Large Load Interconnection: Some Legal and Policy Issues[1]

Click here to download a PDF version of this article.

Introduction: The Challenge

Forecasts of growth in electricity consumption (“load”) across the United States are increasing remarkably.  In just over the last few years, those forecasts have doubled and tripled.[2]  Much of that growth is attributable to large loads, such as data centers, cryptocurrency mining operations, and other heavy-computing installations, which the President has declared vital to American national security and economic competitiveness.[3] Both large energy companies and load developers have made the same point, arguing that “interconnection of new large loads represents a foundational element of the United States’ economic prosperity and national security”[4] and that “continued development of data centers here in the United States [is] a strategic economic and national security imperative.”[5] But at the same time, many existing power generation facilities are retiring.  And the timeline to site, permit, and build transmission facilities has not appreciably shortened.  Nor have transmission and distribution costs decreased.  The developers of large load installations see an opportunity to maximize returns via speed to market and avoidance of all avoidable costs.  But resource inadequacy, regulatory processes, and sometimes-conflicting stakeholder objectives can all stand in the way of speedy connection of large load facilities.  Hence, it is no surprise that this has all created keen interest in how to address the forecasted supply-demand imbalance and has sharpened the financial and policy differences among various industry participants.  This paper will examine some of the drivers of the imbalance, some policy response options, legal issues, and proposed alleviators or solutions.

Drivers of the Forecasted Supply-Demand Imbalance

Electrical infrastructure design, development, and construction is time- and capital-intensive, making load forecasts a key tool for successful electrical grid and market planning.  The recent explosion in artificial intelligence use is fueling intense activity by heavy-computing companies like Google, Amazon, and others to push data center and computing projects as fast as possible.  The massive, forecasted increase in these types of projects has led load forecasters across the country to project load growth at rates that might have been, for most of the history of the electricity industry, nearly unthinkable.  In just a few years, load forecasts have doubled or tripled.  Examples of those increases include:

• The WECC’s 2024 Western Assessment of Resource Adequacy forecasts that load growth in 2025 to 2034 will increase at four times the growth rate seen in 2013-2022, and double the growth rate just recently forecasted in 2022;[6]
• ERCOT, a grid with about 103 GW of installed capacity currently, in 2024 increased its five-year load forecast by over 40 GW from its 2023 forecast;[7]
• MISO has recently projected a need to triple its annual rate of generation capacity additions in order to keep up with load growth for the next 20 years;[8]
• The Lawrence Berkely National Laboratories projected in 2024 that electricity use by data centers could increase by more than 700 percent from 2018 to 2028, from 76 TWh to 580 TWh, representing an increase from 1.9% of total U.S. electricity consumption in 2018 to 12% of U.S. electricity consumption in 2028;[9]
• PJM expects 40 GW of generation to retire between 2022 and 2030, while load is rapidly growing, leading PJM to “project a potential shortfall in generation supply by the end of this decade” (i.e., in five years).[10]

In sum, the demand for electricity across the United States, and especially in regions like northern Virginia,[11] Texas, and others that are particularly attractive to data center operators, is soaring.

But, meanwhile, operators of older, fossil-fueled power generation sources are increasingly retiring those resources.[12]  And new power generation projects are heavily weighted towards wind, solar, and battery installations.  In ERCOT, the grid operator reported in January 2025 that ninety-five percent of the MWs of new generation sources in the development queue through 2030 are wind, solar, or batteries.[13]  While that is probably good news from an air quality standpoint, it can be disconcerting from a resource reliability standpoint, since those fuel sources (sun and wind) are not dispatchable by human  command.  In some circumstances, batteries also make up a large percentage of the new generation in the development queue.  But batteries, while dispatchable, are short-duration resources and do not, of themselves, actually generate any electricity—rather, they simply hold electricity generated at an earlier time by some other source. 

The timeline for designing, studying, funding, developing, building, and interconnecting new generation sources can consume one or two years.  The timelines for retirements of existing fossil-fueled resources can be shorter.  The timelines for designing, studying, permitting, and building transmission facilities can be much longer—three to five years in ERCOT and much longer (6 to 13 years) in FERC-jurisdictional regions.[14]  But the data center developers and hyper-scalers are pushing to get their facilities in operation much faster: in six to twelve months.  Accordingly, the potential for imbalance between electricity supply and demand can be accelerating like a 100-meter sprinter running downhill with a tailwind.[15] That type of confluence of factors could be part of the reason at least one grid operator is concerned about the ability of the grid to serve all the load reliably in the near term: “I don’t have a positive sense that we have enough generation on the books to serve the load that’s expected.”[16]

Moreover, these types of challenges are prompting data center owners, hyper-scalers, and other large load developers to seek solutions that serve their own project’s financial and electrical interests in ways that others think might degrade the grid or shift costs.  Notably, these loads have explored various means of obtaining service directly from an existing or new generation source without using, or minimally using, the transmission grid—a configuration loosely and broadly described as “behind the meter” service. 

“Behind the meter” service can benefit the large load installation by providing reliable power directly from an adjacent source at a competitive price and without need of the time and expense of grid upgrades to serve that load. The “behind the meter” load might also financially benefit from avoiding transmission grid-related costs such as transmission charges and ancillary service charges.[17]  These benefits can accrue to American society as a whole to the extent that such heavy-computing installations are vital to national security as the artificial intelligence race heats up and drives international geopolitical and military competition.

On the other hand, some stakeholders believe that a proliferation of behind-the-meter load will degrade reliability and increase all other customers’ costs.  This belief springs from the concern that large behind-the-meter loads will effectively take the associated generation “off the grid” and thus potentially exacerbate resource adequacy concerns by eliminating that load-adjacent generation from the pool of resources that can serve the grid generally.[18]  Moreover, critics assail the “disappearance” of the generation and associated behind-the-meter load from the grid operators’ vision and control, which they assert might make it harder for the grid operator to anticipate, see, and respond to varying grid conditions, thus potentially jeopardizing reliable continuous operation.[19]

Further, critics of behind-the-meter configurations argue that if such large load operators can avoid paying transmission costs or ancillary services costs, that will just increase the costs to be borne by other grid-using customers, since transmission expansion and other services are inevitable and must be paid for by someone. 

Thus, the electric power system and its stakeholders face abundant physical, financial, and legal challenges, many of which stem from the proliferation of demand for more interconnection of large loads like data centers.  While much attention has been focused on “behind-the-meter” configurations or “co-located” load, the issues are bigger, and in a sense might be boiled down to simply resource adequacy in a world of rapidly exploding large-scale load development.[20]  Unsurprisingly, this coming supply-demand imbalance and its significant financial implications have inflamed controversies and spawned contentious debates and contests at state, regional, and federal regulatory, legislative, and self-regulating bodies—including  the Federal Energy Regulatory Commission, state legislatures, state utility commissions, regional transmission organizations, and independent system operators. 

