets, harm their competitive edge, and risk creating safety and security vulnerabilities…isolating energy consumption specifically for artificial intelligence (AI) model training and development is difficult to accomplish, given the multifaceted role of data centers… AB 222 risks stifling job creation and innovation by discouraging data center expansion."
FISCAL COMMENTS
According to the Assembly Committee on Appropriations, the CPUC will likely experience an increase in applications and proceedings related to data centers as a result of this bill. To accommodate this new workload, the CPUC assumes it will need an additional administrative law judge, at an annual cost of $257,000 (Public Utilities Commission Utilities Reimbursement Account).
VOTES
ASM UTILITIES AND ENERGY: 13-5-0
YES: Petrie-Norris, Boerner, Calderon, Mark González, Harabedian, Hart, Irwin, Kalra, Papan, Rogers, Schiavo, Schultz, Zbur
NO: Patterson, Chen, Davies, Ta, Wallis
ASM PRIVACY AND CONSUMER PROTECTION: 10-1-4
YES: Bauer-Kahan, Bennett, Bryan, Irwin, Lowenthal, Ortega, Petrie-Norris, Ward, Wicks, Wilson
NO: Patterson
ABS, ABST OR NV: Dixon, DeMaio, Flora, Pellerin
ASM APPROPRIATIONS: 11-3-1
YES: Wicks, Arambula, Calderon, Caloza, Elhawary, Fong, Mark González, Hart, Pacheco, Pellerin, Solache
NO: Dixon, Ta, Tangipa
ABS, ABST OR NV: Sanchez
UPDATED
VERSION: May 23, 2025
CONSULTANT: Kristen Koenig / U. & E. / (916) 319-2083 FN: 0000892
Date of Hearing: May 14, 2025
ASSEMBLY COMMITTEE ON APPROPRIATIONS
Buffy Wicks, Chair
ABPCA Bill Id:AB 222
Author:(Bauer-Kahan) – As Amended Ver:May 5, 2025
Policy Committee: | Utilities and Energy | Vote: | 13 - 5 |
Privacy and Consumer Protection | 10 - 1 | ||
|
|
Urgency: No State Mandated Local Program: No Reimbursable: No
SUMMARY:
This bill, among other things, requires publication of energy use associated with development and use of artificial intelligence (AI) and tasks the California Public Utilities Commission (CPUC) with scrutinizing utility costs related to data center development.
More specifically, this bill requires the developer of an artificial intelligence (AI) model, before the AI model is used, to publish information on the energy used to develop the AI model and, annually, to publish information on the amount of energy used by the AI model in the previous calendar year and the percentage of that energy that was generated in California.
The bill also requires the California Energy Commission (CEC), by January 1, 2027, to adopt efficiency standards for data centers and include energy consumption trends for data centers in its integrated energy policy report (IEPR).
Finally, the bill directs the CPUC to consider whether just and reasonable any costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center that an electrical investor-owned utility (IOU) has requested the CPUC allow it to recover and, in doing so, to minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center.
FISCAL EFFECT:
It is likely the CPUC will experience an increase in applications and proceedings related to data centers as a result of this bill. To accommodate this new workload, the CPUC assumes it will need an additional administrative law judge, at an annual cost of $257,000 (Public Utilities Commission Utilities Reimbursement Account).
The author describes the bill’s provisions regarding the CEC not as giving the CEC new authority; rather, the bill is to expedite the CEC’s issuance of energy efficiency standards for data centers pursuant to its existing authority. Therefore, the bill does not create new costs for the CEC, but it does create cost pressure, likely in the tens of thousands to hundreds of thousands of dollars, for funding to allow CEC to undertake the energy efficiency-related work it would have otherwise done by January 1, 2027, absent this bill.
The CEC sees it differently, concluding it will need to adopt new regulations as a result of the bill and enforce them, for which the CEC believes it will need a mechanical engineer at an annual cost of $191,105 and an attorney at an annual cost of $178,548. The appropriate funding source would be the Energy Resources Program Account (ERPA), which faces a structural deficit.
COMMENTS:
- Purpose. The author wants AI-related energy demand made public and asserts California ratepayers should not pay for the costs of infrastructure needed to accommodate development of data centers that enable creation and use of AI. According to the author:
Across California, energy-intensive data centers are being built to support the rapid expansion of the artificial intelligence (AI) industry. These data centers increase energy demand and frequently require expansions to the electrical grid; together, these factors threaten to increase energy costs for Californians. AB 222 creates accountability for data centers by increasing transparency around their energy use, adopting energy efficiency standards, and preventing grid development costs from being passed onto ratepayers. California’s energy costs are already among the highest in the country, and ratepayers should not be forced to bear the additional costs of AI development
- Background. Data centers—facilities that house servers, storage devices and other computer-related infrastructure—are, by their nature, energy intensive: it takes a great deal of electricity to power such large concentrations of electronic equipment. California is home to many data centers, especially in the Silicon Valley region of Santa Clara County. Many expect data center development, and the associated demand for electricity, to continue and accelerate as the AI industry grows.
