AI Energy Costs and the Ratepayer Protection Pledge: Why Nonprofits Should Care
America's AI boom is running on electricity, and someone has to pay for it. The Trump administration's Ratepayer Protection Pledge promises that Big Tech, not ordinary consumers, will foot the bill for the power demands of AI data centers. But energy experts say the pledge is largely unenforceable, and for nonprofits already squeezed by rising operating costs and surging community need, the stakes could not be higher.
On February 24, 2026, President Trump used his State of the Union address to unveil the "Ratepayer Protection Pledge," a commitment from the country's largest technology companies to cover the electricity costs their AI data centers impose on the grid, rather than allowing those costs to flow through to consumers. Amazon, Google, Meta, Microsoft, xAI, Oracle, and OpenAI were all named as participants. A formal White House signing ceremony with the tech companies was scheduled for March 4.
The pledge arrives against a backdrop of sharply rising electricity bills that have become one of the most politically charged issues in the country. U.S. data centers consumed more than 4% of total national electricity in 2024, according to the International Energy Agency, and AI-specific data center demand is projected to grow roughly tenfold between 2022 and 2026. Utilities have responded by requesting record rate increases, and the costs have landed hardest on households and small organizations with no ability to negotiate their own power supply.
For nonprofits, this is not a distant policy debate. Energy is typically the second-highest operating expense for community-serving organizations, behind only salaries. The same communities many nonprofits exist to serve are disproportionately experiencing energy poverty. And the pledge itself, according to energy policy experts and advocacy organizations, may offer far less protection than its name implies.
This article explains what the Ratepayer Protection Pledge actually promises, why critics say it falls short structurally, how AI energy demand has been affecting electricity costs in practice, what legislative responses are taking shape at the federal and state level, and what nonprofit leaders should consider in response.
What the Ratepayer Protection Pledge Actually Says
The core logic of the pledge is straightforward: tech companies operating large AI data centers should "build, bring, or buy" their own power supply rather than drawing from the shared grid and having associated infrastructure and capacity costs distributed across all ratepayers. Energy Secretary Chris Wright stated that "all of the brand-name hyperscalers" had signed on. Anthropic made the most specific public commitment, pledging to cover 100% of electricity price increases that consumers face as a result of its data center operations.
The administration framed the pledge as a natural extension of its goal to maintain U.S. dominance in artificial intelligence while protecting ordinary Americans from the collateral costs of that dominance. The political logic is clear: the same voters who support AI leadership are also experiencing electricity bill shock, and the pledge is designed to convince them these goals are compatible.
Companies Named in the Pledge
- Amazon
- Meta
- Microsoft
- OpenAI
- Oracle
- xAI
What the Pledge Commits To
- "Build, bring, or buy" their own power
- Fund generation capacity for new data centers
- Avoid shifting energy costs to consumers
- Anthropic: cover 100% of consumer price increases
Why Energy Experts Say the Pledge May Not Deliver
The Ratepayer Protection Pledge has drawn skepticism from energy policy experts, consumer advocates, and even some utility regulators. Their critiques center on three structural problems that cannot be resolved through voluntary commitments alone.
Problem 1: Transmission and Distribution Costs Are Not Covered
The most expensive part of serving data centers may be off the table entirely
Grid expert Brandon Owens identified the core structural flaw: "Most of today's cost pressure is coming from transmission, distribution, and system readiness, not energy supply. Those costs remain even if a data center self-supplies generation." In other words, a data center that builds its own solar farm still requires the grid's transmission lines, substations, and reliability infrastructure to deliver power, and those costs are typically socialized across all ratepayers regardless of the generation source.
PJM Interconnection, which operates the grid serving 67 million people across 13 mid-Atlantic and Midwestern states, recently approved $11.8 billion in new transmission projects where data centers are the primary beneficiaries. Under standard utility rate formulas, those costs are distributed across all ratepayers, regardless of whether the data centers self-supply their generation.
Problem 2: No Enforcement Mechanism
A voluntary pledge with no legal force and no penalties for non-compliance
Jackson Voss of the Alliance for Affordable Energy described the pledge as "a handshake agreement between Silicon Valley and the White House" with no legal binding force. There is no legislation attached, no regulatory framework, no penalties for companies that do not follow through, and no mechanism for consumers or advocacy groups to seek redress if electricity bills continue rising.