Of particular note with respect to FERC-jurisdictional markets, the issues are thoroughly addressed in the FERC contested case proceeding styled as PJM Interconnection, LLC, FERC Docket Nos. ER24-2172-000 and ER24-2172-001 (the “Susquehanna” proceeding) and the FERC technical conference proceeding styled as Large Loads Co-Located at Generating Facilities, FERC Docket No. AD24-11-000 (the “Co-Location Technical Conference”).

And, of course, due to the competing viewpoints, incentives, perspectives, arguments, and timelines, policy makers at all those bodies struggle to fully absorb, thoroughly analyze, and reconcile all those competing forces in order to adopt policy outcomes that ably serve the pressing timelines. 

The net result:  significant regulatory uncertainty at the very time that large load developers are pursuing speed to market.[21] 

But humans are imaginative and ever-incented to succeed. Consequently, stakeholders continue to generate and debate many policy options and legal issues, all of which could lead to proposals to solve the myriad issues. 

Policy Response Options

In various federal and state regulatory and legislative proceedings, as well as in ongoing stakeholder discussions at regional transmission organizations, interested companies, regulators, citizens, and policy makers are suggesting a host of policy responses to deal with the challenges posed by large load interconnections.  These policy responses feature a variety of approaches, ranging from constraining the physical options of large load developers/owners, to imposing certain financial constraints on those developers/owners, to mandating particular actions by regulators or grid operators. 

In the category of constraining physical options of large load developers or owners, some suggested responses to the challenge include: flatly prohibiting “behind the meter” configurations; or requiring any data center or large load developer to “bring your own generation” as part of the total project.[22]

The genre of imposing financial constraints on large loads includes: requiring an up-front financial commitment as a condition of even joining the interconnection queue; or requiring the large load owner to pay all or some significant portion of transmission or other grid costs, so that there is no financial “windfall” from a behind-the-meter configuration.

And with respect to directing regulator behavior, some suggestions include: mandates to speed up the processing of the interconnection queue for generation and load; permitting bespoke behind-the-meter arrangements and addressing policy considerations on-the-fly as such arrangements proliferate; and affirmatively promoting behind-the-meter configurations for large loads as a strategy to attract more investment in electricity-consuming and electricity-grid projects.

Each of these suggestions has its own motivating rationale and concomitant impacts, which are discussed in more depth in the last section of this paper. 

Legal Issues

As the discussion progresses regarding large loads and the potential for a coming imbalance between generation and load across the United States, a few pertinent legal issues have come to the fore.  Those legal issues principally revolve around: the scope of federal authority versus state and local authority; the rationales for and legal bases for various ratemaking/cost assignment regimes; governmental authority to impose conditions on large load development/investment; state or transmission operator authority over loads; and consistency of load treatment (especially in the context of cost allocation) with existing regulatory precedent.

With respect to the scope of federal authority versus state regulatory authority, perhaps the most visible fora for discussion for FERC-jurisdictional markets recently have been the Susquehanna Interconnection Services Agreement (ISA) proceeding at FERC and that agency’s ongoing Co-Location Technical Conference aimed at exploring issues pertaining to large load interconnection. 

In Susquehanna, FERC was asked to approve an amended ISA that would expand the amount of behind-the-meter consumption at a nuclear facility.  The amended ISA was supported by not only the plant operator and the load, but also by PJM.  FERC rejected the application, however, on the narrow legal ground that the applicants had not demonstrated that the provisions of the amended ISA that deviate from the pro forma ISA were “necessary for any interest unique to the interconnection of the Susquehanna Customer Facility.” FERC reasoned that because some of those provisions relied heavily on a generally applicable PJM guidance document, that raised questions regarding whether PJM intended to apply those provisions to other similarly-situated interconnection customers, therefore making those provisions insufficiently “necessary for any interest unique to” the Susquehanna configuration.[23] Interestingly, in his dissent to that FERC decision, Chairman Phillips appears to imply that the Commission might have rejected the particular application because it was thinking about co-located load issues on a broader scale in the simultaneous Co-Location Technical Conference.  Whether that is true or not, the proceedings raise overlapping and intriguing legal and policy issues.

Some key legal questions raised in Susquehanna and the Co-Location Technical Conference include:

• Is behind-the-meter load outside of FERC’s jurisdiction (because it does not involve the transmission of electric energy in interstate commerce or the sale of electric energy at wholesale in interstate commerce)?[24]
• Does behind-the-meter load even use any FERC-jurisdictional services?[25]
• Is the behind-the-meter load effectively subject to FERC jurisdiction because the generation serving the load is making wholesale sales in interstate commerce via the grid, in at least some manner, or is making use of FERC-jurisdictional transmission service?[26]
• Does FERC’s grant of authority under the Federal Power Act include the authority to consider resource adequacy issues when evaluating interconnection agreements?[27]
• What does FERC’s rejection of the bespoke behind-the-meter interconnection agreement in Susquehanna really mean and how far-reaching are its impacts?[28]

With respect to ratemaking and cost assignment, some pertinent legal questions include:

• Would it be just and reasonable to use a ratemaking approach that departs from historical precedent?
• Can specialized ratemaking approaches be applied to large loads and remain non-discriminatory?[29]
• What portions, if any, of the electricity service to a behind-the-meter load are appropriately the subject of FERC-jurisdictional rates?[30]

Interesting legal questions in the category of governmental authority to impose conditions on large load development or investment include:

• Can a regulatory agency or RTO require a large load developer to “bring your own generation” as a condition to joining the interconnection queue?
• Can a regulatory agency or RTO require a large load developer to make a certain financial commitment as a condition to joining the interconnection queue, to bolster certainty that the load will truly become reality?
• Can a regulatory agency or RTO impose a requirement for certain transmission cost funding by a large load as a condition of approving interconnection?
• Can a regulatory agency or RTO flatly prohibit behind-the-meter configurations for large loads?[31]

With respect to questions of state authority over large loads, some interesting issues include:

• If the large load is a retail consumer of power, can that load truly disappear from state authority or from the local load-serving-entity’s responsibility?[32]
• Can a state lawfully require a large load developer to “bring your own generation” as a condition to joining the interconnection queue?
• Can a state preclude a nuclear-powered generation facility (or any other facility) from entering into a behind-the-meter arrangement?