This bill requires the developer of an AI model to publish information on the energy used to develop the AI model and, annually, to publish information on the amount of energy used by the AI model. This is a new requirement—as far as this committee knows, an AI developer, by virtue of being an AI developer, is not obligated, today, to publish such information. Writing together in support of this bill, the Center for AI and Digital Policy and the Transparency Center reference AI principals published by the Organization for Economic Co-operation and Development (OECD) that call for, among other things, AI stakeholders “to proactively engage in responsible stewardship of trustworthy AI in pursuit of beneficial outcomes for people and the planet,” and for “AI Actors to commit to transparency and responsible disclosure regarding AI systems.” These bill proponents describe the energy use disclosure requirement as consistent with the OECD AI principals. The committee is also unaware of a similar energy consumption data disclosure requirement applicable to a developer of other energy-intensive applications.
In addition, the bill requires the CEC, by January 1, 2027, to adopt efficiency standards for data centers and include energy consumption trends for data centers in its IEPR. This is not new—state law empowers the CEC to adopt and regularly adopt energy efficiency standards for buildings, in general, and for appliances. The author’s office describes this requirement not as a new authority of the CEC, but, rather, as an assurance the CEC will act on its existing authority within the bill’s specified timeframe.
The bill also requires the CPUC to consider whether just and reasonable any costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center that an IOU has requested the CPUC allow it to recover from ratepayers. This requirement, to an extent, seems to reiterate standard practice: state law allows the CPUC to approve recovery of only those costs that are just and reasonable. What seems to be new is the statutory requirement that the CPUC minimize shifting costs from a data center’s construction or alteration to ratepayers who do not directly benefit. While current CPUC tariffs generally assign these costs to the requesting customer, this bill codifies that principle into statute specifically for data centers.
- Support and Opposition. The bill is support by several environmental advocacy organizations and, as mentioned above, by both the Center for AI and Digital Policy and the Transparency Coalition, which together write:
By requiring AI developers to disclose energy consumption data and establishing efficiency standards for data centers, AB 222 aligns with best practices seen in international AI governance frameworks…and is consistent with practices of other US states facing similar competing concerns. The bill will not only enhance California's leadership in ethical AI and responsible business practices but also prevent undue financial burdens on ratepayers from the construction of new, energy-intensive data centers.
A coalition of industry representatives and associations oppose the bill, asserting:
data centers should not be singled out for different and disparate treatment…The proposed energy reporting requirements could force businesses to disclose sensitive trade secrets, harm their competitive edge, and risk creating safety and security vulnerabilities…isolating energy consumption specifically for artificial intelligence (AI) model training and development is difficult to accomplish, given the multifaceted role of data centers… AB 222 risks stifling job creation and innovation by discouraging data center expansion.
Analysis Prepared by: Jay Dickenson / APPR. / (916) 319-2081
Date of Hearing: May 1, 2025
Fiscal: Yes
ASSEMBLY COMMITTEE ON PRIVACY AND CONSUMER PROTECTION
Rebecca Bauer-Kahan, Chair
ABPCA Bill Id:AB 222 (
Author:Bauer-Kahan) – As Amended Ver:April 7, 2025
PROPOSED AMENDMENTS
SUBJECT: Data centers: energy usage reporting and efficiency standards: electricity rates
SYNOPSIS
Over the past five years, the development of artificial intelligence (AI) systems has accelerated rapidly, with the potential to transform the economy. These systems include large language models, such as OpenAI’s ChatGPT and Google’s Gemini, general-purpose chatbots trained on massive amounts of information and adaptable to a wide range of tasks. Despite their growing influence, how these systems actually operate and the specifics of their training remain largely opaque.
Recently, many AI developers have announced major investments in expanding data centers to support more powerful computing needs required to train increasingly sophisticated AI systems. However, these expansions have revealed that the training and deployment of AI systems are extraordinarily energy- and resource-intensive. This raises important questions about whether the existing electrical grid can handle the significant increase in electricity demand generated by AI technologies. Additionally, there are concerns that the costs associated with ensuring grid resilience could be unfairly passed on to other ratepayers, many of whom may not directly benefit from these technologies.
