Trillions of dollars of AI capital are chasing roughly 100 gigawatts (“GW”) of electric power, and the nation’s electric-sector regulators – the policymakers at the federal and state levels, as well as the administrators of the various RTOs/ISOs – have a critical role to play in helping them secure it.
In this piece we discuss the potential for a conduit between capital in search of incremental grid capacity and those best positioned to provide it: the companies, governments, and markets focused on unleashing the “demand-side revolution.” These actors are working to identify and utilize the latent capacity currently hiding in plain sight across the nation’s electric grid, capacity that can be “unlocked” through the deployment of demand-side resources (resources that reduce demand for power, shift it across time, or inject supply at the grid’s edge) and well-designed market signals.
The golden goose of demand-side resources – representing both the greatest source of potential capacity and the most difficult source to “unlock” – is the mass implementation of energy efficiency measures and distributed energy resources (“DERs”) at the household level. These measures and resources reduce both peak demand and overall energy use. State agencies, municipalities, and nonprofits already operate many programs that install (and offset the cost of) efficiency upgrades nationwide. But as a much-discussed September 2025 study by Rewiring America identified, we have yet to scratch the surface of the potential reductions in demand and energy use available through household efficiency upgrades and distributed resource deployment.
The Rewiring study presents the maximalist case: replacing resistance heaters and central air conditioning nation-wide would unlock 30 GW of capacity. Installing a home battery in every single-family home, and adding a 5kW solar array to those with suitable roofs, would unlock another 109 GW. The growth and maturation of markets for such behind-the-meter assets and efficiency upgrades would drive per-unit cost reductions on the order of 40%. If a data center developer paid for 50% of the upfront cost of such devices (making them cost-efficient for the purchasing consumer), through a sort of “bring-your-own-capacity” model, the cost would be about the same as building a natural gas-fired power plant from scratch.
In the face of this incredible potential, the question begs itself: what are the regulatory structures and price signals needed to capture this extraordinary promise?
Into the demand side
The numbers suggest that household efficiency upgrades could solve the data center power problem nearly by themselves. The problem is that legacy markets, regulations, and financing for electric power are not designed to (and at present, largely cannot) incentivize investment in such upgrades at scale.
Capital will reliably flow from buyers of efficiency-derived capacity to sellers only when sellers can reliably (i) certify their capacity contributions and (ii) sell them, whether that be in competitive markets or via direct contracts or procurements. Regulators must require the revision of market rules, accreditation methodologies, and RTO/ISO and utility tariffs to facilitate both changes. Below, we explain how they should begin.
The electric-sector issues that threaten to slow the AI boom are now well-known. In sum: it is not possible to develop electric generation and transmission capacity quickly enough, and at sufficient scale, to satisfy the projected increase in AI and other data demand for power in the short-term. (And if it were possible, it would be extraordinarily expensive).
So, where from here? As the limits to the electric sector’s ability to increase the supply of electric power become apparent, an explosion of creative thinking concerning the other side of the equation – the “demand side” – is underway.
For decades now, various means of “creating” incremental capacity by reducing or shaping demand have played a role in U.S. grid operations. Demand response programs were first implemented in the 1970s and now exist in nearly every state. Virtual power plants (“VPPs”), imagined by academics in the 1990s and later realized in practice beginning in 2008, expanded and elaborated upon what demand response programs had started, coordinating assets beyond load and accessing more revenue sources. Soon after, the Internet of Things and mass availability of smart meters enabled the rise of the contemporary “residential energy management” firm, which deploys smart devices, efficient appliances, and behind-the-meter assets, and enrolls customers in utility incentive and community generation programs to maximally reduce customers’ electricity bills.
New urgency—and new ideas
The AI boom’s potential effects on the electric power sector first became apparent to industry-watchers in 2023, when utilities and regulators began revising their demand forecasts up—way up. In 2022, the Federal Energy Regulatory Commission (“FERC”) forecasted 23 GW of demand growth nationwide in the next five years. By 2024, that number had ballooned to 128 GW. (This 105 GW increase is more than 20 GW greater than the total generating capacity in the CAISO footprint.)
Developers of data centers, and the tech behemoths that fund them, are willing to spend a lot of money to access power. Their appetites have fueled extraordinary investment in generation facilities of all types, but especially – in a post-OBBBA landscape where clean energy tax credits have effectively been abolished – natural gas- and even coal-fired generating capacity.
As the limits of the electric power sector’s ability to provide incremental capacity on the supply-side have become clearer, though, discussions of demand-side resources, formerly confined to wonky conferences and trade pubs, have jumped to national newspapers. Now (that is, as of June 2026), major leaps are already being made in deploying and scaling more established forms of demand-side resources—including, notably, the June 2 announcement by Google and Voltus of a three-year deal to establish and operate a 100-MW VPP in PJM, in the first commercial deployment of Voltus’s “Bring Your Own Capacity” product. The global VPP market is expected to grow between roughly 400% and 700% in the next ten years.