And with respect to the potential for regulatory precedent to be binding or to be subject to flexibility, some interesting questions include:

• Is FERC’s denial of the revised Susquehanna agreement consistent with its approval of prior iterations of the agreement for behind-the-meter load at that same site involving the same parties?
• Has FERC charted a new course for reviewing and approving interconnection agreements by rejecting the latest Susquehanna agreement during (and perhaps in consideration of) broad discussions across the nation about potentially similar arrangements?
• Must any ratemaking or cost allocation decisions regarding behind-the-meter load be based solely on the “net” load of the behind-the-meter facility?[33]
• Is it consistent with regulatory precedent to treat fully isolated behind-the-meter load as if it is indirectly consuming some grid services?

Proposals to Solve the Issues

People trying to solve one or more of the issues raised by the large load impact on supply-demand imbalance have proposed varied remedies. As in most aspects of life, these proposals create tradeoffs that reflect a policy perspective or assignment of values to different policy outcomes.  And they often raise legal issues.  Some of the proposals getting significant attention include the following, each of which is discussed in more detail in this last section of this paper, with discussions of some possibilities of each being adopted:

• Forbid all behind-the-meter configurations;
• Forbid behind-the-meter configurations at some types of plants (e.g., nuclear);
• Declare that all behind-the-meter loads actually use some grid services, even if indirectly because of the grid connectedness of the dedicated generation;
• Require large loads to “bring your own generation” as a condition to interconnection;
• Give large loads that “bring their own generation” a preference in the interconnection queue;
• Require large loads to demonstrate a financial commitment or require payment of a significant fee as a condition to joining the interconnection queue, to promote more certainty in load forecasting;
• Accelerate large load interconnection and figure out the broader market-affecting policy considerations on the fly; and
• Promote co-location/behind-the-meter configurations as a means of attracting more investment in grid projects.

Forbid All Behind-the-Meter Configurations

Some stakeholders argue that behind-the-meter arrangements should be flatly prohibited.[34]  These are often proponents of the notion that all load should benefit from all grid resources or proponents of the notion that behind-the-meter service degrades reliability by “hiding” the load from the grid operator.[35]  A fundamental rationale for this suggested prohibition is that generation resources should be available to serve all load and a new large load should not be permitted to essentially remove a generation resource from the grid by claiming its entire output for use solely by that large load.[36]  Doing so, they argue, deprives other loads of the benefit that the generation resource should provide to the grid as a whole.[37]  Further, proponents of this ban on behind-the-meter configurations argue that such configurations degrade overall system reliability, to the detriment of all customers, by “hiding” the load (and its associated generation) from the grid operator, lessening the grid operator’s ability to monitor and account for the load’s and generation’s behavior in day-to-day grid operation.  They argue that if the load suddenly drops, then generation surplus will instantly spike, to the surprise of grid operators, causing detrimental instantaneous responses.  Similarly, they argue that if the generator were to trip, the excess load would instantaneously spike, potentially requiring immediate action to keep the grid in balance if the behind-the-meter load does not also instantaneously trip or does not provide its own instantaneous backup. [38] 

Of course, opponents to a ban on behind-the-meter configurations can respond with freedom-to-contract arguments and technical arguments regarding particular mechanisms designed to mitigate or eliminate the reliability risks.  Prohibiting behind-the-meter arrangements could be tantamount to barring market participants from lawfully contracting among themselves in ways that best serve the contracting parties’ financial interests.  Such a prohibition would fly in the face of economic freedom and could impose undue regulation that might amount to a regulatory taking of private property, purportedly for the common good, requiring just compensation for the taking. 

Additionally, as demonstrated by the detailed mechanisms built into the Susquehanna ISA, the contracting parties can design protocols and physical arrangements that would mitigate or eliminate most or all of the reliability concerns raised by proponents of a ban on behind-the-meter configurations.  Thus, the parties to a behind-the-meter arrangement can argue that the configuration serves private parties’ economic interests in an open market and does not pose significant reliability concerns, as the parties to the Susquehanna ISA argued.

Further, behind-the-meter configurations are somewhat commonplace today.  For example, industrial facilities can install their own generation, sometimes in the form of a generator-steam host arrangement. Would that suddenly be prohibited? As another example, in the rapidly developing oilfields, producers will sometimes bring their own on-site generation to ensure they have sufficient power in the interim before a transmission and distribution company can connect to them.[39]  Why would that be prohibited?

Opponents of a ban on behind-the-meter configurations also adamantly argue that such a ban would do nothing to improve resource adequacy and could actually backfire.  One major energy company has argued that:

there is no meaningful resource adequacy difference between loads interconnected behind-the-meter and in-front-of-the-meter. Those who assert that behind-the-meter co-location is taking megawatts off the grid are merely engaging in a thinly-veiled plea to slow down load growth so the markets can address resource adequacy concerns first. That is not only an unlawful basis for denying an interconnection, it is also a recipe for undermining resource adequacy and harming the national interest.[40]

On balance, a total ban on behind-the-meter configurations could be seen as a rather drastic step that tramples upon market participants’ freedom to contract without fully evaluating all potential mitigating strategies or policies that could apply to a particular situation.

Forbid Behind-the-Meter Configurations at Certain Facilities

This suggestion, to prohibit behind-the-meter configurations at certain types of facilities—e.g., nuclear power plants—would be designed to preserve the benefits of that type of plant for the grid as a whole and would seek to avoid reliability or operational complications that might affect operations of that facility.  This type of approach would be more narrowly tailored than a blanket ban on behind-the-meter configurations and thus would probably have a lesser impact on large load development opportunities.  But it is also possible that, if such a ban were applied to nuclear-fueled power plants, it might hit the precise sweet-spot that many data center developer or hyper-scalers are targeting, since nuclear power fits into the no-emissions category of “green” generation that many companies find particularly desirable.  Consequently, this approach would likely trigger the same objections regarding freedom of contracting, regulatory takings, and ignoring technical problem-alleviating solutions that would each ostensibly apply to a blanket ban on behind-the-meter configurations. It also might thwart the drive for advanced nuclear facility development in places like Texas.[41]  For those reasons, a ban on behind-the-meter configurations at nuclear-fueled facilities could, like a blanket ban, seem to go farther than is necessary without fully evaluating case- and site-specific mitigating protocols.

Declare that all Behind-the-Meter Load Actually Uses Some Grid Services

One of the fundamental economic arguments supporting the potential use of behind-the-meter configurations is that by locating wholly behind the meter, the load is totally isolated from, and uses no services provided by, the grid at large.  In other words, the argument is that such load uses only the generation provided by the adjacent power plant and does not use any services provided by the grid that exists beyond the direct connection between the power plant and the co-located load.  And, goes the argument, that co-located load should thus bear no costs that are typically imposed on grid-connected loads because the behind-the-meter load is receiving no grid-related service. 