This author-sponsored bill would address these concerns by requiring developers of the largest and most power-intensive AI systems to publicly disclose the amount of energy consumed during the training of their models. The bill would also require the California Energy Commission (CEC) to adopt energy efficiency standards for data centers to ensure that California’s climate goals are not compromised in the pursuit of technological advancement. Additionally, data centers would be required to provide specific information to the CEC before beginning operations. The CEC would be mandated to include energy consumption trends for data centers in its Integrated Energy Policy Report. Further, the bill would require the California Public Utilities Commission (PUC) to consider whether the costs and expenses associated with new data center developments should be passed on to other ratepayers who may not directly benefit from the facilities.
This bill is supported by the Center for AI and Digital Policy, the Transparency Coalition, and a variety of environmental organizations, including the Sierra Club. It is opposed by various trade organizations, including TechNet, the California Chamber of Commerce, and the California Blockchain Advocacy Coalition.
While the bill currently focuses on reporting requirements related to the energy consumption involved in training AI models, it is important to note that energy usage continues even after deployment. As noted in Comment #8, Committee amendments would expand the reporting requirements to include the energy usage of AI models post-deployment.
This bill was previously heard by the Utilities and Energy Committee, where it passed with a 13–5 vote.
THIS BILL:
- Defines the following terms:
- “Artificial intelligence” or “AI” means an engineered or machine-based system that varies in its level of autonomy and that can, for explicit or implicit objectives, infer from the input it receives how to generate outputs that can influence physical or virtual environments.
- “Computer room” means a room within a building whose primary function is to house electronic equipment and that has a design information technology equipment power density exceeding 20 watts per square foot or 215 watts per square meters of conditioned floor area.
- “Covered model” means an AI model developed using a quantity of computing power exceeding 10^25 integer or floating-point operations.
- “Data center” means a building the primary function of which is to house one or more computer rooms with an energy demand of at least 5 megawatts.
- “Developer” means a person, partnership, state or local government agency, or corporation that designs, codes, trains, or otherwise produces a covered model.
- “Information Technology Equipment” includes computers, data storage, servers, and network or communication equipment located in a computer room.
- Requires a developer before using a covered model commercially, or before making a covered model available for use by a third party, to do both of the following:
- Calculate both of the following:
- The total energy, in megawatt hours, used to develop the covered model.
- The percentage of total energy used to develop the covered model that was generated in California.
- Publish the energy usage data estimated on the developer’s internet website.
- Calculate both of the following:
- Requires energy usage data published to be all of the following:
- Publicly available for as long as the developer continues to use the covered model or continues to make the covered model available for use by a third party.
- Provided at no cost to a user of the developer’s internet website.
- Accessible and comprehensible to a natural person with ordinary cognitive abilities.
- Requires the CEC to adopt efficiency standards for data centers that do all of the following on or before January 1, 2027:
- Are technologically feasible and cost effective.
- Align with California’s greenhouse gas emission reduction targets.
- Require new data centers and substantial alterations to existing data centers to incorporate load-management and load-shifting capabilities, including the ability to participate in demand response programs.
- Requires data centers to report all of the following information to the CEC before initially operating as a data center and following any change that materially affects the correctness of information previously provided to the commission:
- The name of the data center.
- Any physical addresses associated with the operation of the data center.
- The internet website of the data center if the data center has an internet website.
- A telephone number and email address for a point of contact for the commission to communicate with the data center.
- Requires the CEC to include energy consumption trends for data centers in its integrated energy policy report.
- Requires the PUC to determine whether those costs and expenses are just and reasonable in an application by an electrical corporation to recover costs and expenses arising from, or incurred as a result of, the construction of a new data center or a substantial alteration to an existing data center.
- Requires the PUC to minimize the shifting of costs attributable to the construction or alteration of the data center to ratepayers who do not directly benefit from the data center in determining whether costs and expenses are just and reasonable.
EXISTING LAW:
- Mandates the PUC to ensure that all charges from a public utility shall be just and reasonable. (Pub. Util. Code § 451.)
- Mandates the CEC to develop energy policies that conserve resources, protect the environment, ensure energy reliability, enhance the state's economy, and protect public health and safety. (Pub. Res. Code § 25301(a).)
- Requires the CEC to generate an Integrated Energy Policy Report, or IEPR every two years, which will include, among other things, an assessment of resources and a forecast of reliability and energy usage. (Pub. Res. Code § 25302.)
- Grants the CEC the authority to hold public hearings and stakeholder processes to best assess the necessary reporting and efficiency standards for buildings. (Pub. Res. Code § 25402(b)(4).)