Food and Water Watch called the pledge a "theatrical stunt," and Legal Planet's analysis concluded it was "more BS on electricity prices," noting that the administration lacks the authority to compel compliance without congressional action.
Problem 3: States Control Electricity Rates, Not the Federal Government
Jurisdiction is a fundamental barrier to federal rate protection
Electricity rate-setting in the United States is primarily a state function, regulated by state public utility commissions. A federal pledge from the White House cannot override the decisions of state utility regulators, who approve rate cases based on their own legal standards and the evidence before them in individual proceedings.
This means that even if a tech company sincerely intends to cover its energy costs, the legal mechanisms for actually preventing those costs from flowing through to ratepayers run through state regulatory processes that the federal pledge does not address. Consumers in Virginia, New Jersey, Maryland, and other high-impact states have no guarantee that their utility bills will reflect the pledge's intentions.
The Scale of AI Energy Demand: What the Numbers Show
Understanding why this issue matters requires understanding the scale of AI energy demand and how dramatically it has accelerated. The growth has been faster than most forecasters predicted just a few years ago, and the trajectory is not leveling off.
U.S. data centers consumed approximately 183 terawatt-hours of electricity in 2024, representing more than 4% of total national electricity consumption, roughly equivalent to the annual electricity demand of Pakistan. AI-specific data centers are projected to reach 90 terawatt-hours of annual consumption by 2026, roughly a tenfold increase from 2022 levels. By 2028, data centers are expected to represent between 6.7% and 12% of total U.S. electricity consumption, and Goldman Sachs projects that data centers will account for 40% of all U.S. electricity demand growth going forward.
These numbers have translated directly into the electricity rate environment. Utilities filed a record $31 billion in rate hike requests in 2025, more than double the $15 billion requested in 2024. Electricity prices rose 6.9% year-over-year in 2025, more than double the headline inflation rate of 2.9% during the same period. Average monthly energy costs nationwide rose from $196 to $265 between 2022 and 2025, a 35% jump that was nearly three times the overall inflation rate during that period.
U.S. data center electricity use in 2024, equal to Pakistan's annual demand
Utility rate hike requests in 2025, double the prior year's record
PJM capacity price increase in Virginia for 2025-2026, driven by data center demand
The regional concentration of these effects has been particularly striking. Virginia, home to roughly 35% of all known AI data centers worldwide, has experienced some of the most acute rate pressure. Dominion Energy, which serves approximately 450 data centers in the state, approved rate increases of $565.7 million in 2026 and $209.9 million in 2027 based on data center-driven demand growth. The PJM capacity market auction price for 2025-2026 increased 833% in Virginia compared to the prior year. New Jersey saw average electric bills surge more than 20% in 2025 alone, creating visible political backlash that has begun influencing elections.
The Nonprofit Double Burden: Higher Costs, Higher Demand
For nonprofits, AI energy costs create a double burden that deserves direct attention. Organizations face rising operational electricity costs on one side, and rising demand for their services from communities experiencing energy poverty on the other. Both sides of this squeeze have accelerated in the same period.
Energy is typically the second-highest operating expense for nonprofits that operate physical facilities, behind only staffing. Community centers, shelters, food banks, health clinics, houses of worship, and school-serving organizations all carry significant energy loads that they cannot easily reduce without capital investments most lack access to. When electricity rates rise, these costs increase without any corresponding increase in program revenue, donations, or government funding.
The Community Purchasing Alliance, a cooperative purchasing organization serving nonprofits in the DC, Maryland, and Ohio region, documented this directly in 2025. Beginning in June 2025, when PJM's record capacity auction costs began flowing through to consumers, their member organizations faced sudden spikes in electricity supply costs. Electricity suppliers invoked "Change in Law" clauses to pass through the increased PJM capacity costs directly, with Pepco increasing default rates by up to 25% for some customers. This happened simultaneously with government funding cuts, freezes, and stop-work orders that affected a large share of community nonprofits in the first half of 2025.