In light of that argument, one potential means of thwarting at least part of the economic rationale that could lead to proliferation of behind-the-meter arrangements for large loads would be to declare, as a matter of federal or state regulatory policy, that all behind-the-meter load is deemed to be using grid services.  The Independent Market Monitor for PJM has asserted this is true—all load benefits from grid services.[42]  That would lead to behind-the-meter load owners bearing some amount of transmission costs, ancillary services costs, or other costs, thus mitigating or reducing the argument that behind-the-meter loads are free-loading or raising the costs of other customers by avoiding their “fair share” of grid costs.[43] 

This is an approach that presents a classic regulatory scenario—multiple stakeholders with varying economic interests presenting their legal, technical, and policy arguments about how a cost-allocation formula might or might not reflect physical realities of power flows and customer benefits. 

At this point in the debate, it seems at the federal level that policy makers are not yet ready to decide that behind-the-meter load does not use grid services.  Thus, it would appear that the de facto decision to date is that behind-the-meter loads benefit in at least some way from grid services. 

A robust debate in federal and state ratemaking proceedings seems inevitable and worthwhile, and the status quo seems to be that co-located load does, indeed, use or benefit from at least some grid services.  Thus, this approach would seem to be minimally disruptive to the status quo unless the word “some” is functionally interpreted to impose significant, project-altering costs on large load owners. The only question, then, under this approach would be deciding what amount of grid-related costs are truly the “fair share” of the co-located load.[44]

 Require Large Loads to “Bring Your Own Generation” to Interconnect

As noted above, some critics of large load interconnections, especially behind-the-meter interconnections, argue that large loads inappropriately deprive other customers of massive amounts of generation, thus imperiling reliability. A suggested approach to alleviating that objection is to require large load developers to simultaneously build associated generation, so that the net load impact is effectively zero, and resource adequacy is not degraded.  From a policy perspective, agnostic of costs imposed on the load, this approach, of course, would make perfect sense—if every load were to bring its own generation, then resource adequacy would never suffer.  But the policy and legal implications of imposing such a requirement on large load developers as a condition of interconnection are huge.

As a starting point, the notion of requiring a load developer to also build generation as a condition of getting electricity service would be a radical departure from a basic premise of the electricity industry construct in the United States.  That basic premise is that users of electricity can generally be considered to have a right to electricity service as a basic feature of the regulatory compact between the government and regulated (often “franchised”) electricity providers.  Granted, that regulatory compact has been altered in significant respects by the adoption of wholesale competition and, in some states, retail competition.  But the fundamental premise remains—American policy makers generally expect that electricity service will be provided to loads, without need of the load owners building their own power plant (but, of course, some customers can be required to make a contribution-in-aid-of-construction to receive service).  If a regulatory or legislative body were to instead prescribe that loads over a certain size cannot connect to the grid and expect service unless they simultaneously build their own generation, that would signal a material shift in the role to be played by regulated utilities and quasi-regulated power plant owners. 

Accomplishing that shift would likely require enabling legislation.  And the legislative debate would undoubtedly be robust, as data center developers, hyper-scalers, and other large load developers might query why they would be singled out to be required to bring their own generation, when such a requirement has not historically been imposed on any other load developers.  Would such a financial burden be tolerated? Or would it drive data centers and hyper-scaler facilities overseas, thus potentially thwarting America’s geopolitical and defense interests? 

The notion of bringing one’s own generation to support one’s load is certainly achievable as a voluntary matter and might makes sense to some developers, especially in situations where the generation output might also be sold on the market at a profit.  In fact, Google appears to have committed to that approach:

For Google, the promise of co-location rests on matching new load with new generation. This “additionality” component is core to Google’s DNA and a key part of our 24/7 carbon-free energy objective, which requires all new load be served with new carbon-free energy.[45]

But that voluntary approach is a distinct matter from making such generation development a requirement for interconnection.  While the policy goal of not degrading resource adequacy would be laudable, the international competition impacts and free-market distortions that might attend such a requirement could be material. Would the benefits of such a policy outweigh those impacts and justify a marked departure from one of the basic features of American electricity policy? That is a debate that will surely be hotly waged if this type of proposal gains serious traction.

Give Large Loads a Preference in the Interconnection Queue if they Bring Their Own Generation

A potentially less controversial means of encouraging large load developers to bring their own generation would be to establish a quid pro quo:  If you commit to building generation to serve your own large load, you will receive preferential treatment in the interconnection queue.[46]  This would address a key large load objective: speed to market.  But it would come a at cost—to get that speed, you must invest in power generation to serve your own load. 

The American legal system is replete with these types of arrangements, in which a market participant receives some type of benefit for committing to take on some financial obligation.  For example, the federal production tax credits and investment tax credits are built upon the notion that if an investor brings a certain type of facility to the market, that investor will be rewarded by the government in the form of financial credits.  In the electricity industry specifically, the State of Texas recently adopted a form of this construct by creating the Texas Energy Fund, which will provide low-cost loans and completion bonuses to companies who commit to, and follow through on, building dispatchable generation that will ostensibly benefit citizens throughout the state. 

Thus, the notion of a quid pro quo from the government in exchange for a public-benefitting investment by a private company is not unique.  And, consequently, it would probably not be overly controversial, unless the benefit conferred on the company were to be viewed as disproportionate to the sacrifice made by the company to invest in the generation.  While there are inevitably market distortions that arise from governmental incentive applied to competitive market behavior, in this context the prospect of getting speedier connection if the project shows that it will not degrade, and might improve, resource adequacy seems to be an elegant way to address the potentially impending supply-demand imbalance. 

Interestingly, Google expressly advocates for fast-tracking projects that include generation and co-located load, arguing that “new resources that can demonstrate they are backed by committed co-located loads should be fast-tracked for connecting to the grid.”[47]

Require Large Loads to Demonstrate Financial Commitment or Pay a Material Fee to be Included in the Interconnection Queue

As noted at the outset of this paper, load forecasters are projecting massive increases in electricity consumption in the coming years. That has even prompted some grid operators to express concern about reliability if all that load really shows up.  To assist load forecasters with the uncertainty about whether the load will or won’t really develop, some have suggested requiring large load developers to provide evidence of a financial commitment to their project, or to pay a material fee, as a condition to being included in the interconnection queue (i.e., as a condition to being seriously considered in the load forecast). 

As an initial matter, this proposal raises questions about discriminatory treatment of some load developers versus others.  What would be the driving rationale for imposing this financial burden on some developers but not others?  Is it merely the magnitude of the load?  If so, how would the line be drawn between those who must ante up and those who do not?  While legislative bodies have broad discretion to make such judgments and impose such requirements, regulatory agencies often have an obligation to treat similarly situated entities similarly.  Consequently, there would likely be vigorous debates and objections to regulatory bodies trying to impose these types of measures, which might be inherently arbitrary due to being dependent on drawing a line that might not be unassailably grounded in physical impacts to the grid. 