COMMENTS:
1) Author’s statement. According to the author:
Across California, energy-intensive data centers are being built to support the rapid expansion of the artificial intelligence (AI) industry. These data centers increase energy demand and frequently require expansions to the electrical grid; together, these factors threaten to increase energy costs for Californians. AB 222 creates accountability for data centers by increasing transparency around their energy use, adopting energy efficiency standards, and preventing grid development costs from being passed onto ratepayers. California’s energy costs are already among the highest in the country, and ratepayers should not be forced to bear the additional costs of AI development.
2) AI and GenAI. The development of GenAI is creating exciting opportunities to grow California’s economy and improve the lives of its residents. GenAI can generate compelling text, images and audio in an instant – but with novel technologies come novel safety concerns.
In brief, AI is the mimicking of human intelligence by artificial systems such as computers. AI uses algorithms – sets of rules – to transform inputs into outputs. Inputs and outputs can be anything a computer can process: numbers, text, audio, video, or movement. AI is not fundamentally different from other computer functions; its novelty lies in its application. Unlike normal computer functions, AI is able to accomplish tasks that are normally performed by humans.
AI that are trained on small, specific datasets in order to make recommendations and predictions are sometimes referred to as “predictive AI.” This differentiates them from GenAI, which are trained on massive datasets in order to produce detailed text and images. When Netflix suggests a TV show to a viewer, the recommendation is produced by predictive AI that has been trained on the viewing habits of Netflix users. When ChatGPT generates text in clear, concise paragraphs, it uses GenAI that has been trained on the written contents of the internet.
GenAI tools can be released in open-source or closed-source formats by their creators. Open-source tools are publically available; researchers and developers can access their code and parameters. This accessibility increases transparency, but it has downsides: when a tool’s code and parameters can be easily accessed, they can be easily altered, and open-source tools have the potential to be used for nefarious purposes such as generating deepfake pornography and targeted propaganda. By comparison, closed-source tools are opaque with respect to their security features. It is harder for bad actors to generate illicit materials using these tools. But unlike open-source tools, closed-source tools are not subject to collective oversight because their inner workings cannot be examined by independent experts.
3) Data Centers and AI. Data centers are facilities that house critical computing infrastructure, including servers, data storage systems, and network equipment. These facilities support a wide range of functions, from running web-based applications and supporting e-commerce to ensuring cybersecurity operations. Data centers can be located on-site at a company’s premises, operated as colocation facilities where businesses rent space for their servers, or exist as cloud data centers, where companies lease computing infrastructure from third-party providers and outsource the maintenance of physical hardware.[1] According to a Lawrence Berkeley National Laboratory study, data centers currently account for approximately 4.4% of total U.S. energy consumption. However, this energy use is highly concentrated in specific regions, such as Northern Virginia and California’s Bay Area.[2]
Data centers have become a cornerstone of the AI revolution. Training sophisticated AI systems, such as large language models (LLMs), requires tremendous computing power (measured in floating-point operations per second, or FLOPS) and extensive data storage, both of which are provided by modern data centers. Unlike most traditional computing tasks, which rely primarily on central processing units (CPUs), AI systems utilize graphics processing units (GPUs). CPUs are designed for the sequential processing of information, making them ideal for everyday electronic devices. In contrast, GPUs excel at parallel processing, enabling them to handle the massive datasets required to train AI models more efficiently.[3]
AI systems are typically trained to recognize patterns and generate predictions based on statistical analyses of user inputs, processes that demand immense and simultaneous computations. Consequently, GPUs have become the workhorse of AI development. Notably, GPUs are now being specifically engineered for greater stability and efficiency to better meet the operational demands of data centers.[4]
4) AI’s Amped up Energy Boom. Several major companies investing in AI, including Microsoft, Google, and Meta, have announced multibillion-dollar commitments to expand their data center infrastructure and boost computing power.[5] It should be noted that Microsoft has recently pulled back on some of its commitments, likely in response to the federal administration’s evolving tariff policies.[6] Nevertheless, these mass investments are expected to further accelerate the development of AI technologies across the United States. However, they also raise critical questions about the strain new data centers may place on California’s power grid and whether the proliferation of facilities could further stress an already burdened system.
