Data Centers and the US Grid

Microsoft and PowerHouse Hillwood Holding are in disputes over agreements laying out terms and requirements for serving their planned data centers in Wisconsin and Illinois, according to filings at the Federal Energy Regulatory Commission.

The disagreements come after FERC in mid-June found that grid operators’ rules for interconnecting large loads may be inadequate.

In its “show cause” orders issued to the grid operators, FERC detailed five issues it wants regional transmission organizations and independent system operators to address in their large load interconnection rules, including the prevention of cost shifts and transparency into transmission costs. Earlier this month, FERC extended the deadlines the RTOs and ISOs had for responding to the show cause orders until mid-November.

Wisconsin agreements flawed: Microsoft

Agreements filed for FERC approval by American Transmission Co., for example, fail to protect utility customers from the costs for infrastructure needed to serve Microsoft’s data center expansion in Mount Pleasant, Wisconsin, the technology company said in a August 21 filing at FERC.

The four amended Large Load Project Commitment Agreements and the Minimum Transmission Charge Agreement were negotiated by ATC and Wisconsin Electric Power Co., which are affiliated companies, without input from Microsoft, according to the tech company.

“As a result, the agreements predictably contain significant deficiencies — errors, contradictions and inconsistencies — and raise several significant cost-of-service and open access issues,” Microsoft said.

As a signatory to the White House’s Ratepayer Protection Pledge, Microsoft said it is “fully committed” to paying for infrastructure costs related to its data centers.

ATC’s proposed minimum transmission charge agreement lacks a mechanism to prevent WEPCo’s retail customers from paying the costs of the facilities, and large load customers could pay twice, Microsoft said.

Also, a proposed early termination fee would give ATC an unjustified windfall payment, and the transmission company failed to show why it should be allowed to recover project costs through a “construction work in progress” mechanism, Microsoft said.

Microsoft urged FERC to hold “settlement judge” proceedings to resolve the issue.

“Microsoft’s input into the rates, terms, and conditions of these agreements is foundational to ensuring that the agreements can actually facilitate the timely interconnection and operation of this infrastructure necessary to serve Microsoft’s load,” Microsoft said.

The Wisconsin Public Service Commission told FERC that ATC’s proposed large load project commitment agreements were an improvement compared to past practice, but they are “far from fully responsive” to FERC’s show cause order to MISO and “additional work is needed to achieve just and reasonable transmission rates.”

ATC and WEPCo contend that their agreements will protect ratepayers from cost shifts related to building transmission and distribution infrastructure to serve the Microsoft data center campus, according to filings at FERC.

The agreements are “directly responsive” to FERC’s show cause orders, including concerns about shifting costs onto existing customers, according to ATC, which owns the transmission system in eastern and central Wisconsin and in Michigan’s Upper Peninsula.

FERC asked to reject ComEd cancellation notice

Meanwhile, PowerHouse Hillwood on Aug. 14 urged FERC to reject Commonwealth Edison’s “notice of cancellation” that ends a transmission security agreement related to a 1.8-GW, $20-billion data center the company is planning in Joliet, Illinois.

PowerHouse Hillwood contends that the transmission security agreements developed by Exelon’s utilities, including ComEd, vary, unlike pro forma agreements, and reflect the utilities’ monopoly power.

FERC has rules to avoid anti-competitive utility behavior for generator interconnections, but similar rules don’t exist yet for large load interconnections, according to PowerHouse Hillwood, a joint venture between PowerHouse Data Centers and Hillwood Corp.

“Until there are pro forma provisions governing transmission service agreements among PJM, eligible customers, and transmission owners, and other checks and balances to quell utility anticompetitive and unduly discriminatory behavior, the commission must remain vigilant as to the risk that utilities like ComEd will use their monopoly power as a bludgeon,” PowerHouse Hillwood said.

The dispute with ComEd, which sparked a lawsuit by PowerHouse Hillwood, centers on the timing of a security deposit, the company told FERC.

Besides asking FERC to cancel the transmission security agreement, ComEd canceled a retail service agreement that was pending at the Illinois Commerce Commission, according to PowerHouse Hillwood.

The Electric Reliability Council of Texas intends to audit hundreds of data center proposals by Dec. 10, grid officials said August 20, a necessary step for its Batch Zero study process to continue and for grid interconnections of the large loads to resume.

On Aug. 3, Texas Gov. Greg Abbott, R, called for a moratorium on new data center interconnections until questions around their use of energy, water and public funds could be answered.

The delay to the first set of projects to navigate the state’s new large load interconnection process may push back other ERCOT planning efforts, including the grid operator’s long-term load forecast and reliability assessment, officials said at the Public Utility Commission of Texas open meeting.

There are about 300 data centers 75 MW or larger navigating the Batch Zero process, and ERCOT will also perform a community impact review on data centers and crypto facilities of 25 MW and above, to comply with Abbott’s directive, officials said.