Additionally, while the American legal system commonly tolerates a quid pro quo arrangement in which a private entity can be awarded a special benefit in exchange for making a special investment or commitment, there might be questions about whether that legal system would similarly support a requirement to make a special commitment in order to simply gain entry to the system, as opposed to being granted a special benefit. 

If the financial commitment is not overly burdensome, then it might be reasonable to expect that developers will not vociferously object.  But if the financial commitment is not actually burdensome, will it have the desired effect?  Will it weed out the non-serious projects if it doesn’t carry some pain?  These questions will accompany any proposal to require a special class of load developers to bear a special type of financial obligation in order to be treated like all other load developers. The inherent susceptibility to legal and policy challenges might imperil this type of proposal from being widely used.

However, it is definitely worth noting that Google expressly supports this type of approach to requiring some “skin in the game” for large load developers:

Requiring load adjustments due to new large load interconnections to be “commitment-backed” through material, up-front financial commitments is a proactive approach to load forecasting that standardizes the processes for verifying large load additions before inclusion in RTO load forecasts, thereby protecting ratepayers from overbuilding the system on the basis of duplicative or speculative load requests. These up-front financial commitments are akin to “readiness deposits” that FERC recently mandated in Order No. 2023 for all generator interconnection processes. It creates “skin in the game” by requiring sponsors of projects with a large minimum peak demand to post a material financial commitment at the time they apply for interconnection to an EDC system. This effectively mitigates the equally significant risks of over development of infrastructure and potential stranded costs and under investment in critical infrastructure to accommodate these important sources of economic growth. Material up-front financial commitments can reduce speculative or duplicative load requests that may be adding uncertainty to load forecast estimates.

We believe this approach should be standard for all new large load interconnection requests and have been working with our state utility commission and utility partners to adopt up-front financial commitments in regions where we operate.[48]

Accelerate Large Load Interconnections Now and Address Policy Questions Going Forward

In light of the benefit that large loads like data centers and hyper-scale facilities can provide to American competitiveness, some proponents have suggested that the right path forward is to encourage and accelerate interconnection of such loads and work on the policy considerations in flight.[49] This is presented as a reasonable path forward because it is important to get more development going, not less, and it will be impossible or at least impractical to try to get all the policy calls just right before moving ahead. In sum, this approach seems founded on the notion of not letting the perfect be the enemy of the good.  Moreover, large loads, even behind-the-meter large loads, are already part of the American electricity system. Thus, an argument can be made that interconnecting those loads has not broken anything and the benefits of making forward progress now outweigh the detriments that would come from stalling while seeking optimal policy outcomes. 

In a sense, this is a significant rebuttal to the Susquehanna outcome. That decision can be viewed, in context, as a pause on expansion of large load interconnection, at least in behind-the-meter configurations, while policy makers wrestle with the increased attention focused on such loads and the resource adequacy and ratemaking fairness concerns being raised in response to that load growth.  If it is legitimate to view large computing loads as vital to American competitiveness and potentially to American national security,[50] then might it also be legitimate to question whether protracted policy debates and fights over who pays for what are appropriate while the rest of the world moves ahead?  And if the grid today already has such installations without anyone being materially harmed, it might very well be appropriate to question why policy makers would want to push “pause.” 

It will be interesting and enlightening to see if federal and state regulatory bodies begin to feel some pressure from legislators and other elected leaders to move the ball forward, even without perfect answers to all the policy questions.  That tack will probably serve America better then letting these matters become balled up in litigated proceedings.  Proceedings like FERC’s technical conference likely serve a useful purpose by helping elevate public and political attention on these issues.

Given the significant financial interests at play, the attention from the White House, the national trade press coverage, and other factors, it is reasonable to predict that a regulatory stalemate will not be sustainable. Thus, a path of moving large load interconnections forward while simultaneously providing fora for resolving key policy, ratemaking, and technical considerations seems quite likely.

Promote Co-Location of Load and Generation as a Means of Attracting More Investment

This notion of affirmatively promoting co-location or behind-the-meter configurations as part of a deliberate plan to attract more investment in grid facilities by an ever-expanding group of players could be a subset or corollary to the path just discussed. That is, as part of moving the ball forward, policy makers could seek to affirmatively encourage more Googles and Amazons to invest in even more electricity grid projects to keep America competitive as regulators continue to work on the associated policy and ratemaking issues.[51]  Given the high financial stakes involved in ensuring sufficient electricity to power all our devices, homes, cars, personal computers, and AI machines, this approach will probably naturally follow, as a complement to a strategy of moving forward while refining the policies and adapting on the fly.

[1] Author: Bill Moore, attorney, Butler Snow LLP.  The views, analyses, and interpretations presented herein are attributable solely to the author personally, and must not be represented as the views, analyses, or interpretations of Butler Snow LLP or any of its clients.

[1] “Virginia Legislators Introduce Bills to Deal with Data Center Growth“, RTO Insider (Jan. 14, 2025) (reporting that by one observer’s estimation, “unconstrained growth of data centers could lead to peak demand of 60 GW by 2050, which is nearly three times its current peak of 22 GW”).

[2] “Executive Order on Advancing United States Leadership in Artificial Intelligence Infrastructure.“ The White House (Jan. 14, 2025) (hereafter “AI Executive Order”) (“Recent advancements in AI demonstrate its rapidly growing relevance to national security . . . .  Advances at the frontier of AI will also have significant implications for United States economic competitiveness.  These imperatives require building AI infrastructure in the United States on the time frame needed to ensure United States leadership over competitors . . . .”).

[3] Post-Conference Comments of Vistra Corp., Large Loads Co-Located at Generating Facilities, FERC Docket No. AD24-11-000 (the “Co-Location Technical Conference”), at 5 (Dec. 9, 2024).

[4] Post-Technical Conference Comments of Google LLC, Co-Location Technical Conference, at 1 (Dec. 10, 2024).

[5] Western Assessment of Resource Adequacy, WECC, available at https://feature.wecc.org/wara/  (“Current resource plans forecast staggering demand growth over the next decade. For the Western Interconnection, annual demand is forecast to grow 20.4%, from 942 TWh in 2025, to 1,134 TWh in 2034. That growth rate is more than double the 9.6% growth forecast in resource plans filed in 2022, and over four times the historical growth rate of 4.5% between 2013 and 2022. . . . . A major driver for the increase in demand over the next 10 years is the expansion of large loads like data centers, manufacturing facilities, and cryptocurrency mining operations.”).

[6] “ERCOT Enters New Era of Planning to Meet Future Economic Growth” ERCOT Press Release (Apr. 23, 2024) (“With an estimated additional 40,000 MW of load growth by 2030 as compared to last year’s forecast, the focus will ensure efficient and innovative approaches in planning are leveraged to prepare the ERCOT market for this accelerated growth.”).