As noted above, AI development demands far more FLOPS than traditional computing. This stems from the immense energy requirements of GPUs, which are used to process vast datasets simultaneously. AI models must be trained and retrained through numerous iterations to ensure proper functionality, requiring continuous, around-the-clock electricity for weeks or even months. Additionally, data centers must expend significant energy and water resources to keep these facilities cool enough to prevent equipment from overheating, compounding their already substantial energy needs.[7]
The California Chamber of Commerce, along with a coalition of trade groups, argues that modern data centers have become more energy-efficient, stating:
Data centers are highly efficient facilities that enable energy savings and efficiencies economy-wide. In 2010, nearly 80 percent of data center computing was conducted in smaller traditional computer centers, largely owned and operated by non-technology companies. By 2018, approximately 89% of data center computing took place in larger cloud data centers. By centralizing computing resources, data centers have leveraged innovations in design, equipment, and technology to maximize energy efficiency. While electricity consumption at data centers rose 6% from 2010 to 2018, computing output jumped 550%, marking significant gains in efficiency and productivity.
While it is true that hardware improvements have led to greater energy efficiency, the data cited does not account for the explosive energy demands of the AI sector. Data centers accounted for 4.4% of U.S. electricity consumption in 2023, and Lawrence Berkeley National Laboratory projects that figure will rise to between 6.7% and 12% by 2028.[8] On a granular level, OpenAI’s GPT-4, released in 2024, consumed 50 gigawatt-hours of energy during training, 50 times more than GPT-3, which is equivalent to the annual energy consumption of 3,600 California homes.[9] This soaring demand has prompted AI companies to pursue alternative energy strategies. For instance, Microsoft recently struck a deal with Constellation Energy to reopen the Three Mile Island nuclear power plant off the coast of South Carolina to power its AI operations.[10]
Some progress in energy-efficient AI development has been demonstrated by the Chinese AI company DeepSeek with the release of its DeepSeek-R1 chatbot. This model performs comparably to major LLMs like GPT, Gemini, and LLaMA but was significantly less energy-intensive and cheaper to train. DeepSeek achieved these efficiencies by training specific parts of the model individually rather than all at once, a technique that could reduce overall energy usage across the industry.[11] However, paradoxically, as the energy and cost barriers to training AI models lower, more players could enter the market, potentially increasing overall energy consumption despite individual gains in efficiency.[12]
Energy use does not stop at model training. Once deployed, AI models engage in inference, the process of generating predictions or answering inquiries based on user inputs, which remains highly energy-intensive. For example, a single query on ChatGPT requires ten times more energy than a traditional Google search.[13] Generating one AI-created image can consume energy equivalent to fully charging a smartphone. Because these systems operate continuously, the cumulative energy impact is significant. Notably, inference energy consumption has not seen the same efficiency gains as training. DeepSeek’s model, for instance, uses a comparable amount of energy per inquiry as other major AI models.[14]
5) What this bill would do. As energy consumption associated with the development and deployment of AI models continues to grow, this bill would establish several mechanisms to increase transparency regarding the energy required to train AI systems and the overall energy usage of data centers. Additionally, the bill would empower the PUC to prevent the costs of building or renovating data centers from being unfairly shifted onto other ratepayers who may not directly benefit.
Specifically, the bill would require developers of AI models that use a computing capacity exceeding 10²⁵ integer or FLOPS to calculate the total amount of energy consumed during the training of the model, including the portion specifically used in California, prior to releasing the model to the public. Developers must then publish this information prominently on their websites, ensuring it remains accessible for as long as the model is in use. The disclosures must be freely available to the public and presented in a way that is easy to understand.
The bill would also require the CEC to adopt energy efficiency standards for data centers that are technologically feasible and aligned with the state’s greenhouse gas emissions goals. In addition, the CEC would be tasked with adopting standards to encourage load management and load-shifting capabilities, such as participation in demand response programs, for both new and existing data centers.
Furthermore, before beginning operations, new data centers would be required to report key information to the CEC, including the center’s name, address, website, and a point of contact. The CEC would also be required to track and report on energy consumption trends for data centers as part of its Integrated Energy Policy Report.
Finally, the bill would direct the PUC to assess whether the costs and expenses incurred by electrical corporations for the construction or substantial modification of data centers are reasonable. Crucially, the PUC would be required to minimize the shifting of these costs onto ratepayers who do not directly benefit from the data center’s operations.
6) Trade Secret? One major question regarding AI companies’ energy usage is whether it is trade secret. The Los Angeles Business Federation writing in opposition argues:
The proposed energy and water disclosure requirements would force companies to share highly sensitive operational data that could reveal proprietary business information. For multi-tenant data centers—serving hundreds of diverse customers, including financial services, cloud providers, and public institutions—this level of disclosure jeopardizes competitive positioning and creates a disincentive to expand or invest in California.