On the demand side, the communities many nonprofits serve are among those most vulnerable to energy cost increases. Approximately 21 million U.S. households were behind on their utility bills as of mid-2025. Total outstanding utility bill debt reached $25 billion in June 2025, up from about $15 billion in early 2022. Utility shut-offs surpassed 3.5 million in 2024. One in seven American families lives in energy poverty, spending 10% to 20% or more of household income on utilities. These are the people who show up at food banks, emergency shelters, utility assistance programs, and other community services, and their numbers are growing.
The Compounding Pressure on Nonprofits
- Rising electricity rates increase the cost of operating physical facilities, from shelters to community centers to health clinics
- Energy poverty in served communities drives higher demand for utility assistance programs, food banks, and emergency services
- Federal and state funding cuts have coincided with the electricity cost spike, leaving organizations with fewer resources to absorb both pressures simultaneously
- Utility suppliers' "Change in Law" clauses allow pass-through of grid cost increases with minimal notice, disrupting budget planning
- Capital-constrained organizations cannot easily invest in efficiency upgrades that would reduce long-term energy costs
Legislative Responses: What Is Actually Being Proposed
While the White House pledge is voluntary and faces structural limitations, binding legislative action is taking shape at both the federal and state level. The pace of state-level activity in particular has accelerated significantly in 2026, reflecting how politically salient the electricity cost issue has become across party lines.
Federal Legislative Action
Representative Greg Landsman of Ohio introduced the Protecting Families from AI Data Center Energy Costs Act in December 2025. The bill would require FERC, the Federal Energy Regulatory Commission, to hold a technical conference specifically focused on protecting residential ratepayers from increased costs associated with large AI loads. Senator Chris Van Hollen of Maryland introduced companion legislation in the Senate through the Power for the People Act, which would require data centers to cover their own energy infrastructure costs rather than having them distributed across consumer bills.
Senators Warren, Van Hollen, and Blumenthal launched a formal investigation into Big Tech's role in driving up consumer electricity costs, sending letters directly to major data center operators requesting documentation of their energy use and the costs flowing to ratepayers. NBC News reported the introduction of the first bipartisan Senate legislation specifically aimed at curbing utility bill hikes connected to data centers, signaling that the issue is beginning to transcend the usual partisan divide.
State-Level Activity
The most significant action has been at the state level. More than 300 bills were filed across 30 or more states in just the first six weeks of 2026 alone, representing a dramatic shift from incentive-focused data center policy to regulatory and accountability-focused policy. Seven states, including Indiana, Maryland, Minnesota, Missouri, Oregon, Texas, and Utah, had already enacted legislation by early 2026 requiring large electricity users to pay for the costs of increased generation rather than socializing those costs.
California Governor Newsom signed SB 57, directing the state's Public Utilities Commission to study data center impacts on ratepayers, though Big Tech lobbied to prevent further regulatory action. Georgia introduced legislation prohibiting utilities from passing fuel, generation, or transmission costs of data centers to other customers. Virginia, the state with the greatest concentration of data centers, created a new rate class for the largest electricity users effective January 2027, which will increase data center customer rates by 15.8% while reducing typical residential rates by approximately 3.4% once it takes effect.
Federal Action
- H.R. 6529: Protecting Families from AI Data Center Energy Costs Act
- Power for the People Act (Senate companion)
- Senate investigation by Warren, Van Hollen, and Blumenthal
- First bipartisan effort to curb utility bill hikes
State Action
- 300+ bills filed in 30+ states in early 2026
- 7 states enacted cost-responsibility legislation
- 18+ states proposed special rate classes for large energy users
- 14 states have towns with data center moratoriums
What Big Tech Is Actually Doing on Energy
Separate from the pledge, the major technology companies have been making substantial investments in their own power supply, driven by a combination of genuine sustainability commitments, economic incentives to secure stable low-cost power, and a desire to get ahead of regulatory requirements. These investments are real and significant, though energy analysts note they do not fully resolve the ratepayer cost question.
Google announced on February 24, 2026, the same day as the State of the Union, a deal to bring nearly 2 gigawatts of clean power online through a contract with Xcel Energy, alongside a separate 150-megawatt geothermal contract. Google has adopted what it calls a "bring your own power" model, arranging direct power purchase agreements rather than relying on standard utility grid service. Microsoft announced a $10 billion renewable energy deal with Brookfield Asset Management, deploying more than 10.5 gigawatts of renewable capacity beginning in 2026, equivalent to roughly 10 nuclear power plants, while also entering the world's first fusion energy purchase agreement with Helion Energy.