[7] “MISO Assessment Calls for 17 GW in New Resources Annually” RTO Insider (Dec. 18, 2024) (MISO’s Armando Figueroa Acevedo said a 17 GW/year rate would require members to add more than three times their recent average additions of 4.7 GW/year.”).

[8] “2024 United States Data Center Energy Usage Report,” Lawrence Berkely National Laboratory (Dec. 2024), available at: https://eta-publications.lbl.gov/sites/default/files/2024-12/lbnl-2024-united-states-data-center-energy-usage-report.pdf.

[9] Statement of Aftab Khan on Behalf of PJM Interconnection, L.L.C., Resource Adequacy and Expected Load Growth Reliability Technical Conference, FERC Docket No. AD24-10-000 (Oct. 16, 2024).

[10] The Associated Press reported in December 2024 that electricity consumption by data centers in Virginia could double from 2023 to 2030, consuming 46% of the total electricity generated in that state in 2030.

[11] “MISO Assessment Calls for 17 GW in New Resources Annually “RTO Insider (Dec. 18, 2024) (“Between 2029 and 2043, MISO expects 27 GW in thermal retirements and 11 GW in thermal additions, leading to a net loss of 16 GW.”).

[12] ERCOT Monthly Operational Overview for January 2025 (Feb. 15, 2024), available at: https://www.ercot.com/files/docs/2024/02/16/ERCOT-Monthly-Operational-Overview-January-2024.pdf.

[13] ERCOT CEO Board Update (Apr. 23, 2024) available at: https://www.ercot.com/files/docs/2024/04/22/5%20CEO%20Update.pdf.

[14] A “key takeaway” presented by ERCOT CEO Pablo Vegas to the ERCOT Board of Directors in April, 2024 is: “The forecasted pace of load growth could exceed the pace at which transmission capacity can be built to support it.” ERCOT CEO Board Update (Apr. 23, 2024) available at: https://www.ercot.com/files/docs/2024/04/22/5%20CEO%20Update.pdf.

[15] “ERCOT Faces Uphill Battle to Meet Large Loads,” RTO Insider (Jan. 6, 2025) (quoting Woody Rickerson, Chief Operating Officer at the Electric Reliability Council of Texas).

[16] The PJM IMM asserts that avoiding these costs and regulatory processes is “the core feature” of the “behind-the-meter” configuration:  “The core feature of the co-located approach is avoiding the costs associated with both state and federal regulation and avoiding the PJM planning process.” Post-Technical Conference Comments of the Independent Market Monitor for PJM, FERC Docket No. AD24-11-000 (the “Co-Location Technical Conference”), at 5 (Dec. 16, 2024).  But Google argues that avoiding those costs is absolutely not the driving rationale—rather, the driving rationale is simply seeking to obtain power in resource-constrained electricity markets as quickly as the loads are becoming “live”: “Thus, co-location has emerged as a rational response to the current market insufficiency: new generation cannot be interconnected quickly enough to match the pace of new load additions. Put differently, co-location is driven by a need to secure sufficient resources to ensure our timely growth, not a desire to avoid infrastructure costs.”  Post-Technical Comments of Google, LLC, Co-Location Technical Conference, at 3 (Dec. 10, 2024).

[17] The PJM IMM has raised this concern: “Given that PJM’s excess reserves are currently less than 1,000 MW, removing a relatively small amount of MW to serve co-located load would have a significant impact on reliability and could easily result in a shortfall in reserves in PJM.” Post-Technical Conference Comments of PJM IMM, Co-Location Technical Conference, at 3.

[18] Exelon has, for example, raised this concern: “[T]he interconnection of load not regulated, or even fully visible, to the nation’s utilities and grid operators pose[s] real security challenges, not only issues of economic fairness.” Joint Public Service Parties’ Post-Conference Comments, Co-Location Technical Conference, at 2 (Dec. 9, 2024).

[19] The Data Center Coalition has made this point: “Even without co-location, data centers would still require access to power, and the broader grid challenges would persist.  As such, it is crucial that we focus on addressing these larger reliability and interconnection issues which are a symptom of broader grid constraints.” Post-Technical Conference Comments of the Data Center Coalition, Co-Location Technical Conference, at 4 (Dec. 9, 2024). Similarly, Exelon has argued that the real concern is resource adequacy, not co-location: “As stakeholders overwhelmingly agreed during the Technical Conference, the true challenge in supporting data centers is finding sufficient generation to meet exploding demand growth, a challenge that is fundamentally about resource adequacy, not co-location. Serving the forecasted growth of this new load requires a holistic approach to procuring generation and transmission, in addition to coordinated, inter-agency national security policies.” Joint Public Service Parties’ Post-Conference Comments, Co-Location Technical Conference, at 4 (Dec. 9, 2024).

[20] Amazon has highlighted this present uncertainty to FERC: “At present, there is significant uncertainty as to whether the Commission will permit Amazon Energy and other market participants to continue pursuing co-location arrangements, and if so, how to structure such arrangements, which presents a significant challenge for planning load and making long-term investments. This uncertainty affects not only load seeking to co-locate with existing generators, but also future load seeking to co-locate with future generators.”  Post Technical Conference Comments of Amazon Energy LLC, Co-Location Technical Conference, at 2 (Dec. 9, 2024). Calpine Corporation has, as well: “Clear guidance in response to pending filings will foster regulatory certainty, which in turn will promote a more stable investment environment for business leaders and investors who are on the front lines of our nation’s effort to be a global AI powerhouse. Though a more comprehensive effort to update rules and policies may be on the horizon, the Commission should not wait to act.”  Comments of Calpine Corporation, Co-Location Technical Conference, at 4 (Dec. 9, 2024).

[21] The PJM IMM supports the “bring your own generation” idea: “[T}he addition of large data centers would be easier if the data centers brought new generation to the market in addition to new load.” Post-Technical Conference Comments of PJM IMM, Co-Location Technical Conference, at 3.

[22] Susquehanna, Order Rejecting Amendments to Interconnection Services Agreement, at ¶¶85-87 (Nov. 1, 2024).