Under California law, a “trade secret” is defined as information, including a formula, pattern, compilation, program, device, method, technique, or process, that (1) derives independent economic value, actual or potential, from not being generally known to the public or to others who could obtain economic value from its disclosure or use, and (2) is subject to reasonable efforts to maintain its secrecy.[15] To evaluate the claim that energy usage constitutes a trade secret, there must be verifiable evidence that such information is essential to protect and that its disclosure would reasonably harm a company's financial interests.
It is undisputable that energy consumption is integral to the training of AI models. However, the amount of energy consumed does not necessarily correlate with the quality or innovation of the models themselves. For example, DeepSeek developed a model comparable to other leading AI systems while using significantly less energy. Additionally, Google disclosed the total energy usage for training its large language models in 2022 and 2023, without suffering any apparent loss of competitive advantage.[16] On the contrary, DeepSeek’s disclosure enhanced its competitive position by demonstrating greater efficiency relative to its rivals.
While companies like Meta and Apple have not disclosed the specific energy usage for training their AI models, they have regularly reported the overall electricity usage of their data centers.[17] With sufficient knowledge of industry standards and operations, competitors and observers could likely estimate the energy consumed for AI development based on these broader disclosures. Thus, while some concerns over proprietary information are understandable, the argument that disclosing energy usage alone constitutes a protectable trade secret appears to lack strong support in practice.
7) Energy disclosures a security risk? The other aspect of this bill that sits in this committee’s jurisdiction is whether disclosures of energy usage are a threat to data centers’ cybersecurity. The Data Center Coalition, alongside a coalition of trade groups, argues that energy disclosure could endanger cybersecurity, stating:
Sharing of this information may leave this critical infrastructure more exposed to cyber threats, industrial espionage, and potential exploitation by foreign adversaries. In Virginia, the largest data center market in the U.S., a similar measure to impose such reporting requirements was introduced during this year’s legislative session but was not advanced due to national security concerns.
The letter references the Virginia bill HB 2035, which opponents assert did not advanced due to national security concerns. However, the bill was withdrawn by its author during a hearing in the Virginia Labor and Commerce Subcommittee #3, and the archives on Virginia’s legislative website do not provide any evidence linking the bill’s withdrawal to national security issues.[18] Even if national security concerns were cited as the reason for halting the bill, it’s important to note that Virginia’s data centers, particularly those located around Washington, D.C., likely have a more significant role in national security than AI models developed and marketed for public use. It remains unclear how disclosing the amount of energy used by an AI model developer could expose cybersecurity to threats from foreign adversaries.
This bill does require that new data centers report certain identifying characteristics to the CEC, but it does not mandate that this information be made publicly accessible. Under current law, any building larger than 50,000 square feet, including data centers, must disclose its energy usage to the CEC for benchmarking purposes, with 59 such buildings reported in 2023.[19] This information is accessible via an interactive map on the CEC website and has not led to any reported major security risks. And, again, the information required under this bill would be shared only with the CEC, not the public.
8) Amendments. As noted earlier, the energy consumption of an AI model extends well beyond its deployment. Recently, Sam Altman, CEO of OpenAI, shared on the social media platform X that even something as simple as saying “please” and “thank you” to their models has cost OpenAI “tens of millions of dollars” due to the additional compute power required to process those extra words.[20] Furthermore, an investigative piece by The Washington Post, in collaboration with researchers from the University of California, Riverside, found that a single 100-word AI written email requires 0.14 kilowatt-hours (kWh) of electricity, which is enough to power 14 LED light bulbs for one hour. If a person were to send such an email once a week for an entire year, this would amount to 7.5 kWh, the equivalent of powering 9.3 households in Washington, D.C., for one hour.[21]
Given the significant energy demands associated with operating AI models, the author has agreed to an amendment requiring developers of covered models to annually disclose the amount of energy used in their operation. This amendment would apply only to the developers of AI models under their control and would not extend to third parties using open-access models. The Committee amendment is as follows:
25345.1. (a) Before using a covered model commercially, or before making a covered model available for use by a third party, a developer shall do both of the following:
(a) Calculate(1) Estimate both of the following:
(1A) The total energy, in megawatthours, used to develop the covered model.
(2B) The percentage of total energy used to develop the covered model that was generated in California.
(b) (2) Publish the energy usage data estimated pursuant to subdivision (a) paragraph (1) on the developer’s internet website.
(b) Beginning January 1, 2027, and annually on January 1 thereafter, a developer shall do both of the following:
(1) Estimate both of the following:
(A) The total energy, in megawatthours, used by the developer to operate the covered model during the previous calendar year.
(B) The percentage of the total energy estimated pursuant to subparagraph (A) that was generated in California.
(2) Publish the energy usage data estimated pursuant to paragraph (1) on the developer’s internet website.