Across the sector, tech companies have signed contracts for more than 10 gigawatts of possible new nuclear capacity and have announced plans to finance more than 20 gigawatts of small modular reactors. The four largest cloud and AI companies collectively accounted for close to half of the clean power capacity added globally in 2025.
The important caveat that energy analysts consistently raise is that even with these investments, natural gas and coal are expected to meet more than 40% of the additional electricity demand from data centers through 2030. Long lead times for nuclear projects, policy uncertainty around renewables, and the sheer scale and speed of data center buildout mean the grid as a whole is being strained in ways that individual company commitments cannot fully address. The transmission and distribution infrastructure question, as noted above, remains separate from generation source questions entirely.
Major Tech Energy Commitments
What the leading AI companies have announced beyond the voluntary pledge
- Google: Nearly 2 GW of clean power via Xcel Energy deal; "bring your own power" model through direct purchase agreements; 150 MW geothermal contract
- Microsoft: $10 billion renewable deal with Brookfield Asset Management; 10.5 GW of renewables from 2026; first-ever fusion energy purchase agreement with Helion Energy
- Sector-wide: 10+ GW of nuclear contracts signed; 20+ GW of small modular reactors in planning; tech companies covered nearly half of global clean power added in 2025
- Limitation: Natural gas and coal still expected to supply 40%+ of additional data center demand through 2030; grid transmission costs remain socialized regardless
A Genuinely Bipartisan Issue
One of the more striking aspects of the AI energy cost debate is how thoroughly it has scrambled conventional political alignments. Opposition to data center expansion and energy cost socialization has become a genuinely bipartisan issue in ways that few technology policy questions have been in recent years.
Fortune reported in late 2025 on a "grassroots NIMBY revolt turning voters in Republican strongholds against the AI data center boom." CNBC noted that Bernie Sanders and Ron DeSantis, who agree on very little, have both emerged as leading skeptics of the data center expansion model. One community organizer described the dynamic as "not red versus blue, but people who live here versus people who want to industrialize where we live."
Community opposition has had measurable economic impact. Opposition to data center development rose 125% in the second quarter of 2025 alone, and an estimated $98 billion in data center projects have been blocked or delayed by community opposition. Some 142 activist groups across 24 states are organizing resistance. Cities and towns in at least 14 states have passed moratoriums on new data center development.
The political stakes became visible in electoral results. Two Democrats were elected to Georgia's utility commission in 2025 with data center electricity costs as a central campaign issue. The Trump administration's fundamental challenge is that its goal of maintaining U.S. AI dominance, which requires massive data center buildout, directly conflicts with the economic interests of its own voters who are experiencing electricity bill increases as a result.
Energy policy analysts identify the next uncapped PJM capacity auction, scheduled for June 2026, as a potential flashpoint for another significant price spike unless structural reforms are enacted before then. Consumer advocates and state regulators are preparing for that scenario, and the political pressure on both parties to act is likely to intensify regardless of how the voluntary pledge plays out.
What Nonprofit Leaders Should Consider Now
The Ratepayer Protection Pledge is not a reason to assume the electricity cost environment will stabilize on its own. Given the structural limitations experts have identified, and given the timeline for any legislative solutions to move through Congress and state legislatures, nonprofit organizations should treat this as an ongoing operational challenge rather than a resolved policy question.
This does not require nonprofit leaders to become energy policy experts. It does require that they understand their organization's energy cost exposure, explore the tools available to manage it, and engage in the policy conversations where their voice can make a difference.