[23] For example, in the Susquehanna proceeding, in its Order Rejecting Amendments to Interconnection Service Agreement (Nov. 1, 2024) (the “Susquehanna Order”), FERC describes these jurisdiction-based arguments made by intervenors Constellation and Vistra: “Constellation and Vistra contend that Exelon and AEP raise issues outside of the Commission’s jurisdiction. Constellation and Vistra contend that Susquehanna’s sale of electricity to the Co-Located Load is a matter of state law and is not under the purview of the Commission because it neither comes from, nor enters, the interstate transmission system. Constellation and Vistra also argue that the Commission cannot find jurisdiction over the power sale arrangement simply due to the fact that Commission-jurisdictional interconnection facilities may be used in serving the Co-Located Load because the scope of the Commission’s jurisdiction over the interconnection facilities is limited to the service being provided to the generator and the design and operational measures necessary to assure the Commission that no Commission-jurisdictional service is being provided to the Co-Located Load. Constellation and Vistra state that PJM has not created a new type of transmission service as a generator’s behind-the-meter delivery of electricity to the Co-Located Load does not require transmission service under the PJM tariff. Constellation and Vistra contend that such delivery is entirely intrastate, and thus, beyond the Commission’s jurisdiction. Susquehanna also argues that the Co-Located Load’s behind-the-meter delivery facilities and service are not Commission-jurisdictional. Constellation and Vistra claim that the transmission facilities associated with the Co-Located Load are not Commission-jurisdictional facilities because the “facilities [are] used only for the transmission of electric energy in intrastate commerce.” Constellation and Vistra thus argue that Exelon and AEP’s concerns about the Commission’s open access requirements are not relevant.” Susquehanna Order at ¶ 34.

[24] The PJM IMM adamantly insists that “behind-the-meter” load uses grid services: “Contrary to assertions by some supporters of co-located load arrangements, it is not possible for co-located load to be off the grid. All load, including co-located load, is on the grid, affects the grid, and benefits from the grid. This is not a complicated question requiring detailed analysis or explanations. The units that provide power to co-located load are on the grid and benefit from all grid services and could not provide service to co-located load without the grid. As a result, decisions about co-located load in PJM affect all PJM customers.” Post-Technical Conference Comments of the PJM Independent Market Monitor, Co-Location Technical Conference, at 2 (Dec. 16, 2024).

[25] PPL argues that this is sufficient to give FERC jurisdiction: The Commission may not have jurisdiction over the retail customer; but it does have jurisdiction over the wholesale generator and the wholesale market it sells power (and perhaps other services) to.” Post Technical Conference Comments of PPL Electric Utilities Corporation, Co-Location Technical Conference, at 14 (Dec. 9, 2024).   The PJM IMM also weighs in: “The co-located load model would rely on the grid for backup while asserting that it is isolated from the grid.”  Post Technical Conference Comments of the Independent Market Monitor for PJM, Co-Location Technical Conference, at 3 (Dec. 16, 2024).

[26] Vistra Corp. questions whether FERC can lawfully consider resource adequacy: “Adopting punitive policy measures for certain types of load in an effort to address such resource adequacy concerns would be ill-conceived and perhaps unlawful.” Further, Vistra cautions that the role of the regulatory agency is to focus on transmission and interconnection services, while market players are incented to address resource adequacy in response to market forces: “The interconnection process should remain focused on interconnection service and the transmission facilities needed to accommodate it, while markets remain focused on aligning supply and demand. To do otherwise risks not only producing ill-conceived, and perhaps unlawful, policies for co-location arrangements but also undermining the efficacy of the Commission’s regulatory framework for those markets and the nation’s ability to maintain resource adequacy.” Post-Conference Comments of Vistra Corp., Co-Location Technical Conference, at 8, 11 (Dec. 9, 2024).

[27] For example, is one upshot of Susquehanna that it empowers transmission utilities to determine what is or is not acceptable? “The Commission’s actions on co-located load arrangements must ensure that transmission-owning utilities are not discriminatory gatekeepers to new arrangements that meet AI needs. Congress and the Commission have a long record of ensuring that transmission-owning utilities are not empowered to unlawfully restrict competitive access to the grid.”  Comments of Calpine Corporation, Co-Location Technical Conference, at 6 (Dec. 9, 2024).

[28] Amazon has raised this concern: “It is important to recognize that co-located load is not relying on the transmission system in the same way as network load, and therefore, should be not treated as network load from a cost allocation perspective. The Commission should ensure that co-located loads are not assigned charges for grid services that they do not use. Co-location arrangements also ensure that the full cost of any electrical infrastructure needed for load to take power directly from an interconnected generator will be borne by the co-located load, which ensures that those costs are not passed along to other ratepayers.” Post Technical Conference Comments of Amazon Energy LLC, Co-Location Technical Conference, at 3 (Dec. 9, 2024).

[29] Exelon asserts these sales are beyond FERC’s authority: “Sales to end-use load and the interconnection of load are matters of state jurisdiction, and FERC is neither empowered nor well-suited for picking and choosing winners and losers solely based upon interconnection configuration.”  Joint Public Service Parties’ Post-Conference Comments, Co-Location Technical Conference, at 8.

[30] PPL argues that there is no authority to issue such a prohibition under federal law: “Specifically, the sale of energy to the data center is not a “sale of electric energy at wholesale in interstate commerce,” and thus not obviously within the Commission’s jurisdiction. For better or worse, the Commission likely does not have jurisdiction over the load and cannot issue a requirement that it interconnect with the utility directly.”  Post Technical Comments of PPL Electric Utilities Corporation, Co-Location Technical Conference, at 13 (Dec. 9, 2024).

[31] “Will co-located load arrangements remove the load from retail jurisdiction, thus impacting exclusive state authority over retail sales and the rights and obligations of retail electric distribution utilities to serve under applicable certified service territories acts, and eliminating significant revenue (e.g., kwH tax) and contribution to other socialized costs?” Comments of Buckeye Power, Inc., Co-Location Technical Conference, at 13 (Dec. 9, 2024).

[32] “Vistra agrees that co-located loads that take service from the grid should pay their ‘fair share.’ However, where the rubber meets the road on this issue is in how the Commission determines what constitutes a ‘fair share’—and whether the ‘fair share’ is, in some cases, zero. The guiding principle on this issue should be the Commission’s long-standing policy that the system charges for a load that offsets its use of the system through co-located generation should be based on the load’s actual use of the system on a net basis.” Post-Conference Comments of Vistra Corp., Co-Location Technical Conference, at 14 (Dec. 9, 2024).

[33] “The only question is how to serve the potentially very large total increases in load in a way that does not threaten reliability or the ability of PJM markets to reliably serve all load at the lowest possible cost. The co-located load model is clearly not the answer.”  Post Technical Conference Comments of the Independent Market Monitor for PJM, Co-Location Technical Conference, at 1 (Dec. 16, 2024).

[34] “The co-located load is essentially invisible to the RTO/ISO and thus is not considered as part of long-term grid planning processes. This is problematic both for ensuring reliable grid operations and fair cost allocation if the load were to need to come onto the grid for short periods (due to relay failures or other technical issues) or long periods (due to generator shut downs) but the grid has not been designed to accommodate the load.”  Comments of Buckeye Power, Inc., Co-Location Technical Conference, at 14 (Dec. 9, 2024).