(2c) Energy usage data published pursuant to paragraph (1) this section shall be all of the following:
(A1) Publicly available for as long as the developer continues to use the covered model or continues to make the covered model available for use by a third party.
(B2) Provided at no cost to a user of the developer’s internet website.
(C3) Accessible and comprehensible to a natural person with ordinary cognitive abilities
ARGUMENTS IN SUPPORT: California Environmental Voters, write in support:
The rapid development of the AI industry is fueling a boom in data center construction in California. These facilities, which house the servers and hardware needed to train and operate AI, require vast amounts of electricity. According to a recent article in the Los Angeles Times: “In Santa Clara — the heart of Silicon Valley — electric rates are rising as the municipal utility spends heavily on transmission lines and other infrastructure to accommodate the voracious power demand from more than 50 data centers, which now consume 60% of the city’s electricity. And earlier this year, Pacific Gas & Electric told investors that its customers have proposed more than two dozen data centers, requiring 3.5 gigawatts of power — the output of three new nuclear reactors. ”
Large AI models require a tremendous amount of energy to develop. At the same time, the rapid expansion of the AI industry is driving the widespread construction of energy-intensive data centers across the state. The increase in energy demand due to AI, paired with grid infrastructure expansion necessitated by the construction and expansion of data centers, threatens to increase energy costs for Californians
AB 222 creates transparency around energy usage by developers of artificial intelligence (AI) models and data centers, creates energy efficiency standards for data centers, and prevents the cost of constructing new data centers from being passed onto ratepayers.
ARGUMENTS IN OPPOSITION: In opposition to the bill, Technet alongside a coalition of trade groups argues:
Mandating the disclosure of detailed proprietary and sensitive operational details indicating computer workloads, energy consumption and changes, water consumption and changes, water sourcing, and other elements can be used by business competitors to deduce trade secrets and inform competitive strategies. Sharing of this information may leave this critical infrastructure more exposed to cyber threats, industrial espionage, and potential exploitation by foreign adversaries. In Virginia, the largest data center market in the U.S., a similar measure to impose such reporting requirements was introduced during this year’s legislative session but was not advanced due to national security concerns.
Additionally, isolating energy consumption specifically for artificial intelligence (AI) model training and development is difficult to accomplish, given the multifaceted role of data centers. These facilities underpin a diverse array of services, transactions, and applications across numerous industries. Supporting the training and development of AI models is merely one component of the broader functions supported by data centers.
User demand by industry also varies by market. For instance, a 2024 analysis reveals that while the Los Angeles data center market allocates a significant portion (55%) to cloud computing, alongside supporting sectors like technology, telecom, and healthcare, the Northern California market demonstrates an even greater concentration, with approximately 70% of its capacity dedicated to cloud compute.
Data centers fuel economic growth, supporting jobs in technology, construction, engineering, and maintenance. Data centers also attract investment from cloud providers, AI developers, and enterprise IT, strengthening California’s digital economy. AB 222 risks stifling job creation and innovation by discouraging data center expansion.
California's position as a global technology leader depends on the vital infrastructure provided by data centers underscores the need for careful consideration of proposed legislation that could impact this critical industry. We strongly urge the Committee to resist advancing AB 222 or other measures that impose disparate reporting requirements, as these create unnecessary burdens and raise serious concerns related to privacy, security, feasibility, and competition, while failing to deliver tangible benefits to the state.
REGISTERED SUPPORT / OPPOSITION:
Support
350 Bay Area Action
California Environmental Voters (formerly Clcv)
Compute Exchange INC.