VPPs have established a foothold in federal-jurisdictional energy markets, and are poised for explosive growth as a sector, because federal regulators required (with FERC’s Order No. 2222 (2020)) that grid operators allow them to participate in those markets. Order No. 2222 also enabled energy efficiency resources to participate in those same markets, but the results have been lackluster.
Bringing energy efficiency and household-level DERs to the party
Regulators at both the state and federal levels, as well as the administrators of RTO/ISO markets, all will have a role to play in facilitating the growth of household-level energy efficiency resources and DERs. Below, we discuss just a few of the routes they might take to do so.
Utility “shared savings” legislation. State-regulated distribution utilities generate their profits by making capital investments in new or upgraded distribution infrastructure (e.g., power lines and substations). This model can put their interests at odds with their ratepayers, who would prefer that such investments be avoided or deferred where possible—including, for example, through the deployment of energy efficiency resources and distributed generation or storage capacity. The introduction of “shared savings” programs through state-level legislation can align the interests of utilities and ratepayers by giving utilities a cut of the savings created by such deployments. Take, for instance, the Brooklyn Queens Demand Management (“BQDM”) program, in which ConEd avoided a $1 billion substation upgrade by facilitating the deployment of DERs, energy efficiency measures, and demand response resources at substantially less cost. Savings created were shared by ConEd and the ratepayers; going forward, ConEd has proposed that it will collect 30% of net benefits achieved through BQDM and similar programs, while ratepayers will collect 70%. A successful utility shared savings program will account for, and compensate utilities and ratepayers alike for, the full stack of savings created. These include avoided or delayed distribution upgrades, yes, but they also include avoided or delayed energy, environmental, and capacity investments as well. The fuller the accounting, the greater the incentive for regulated utilities to pursue investments that qualify for shared savings creatively and aggressively.
Capacity market reform. We recently
wrote about the American Efficient, LLC case, decided by FERC in April 2026. The scale of the fraud undertaken by American Efficient – roughly half a billion dollars collected via the PJM and MISO capacity markets – reveals the pitfalls of a light-touch regulatory approach to energy efficiency resources. Market administrators cannot simply abdicate responsibility for verifying the capacity “contributed” by energy efficiency resources, but assessing those contributions fairly and accurately – and awarding market revenues accordingly – is a trickier game than the one those administrators have played historically. Compare energy efficiency resources to VPPs to understand why. When VPPs deploy their capacity to reduce demand, their contribution is straightforward to calculate because it is analogous to dispatching a generator: how much power is being contributed? And for how long a period? Energy efficiency resources, on the other hand, can shift MWs of demand temporally or eliminate them entirely—forever. Accounting for their capacity contributions therefore must involve projecting the counterfactual: what demand
would have been absent the installation of efficiency resources. This is exactly what we suggest market administrators in search of incremental capacity do. While FERC has jurisdiction to require RTOs/ISOs to revise the manner in which they administer their capacity markets, RTOs/ISOs are free to move forward themselves. They should do so through detailed, stakeholder-driven plans to require large loads seeking interconnection to the wholesale grid to bring their own capacity at least to a great extent through the leveraging of demand-side resources. Such reforms will require FERC-approved tariffs – a not-for-the-faint-of-heart, and potentially time-consuming process; we have precious little time, in fact, so haste is called for.
Other federal-jurisdictional tweaks. As discussed above, household-level energy efficiency measures and DER upgrades provide value by avoiding distribution and generation investments. They also provide value at the transmission level and during the interconnection process, by freeing up headroom and reducing the need for transmission maintenance and/or expansion. As is already done with respect to the costs and benefits incurred by other resource types, this value creation should be accounted and compensated. Thus, the role of RTOs/ISOs in projecting the counterfactual case (i.e., the case without household-level energy efficiency and DER investments) extends beyond the capacity market and into transmission and interconnection planning. And again, FERC can require such changes—but RTOs/ISOs are free to move forward themselves.