Operational Steps
- Audit your organization's energy use by facility and identify your highest-cost loads
- Explore cooperative purchasing arrangements through nonprofit energy cooperatives or shared service organizations in your region
- Review your utility contracts for "Change in Law" or force majeure clauses that allow cost pass-throughs and understand when you have renegotiation opportunities
- Research green energy grant programs and efficiency financing that does not require upfront capital, including USDA Rural Development programs and LIHEAP-adjacent nonprofit resources
- Build energy cost variability into your budget assumptions for the next 12 to 24 months, rather than treating current rates as a stable baseline
Advocacy and Policy Engagement
- Contact your state representatives about data center cost-responsibility legislation, especially if your state has not yet acted
- Participate in state public utility commission proceedings on rate cases, where nonprofit voices are welcomed and can be influential
- Connect with your regional nonprofit association to coordinate advocacy, as collective voices carry more weight in regulatory proceedings than individual organization comments
- Document and share the impact of rising electricity costs on your programs and the communities you serve, as concrete impact stories are the most persuasive advocacy tool in utility proceedings
- Monitor the June 2026 PJM capacity auction outcome, as it will signal whether the current rate environment is likely to stabilize, worsen, or improve
For nonprofits that are themselves beginning to use AI tools in their operations, there is also a specific question worth considering: to what extent are your own AI tool subscriptions contributing to the energy demand driving up these costs? This is not a reason to avoid AI tools, which can provide real value as described in getting started with AI for nonprofits and our guide to AI strategic planning. But it is a reason to be thoughtful about energy efficiency in your technology choices, to favor providers with strong renewable energy commitments, and to understand that the policy conversations happening right now will shape the cost environment your organization operates in for years to come.
The Bigger Picture: AI's Infrastructure Moment
The Ratepayer Protection Pledge and the debates surrounding it reflect a larger inflection point in how society is choosing to pay for AI infrastructure. For decades, the costs of digital infrastructure, from broadband networks to cellular towers to cloud computing facilities, have been distributed through a mix of user fees, public subsidies, and cross-subsidies within utility rate structures. AI data centers represent a step-change in the scale of that infrastructure, and the existing distribution mechanisms are straining under the weight.
The question of who pays for AI's energy demands is ultimately a question about how the benefits of AI are distributed relative to its costs. If the economic gains from AI accrue primarily to technology companies and their investors, while the electricity cost increases flow to ratepayers across income levels, the distributional outcome is regressive. This is precisely the concern that has generated bipartisan political anger and accelerated legislative activity at the state level.
For the nonprofit sector, which exists in significant part to address economic inequity and serve communities with limited resources, this question has both operational and mission dimensions. The operational dimension, managing rising energy costs, is real and immediate. The mission dimension, whether the AI transition is happening in a way that is fair to the communities nonprofits serve, is equally important and often overlooked in discussions that focus narrowly on technology adoption.
Understanding both dimensions, and being prepared to act on both, is the kind of informed leadership that the moment calls for. The Ratepayer Protection Pledge may or may not deliver meaningful protection. What will definitely matter is whether nonprofit leaders, community organizations, and advocacy groups engage substantively in the policy processes that will determine the answer, and whether they do so before the costs become even more difficult to reverse.
Conclusion
The Ratepayer Protection Pledge is a real political signal that the electricity cost impacts of AI data centers have become impossible for Washington to ignore. The participation of every major hyperscaler indicates that the industry understands the reputational and regulatory risk of being seen as the cause of rising utility bills for ordinary Americans. But a voluntary pledge without enforcement mechanisms, without coverage of transmission and distribution costs, and without the legal authority to override state utility regulation is not the same as a solution.
The structural forces driving electricity costs higher, massive AI infrastructure investment, accelerating data center demand, grid transmission constraints, and utility rate cases that socialize infrastructure costs, are not resolved by a White House signing ceremony. The next PJM capacity auction in June 2026 will be a clearer indicator of where the rate environment is heading than any voluntary pledge.
Nonprofits have both a practical interest in how this resolves and a legitimate voice in shaping the outcome. The communities most affected by energy poverty are the communities nonprofits serve. The policy processes that will determine whether data center costs are fairly allocated are open to public participation at the state level. And the organizations that document, communicate, and advocate around this issue now will be better positioned to protect their operations and their communities as the AI energy reckoning continues.
For a broader look at how nonprofits can navigate AI adoption while managing its organizational implications, explore our resources on overcoming AI resistance in nonprofit teams and building internal AI champions. The technology decisions and the policy environment are both parts of the same picture, and leading organizations are attending to both.
Navigate AI's Challenges With Expert Guidance
From energy costs to technology adoption, the AI landscape is changing fast. One Hundred Nights helps nonprofits make smart, informed decisions about AI at every step.