Sierra Club
Sustainable Rossmoor
The Center for Ai and Digital Policy
Transparency Coalition.ai
Opposition
2 Individuals
Bay Area Council
CA Blockchain Advocacy Coalition
CalChamber
California African American Chamber of Commerce
California Hispanic Chambers of Commerce
California Hispanic Chambers of Commerce (CHCC)
California Pacific Asian Chamber of Commerce (CalAsian Chamber)
California State Association of Electrical Workers
Central Valley Business Federation
Chamber San Mateo County
Coalition of California Utility Employees
Data Center Coalition
Los Angeles County Business Federation
Los Angeles County Business Federation (BIZFED)
San Diego Regional Chamber of Commerce
San Mateo County Economic Development Association (SAMCEDA)
Santa Barbara South Coast Chamber of Commerce
Silicon Valley Leadership Group
Techca
Technet
Vica
Analysis Prepared by: John Bennett / P. & C.P. / (916) 319-2200
Stephanie Susnjara and Ian Smalley, “What is a data center?”, IBM (Sept. 4, 2024), accessed at http://ibm.com/think/topics/data-centers. ↑
Arman Shehabi et al. “2024 United States Data Center Energy Usage Report”, Lawrence Berkeley National Laboratory (Dec. 2024), accessed at https://www.energy.gov/articles/doe-releases-new-report-evaluating-increase-electricity-demand-data-centers. ↑
Amazon “What’s the Difference Between GPUs and CPUs?” Amazon Web Services, accessed at https://aws.amazon.com/compare/the-difference-between-gpus-cpus/. ↑
Ibid. ↑
Karen Weise, Laura Bult, James Surdam and Ramon Dompor, “How A.I. Companies Are Turning Into Energy Companies”, The New York Times (Mar. 17, 2025), accessed at https://www.nytimes.com/video/business/energy-environment/100000010036088/how-ai-companies-are-turning-into-energy-companies.html. ↑
Joe Wilkins, “Something's Gone Wrong With Microsoft's Huge AI Data Center Investments”, Futurism (Apr. 3, 2025), accessed at https://futurism.com/microsoft-huge-data-center-investments-tariffs. ↑
David Gelles, “The A.I. Power Grab”, The New York Times (Oct. 22, 2024), accessed at https://www.nytimes.com/2024/10/22/climate/ai-big-tech-emissions.html. ↑
Arman Shehabi et al. “2024 United States Data Center Energy Usage Report”, Lawrence Berkeley National Laboratory (Dec. 2024), accessed at https://www.energy.gov/articles/doe-releases-new-report-evaluating-increase-electricity-demand-data-centers. ↑
Tejas Dessai, “AI Hinges on Rapid Electrification”, Global X (Feb. 27, 2025), accessed at https://www.globalxetfs.com/articles/ai-hinges-on-rapid-electrification. ↑
C. Mandler, “Three Mile Island nuclear plant will reopen to power Microsoft data centers”, NPR (Sept. 20, 2024), accessed at https://www.npr.org/2024/09/20/nx-s1-5120581/three-mile-island-nuclear-power-plant-microsoft-ai. ↑
Lauren Laws, “Why DeepSeek could be good news for energy consumption”, University of Illinois Urbana-Champaign (Feb. 6, 2025), accessed at https://grainger.illinois.edu/news/stories/73489. ↑
Greg Rosalsky, “Why the AI world is suddenly obsessed with a 160-year-old economics paradox”, NPR (Feb. 4, 2025), accessed at https://www.npr.org/sections/planet-money/2025/02/04/g-s1-46018/ai-deepseek-economics-jevons-paradox. ↑
Marcel Salathé, “Does ChatGPT use 10x more energy than a standard Google search?”, Substack (Jan. 15, 2025), accessed at https://engineeringprompts.substack.com/p/does-chatgpt-use-10x-more-energy?utm_campaign=post&utm_medium=web. ↑
Christa Marshall, “‘Game changer’? What ‘DeepSeek’ AI means for electricity.”, POLITICO (Jan. 29, 2025), accessed at https://www.eenews.net/articles/game-changer-what-deepseek-ai-means-for-electricity/. ↑
Civ. Code §§ 3426–3426.11. ↑
Eric Masanet, Nuoa Lei, and Jonathan Koomey, “To better understand AI’s growing energy use, analysts need a data revolution”, Joule 8, 1–10 (September 18, 2024), accessed at https://www.sciencedirect.com/science/article/abs/pii/S2542435124003477. ↑
Ibid. ↑
Bill information can be found at https://lis.virginia.gov/bill-details/20251/HB2035 and stream of hearing can be found at https://virginiageneralassembly.gov/house/chamber/chamberstream.php. ↑
AB 802 Williams, Ch. 590, Stats. 2015. ↑
Joe Wilkins, “Sam Altman Admits That Saying ‘Please’ and ‘Thank You’ to ChatGPT Is Wasting Millions of Dollars in Computing Power”, Futurism (Apr. 19, 2025), accessed at https://futurism.com/altman-please-thanks-chatgpt. ↑
Pranshu Verma and Shelly Tan, “A bottle of water per email: the hidden environmental costs of using AI chatbots,” The Washington Post (Sept. 18, 2024), accessed at https://css.washingtonpost.com/technology/2024/09/18/energy-ai-use-electricity-water-data-centers/. ↑
Discussed in Hearing
Senate Standing Committee on Appropriations
Senate Standing Committee on Judiciary
Senate Standing Committee on Energy, Utilities and Communications
Assembly Floor
Assembly Standing Committee on Privacy and Consumer Protection
Assembly Standing Committee on Utilities and Energy
Bill Author
Bill Co-Author(s):