Movement at the state level. Debates are swirling, and bills circulating, around state capitols as to how to protect ratepayers from the impacts of large load additions. Some of those efforts have begun to bear fruit in 2026. This series of state-level initiatives may, in time and with some coordination, converge on a regulatory model linking data center demand growth directly to investment in household-level capacity: In April 2026, in Virginia, Governor Spanberger signed into law a bill (HB284/SB371) that creates a capacity reduction credit which utilities can offer to data center developers in exchange for a faster path to interconnection. Virginia’s utilities are due to propose frameworks for compliance by January 2027. The bill is part of a flurry of actions by Spanberger’s administration to reduce (or at least slow the increase of) energy costs across the state, which is home to Data Center Alley. In July 2026, in New Jersey, Governor Sherrill signed into law a bill that (in addition to creating a new ratepayer class and rate structure for data centers consuming 50 MW or more), also establishes a first-of-its-kind retail program – outside the PJM wholesale market – through which a large load offsets its capacity obligation by paying to reduce demand elsewhere on the system. Also in July 2026, in New York, Governor Hochul’s Executive Order No. 62 establishes a one-year moratorium on large data center development while the state’s Department of Public Service considers the creation of a fund, paid into by data center developers, that would support the procurement of new clean energy supply (including customer-sited DERs). Separately, the New York Public Service Commission (“NYPSC”)’s ongoing “Energize NY Development” proceeding is examining opportunities to modernize interconnection for large loads while protecting existing ratepayers from cost increases and hewing to the state’s climate laws. New York State Department of Public Service staff are expected to issue a white paper in early 2027 addressing these issues.
Hashing it out in the dockets Several proceedings now underway at FERC represent opportunities to facilitate investment in large-scale energy efficiency deployments. For example, take Docket Nos. AD24-11 (concerning rules around co-locating large loads with generation capacity) and EL26-63 (a complaint filed by the Maryland Office of People’s Counsel against PJM, alleging that data center development in Northern Virginia is unjustly driving up transmission-related costs in neighboring states). Both proceedings ask FERC to explain how, and how much, large loads (and even large loads co-located with generation) rely on the transmission system. Under longstanding beneficiary-pays principles, such reliance comes with a price tag—and the price tag associated with regional (or even local) transmission upgrades can be enormous. Data center developers, ordinary electricity consumers, and grid operators could all benefit from a more granular accounting of large loads’ impacts on the transmission system—and such an accounting will set the stage for increased investment in energy efficiency. Where exactly does the addition of a large load cause the need for additional transmission investment? What is the precise potential effect on proximal distribution and transmission lines? Accounting that features a stronger emphasis on electrical proximity will make it easier to propose, and quantify the effects of, geographically-targeted energy efficiency investments aimed at offsetting the impact of large loads on the transmission system. Then there’s Docket No. RM26-4, FERC’s large load interconnection rulemaking, in which stakeholders are debating whether and how large loads served by the transmission grid should be required to contribute to system reliability. Here, too, is an opportunity to advance arguments for energy efficiency investments. By freeing up headroom, energy efficiency deployments can avoid transmission system investments and make large load interconnection cheaper and faster—but this will only happen if regulators quantify those capacity contributions accurately and with an eye for their actual, local effects on system topography. June 2026 saw FERC take the unusual step of opening six new dockets, one for each FERC-jurisdictional RTO/ISO, ordering each to explain to FERC how its interconnection processes are just and reasonable with respect to large load additions (or, otherwise, to begin drafting tariff revisions). These dockets represent ideal venues to make more particularized versions of the above-described argument: that energy efficiency deployments can make large load interconnection cheaper and faster. Or consider the various Order No. 2222 compliance dockets, in which regional grid operators continue to develop their proposals for DER participation in wholesale markets. With the April 2026 American Efficient ruling, FERC ended a chapter of light-touch regulation of energy efficiency participation in capacity markets, but Order No. 2222 still requires that grid operators facilitate such participation. There is room for household-level energy resource advocates to begin writing the next chapter – in stakeholder participation processes, in comments on Order No. 2222 compliance filings, and even via a Section 206 filing – by encouraging RTOs/ISOs to facilitate meaningful and targeted participation in the federal-jurisdictional markets to mitigate the detrimental effects of skyrocketing load requirements. Where certainty goes, money will flow When frameworks emerge that will credibly and durably compensate the full value to the grid provided by the deployment of household-level energy efficiency measures and DERs, private sector investment in such deployment will follow—because the value is real, and difficult to come by any other way. As discussed, hyper-scalers are already brokering first-of-a-kind deals to organize and activate existing grid capacity in the form of VPPs. Why aren’t they making similar investments in the deployment of say, ten thousand home batteries or one hundred thousand energy-efficient appliances? . . . Because the mechanisms by which the grid capacity and dollars freed up by such investments will be recognized and compensated are not yet in place. Complexity is the primary obstacle. Grid operators will need to develop methods for estimating counterfactual demand scenarios (i.e., demand absent energy efficiency investments). As in the demand-response context, regulators will need to avoid and police the double-counting of benefits. Calculating the effect of thousands of energy-efficiency devices on system topography will be a challenge. And the actual targeted installation of such quantities of devices may require government- or utility-led coordination. When effective mechanisms for incentivizing investment are established, the infrastructure that already exists for deploying energy efficiency devices and DERs, operated by municipalities, states, nonprofits, and private firms, can be supercharged. This process will begin when regulators and market administrators signal serious interest in, and long-term commitment to, recognizing and realizing the full potential of the demand-side revolution.[PR