[35] PPL Electric argues for all load to be required to be “front of the meter”: “When the new load interconnects behind the meter, the load takes all or a portion of an existing generator off the system. This may increase prices to a greater extent. While this may incentivize new generators to come onto the system, those will be paid for by network load that was previously served by the generator now serving the behind-the-meter load. Thus, network load will pay more for the same service. Further, transmission and distribution upgrades may not be planned for, and if they are, may not be paid for by the new load or the generator it locates behind. These issues can be mitigated through changes to load forecasting and through the generator interconnection process (as was attempted in the Susquehanna ISA), but they will always be less precise than if the process worked as intended and the load connected in front of the meter.”  Post Technical Conference Comments of PPL Electric Utilities Corporation, Co-Location Technical Conference, at 16 (Dec. 9, 2024).

[36] “Large load that is co-located with a generator will essentially remove that generation (assuming it is existing generation), which is typically large baseload generation needed for grid continuity, off the grid so that it can no longer be relied upon by everyone else. While the ultimate balance of demand and supply is arguably the same whether the load is co-located with the generation and behind the transmission meter rather than in front of the transmission meter, a colocation arrangement has the potential to accelerate the pace with which the incremental demand arrives on the system, potentially without any additional incremental supply—exacerbating resource adequacy concerns.”  Comments of Buckeye Power, Inc., Co-Location Technical Conference, at 12 (Dec. 9, 2024).

[37] PPL presents this hypothetical: “To lay out a possible scenario: A large load interconnects with a large generator “behind the meter” while the generator remains interconnected with the larger grid. The load is invisible to the local utility (and the regional grid operator / balancing authority standing behind it). Even if they are generally aware of its existence, from a real-time operating perspective, the load appears only as diminished output of the generator. That generator then trips offline without the transmission system disconnecting it. The large invisible load “appears” on the system, drawing power from the grid. Instead of the generator going to zero output, as the balancing authority’s model expects in an unplanned outage, it now appears as a large, unplanned, unmodeled electricity sink. If the outage occurs close to a system peak, there may not be reserves available to cover this load and the grid operator would have to respond, potentially by dumping load. Worst of all, because the behind-the-meter load is not modeled as load in the system, it cannot be dumped. Thus, normal network load would be dumped instead, despite having paid grid costs avoided by the behind-the-meter load.”  Post Technical Conference Comments of PPL Electric Utilities Corporation, Co-Location Technical Conference, at 9-10 (Dec. 9, 2024).

[38] In Texas, this arrangement is specifically authorized by statute.  Tex. Utils. Code Section 31.002(4-b), (6)(J)(v).

[39] Post-Conference Comments of Vistra Corp., Co-Location Technical Conference, at 12 (Dec. 9, 2024).

[40] See, e.g., the Texas Governor’s November 18, 2024, press release stating that Texas is “ready to be No. 1 in advanced nuclear power.” Available at: https://gov.texas.gov/news/post/governor-abbott-puct-release-texas-advanced-nuclear-reactor-working-group-report.

[41] Post-Technical Conference Comments of the Independent Market Monitor for PJM, FERC Docket No. AD24-11-000, at 2 (Dec. 16, 2024) (“All load, including co-located load, is on the grid, affects the grid, and benefits from the grid. This is not a complicated question requiring detailed analysis or explanations. The units that provide power to co-located load are on the grid and benefit from all grid services and could not provide service to co-located load without the grid. As a result, decisions about co-located load in PJM affect all PJM customers.”).

[42] PPL asserts this is the right outcome: “But, it is clear that behind-the-meter load still benefits from the transmission and distribution systems that interconnect it with the larger grid. It also benefits from voltage, frequency, and the other “ancillary” services that the grid requires to function—and provides to all interconnected customers. Thus, equitable ratemaking requires that behind-the-meter loads share these costs.”  Post Technical Conference Comments of PPL Electric Utilities Corporation, Co-Location Technical Conference, at 12 (Dec. 9, 2024).

[43] At least one stakeholder views this as a relatively straightforward matter to decide: Vistra Corp. asserts that “as long as grid service charges are based on a co-located load’s actual use of the system as determined based on its net load, consistent with well-established Commission policy and precedent, this issue can be resolved relatively quickly.”  Post-Conference Comments of Vistra Corp., Co-Location Technical Conference, at 3 (Dec. 9, 2024).

[44] Post-Technical Conference Comments of Google, LLC, Co-Location Technical Conference, at 2 (Dec. 10, 2024).

[45] The PJM IMM has advocated for this preference: “RTOs/ISOs in general, should do comprehensive planning and explicitly plan for large load additions in the same way that they plan for new generation and new transmission. There should be an orderly queue, including milestones to verify additions, and large loads should not be added until they can be reliably served without disrupting markets for other customers. Large load additions that bring matching generation online would be preferred over large load additions that did not bring matching generation, but both would be part of the queue.”  Post Technical Conference Comments of the Independent Market Monitor for PJM, Co-Location Technical Conference, at 5 (Dec. 16, 2024).

[46] Post-Technical Conference Comments of Google, LLC, Co-Location Technical Conference, at 5 (Dec. 10, 2024).

[47] Post-Technical Conference Comments of Google LLC, Co-Location Technical Conference, at 7 (Dec. 10, 2024).

[48] Calpine Corporation advocates this path:  “Calpine urges the Commission to adopt a “yes and” approach to emerging AI needs. There can be no doubt that the Commission’s policies and some tariff rules will need to evolve in the medium- and long-term to address the increasing prevalence of large, co-located loads and associated supply arrangements and facility configurations. However, the general need for regulatory updates and reforms cannot mean that pending co-location arrangements will be placed on hold indefinitely while the Commission works its way through protracted administrative litigation and rulemaking proceedings and while RTOs engage in stakeholder processes to implement tariff updates.” Comments of Calpine Corporation, Co-Location Technical Conference, at 2 (Dec. 9, 2024).  Vistra Corp. is also a proponent of this approach: “[T]he Commission need not proactively identify every conceivable issue before permitting new co-location arrangements on the grid. Importantly, none of the issues identified to date warrant delaying the interconnection process for co-located load configurations.”  Post-Conference Comments of Vistra Corp., Co-Location Technical Conference, at 2 (Dec. 9, 2024).

[49] See, e.g., President Joe Bidens’s Jan. 14, 2025 Executive Order (“Recent advancements in AI demonstrate its rapidly growing relevance to national security”).

[50] Vistra Corp. supports this argument, asserting that “co-location arrangements of all types—whether for data center load or other large commercial and industrial loads—ultimately support resource adequacy by attracting new investment capital.” Post-Conference Comments of Vistra Corp., Co-Location Technical Conference, at 5 (Dec. 9, 2024).