ERCOT interconnection queue requests total about 474 GW, and approximately 90% are data centers, according to Abbott’s letter to the grid operator. However, a substantial portion of that may be speculative or duplicative, experts say, as the queue is more than five times ERCOT’s record peak electricity demand.

ERCOT is in the process of developing a request for information template that will go to all the large loads that are provisionally qualified in Batch Zero, Chad Seely, the grid operator’s senior vice president of regulatory policy, and its general counsel and chief compliance officer, said at the open meeting.

“We expect to start sending those out as early as the end of the month and into the first part of September, and then as we move forward into October and November, there may be additional rounds of RFIs,” Seely said.

“Our goal is to head toward a December 10 filing,” Seely said, including a comprehensive report on the verification and audit process delivered to the PUCT a week before its open meeting that month.

The PUCT approved good cause exceptions for ERCOT to an Aug. 7, 2026, Batch Zero classification deadline that Abbott’s moratorium interrupted. The grid operator has not yet requested regulators adjust a deadline to deliver the Batch Zero interconnection study results “because ERCOT does not yet know how the delay will impact the study timeline,” according to a presentation the grid operator filed with the commission. 

“We will not have the study done by April 9, 2027. I think what we're saying is, we're still working on what that new timeline might be,” Seely said.

The Texas legislative session “could delay things further,” Jefferies equity analyst Julien Dumoulin-Smith said in a research note today.

Seely also said that ERCOT, this week, sent out RFIs to utilities for an update on medium-sized loads.

“We have an April snapshot that has about 157 medium data center and crypto facilities, representing almost 8,800 MW, that is expected to come on the grid at some point,” he said. “We've asked those utilities to update that information so we have a real time baseline as we start to move forward with the community impact RFI.”

Seely also noted that ERCOT has 17 large loads, mostly data centers, that have already gone through a stability assessment and are scheduled to come online later this year. The projects would have total peak demand of about 6.6 GW, ramping up over about five years.

“Of course, we have not approved those energization requests ...  these loads would still have to go through some form of the verification and audit process, and they will get the community impact [RFI] as well, as we move forward here over the next several weeks,” Seely said. “But they are the furthest along in the process where they've made it through all the ERCOT gates, except for that last gate, which is the approval to energize.”

Pennsylvania will give data center projects preferential permitting treatment if they commit to a set of power supply, environmental and cost-responsibility requirements under an executive order issued August 18 by Gov. Josh Shapiro, D.

The state Department of Environmental Protection will develop the new review process. To receive a faster state permitting review, data center projects in Pennsylvania must source their electricity from new power supplies, including from growing amounts of firm clean power, according to the order, which affects proposals for projects with a peak demand of more than 25 MW.

The order “should further close the door” to the idea that independent power producers like Talen Energy, Vistra and PSEG Power will be able to sell power from their existing generating assets under long-term contracts to data centers in Pennsylvania, Jefferies equity analysts said in a note on Wednesday. However, Talen’s legacy deal to sell power from its majority-owned Susquehanna nuclear power plant to an Amazon data center appears safe, according to the note.

Shapiro’s executive order comes amid a wave of action by state and local governments aimed at setting limits on data center development. At least 81 cities and counties have moratoria on data center development, according to a National League of Cities database unveiled Wednesday.

Under the permitting framework set by the executive order, data center developers that sign a consent order and agreement requiring them to abide by the state’s infrastructure development standards unveiled in February will receive preferential treatment compared to those that don’t.

For example, applicants that execute a consent order and agreement will have their projects reviewed by DEP on a rolling basis, whereas for developers that don’t sign, the state won’t consider their applications until they have all local permits in hand as well as any needed water withdrawal or wastewater discharge authorizations.

“We have so much speculation in Pennsylvania — like, gold rush speculation on these data centers,” Katie Blume, political and legislative director for Conservation Voters of Pennsylvania, told Utility Dive in an interview. “A lot of this [order] is going to be weeding out those bad actors because they're not going to want to spend five years in the permitting process.”

But Dan Diorio, executive vice president for state policy and government affairs for the Data Center Coalition, sounded a note of caution in an emailed statement.

“It’s important that rules are not changed midstream impacting ongoing investment in verified and responsible data center projects,” Diorio said. “Data centers take compliance and accountability seriously, building only where they are authorized to do so under local, state, and federal rules and regulations.”

The infrastructure development standards include requiring data centers to pay for “all costs caused in whole or in part by the interconnection, service, or load of a [data center] project, including any costs associated with energy and ancillary services, transmission, distribution, network upgrades.”

Besides requiring data centers to be supplied by new generating resources from the same local PJM Interconnection zone where the facilities are built, they must be supplied from growing amounts of firm clean energy such as solar, advanced nuclear and battery storage, Shapiro said when he signed the executive order.

The firm clean energy requirements ramp up from 10% on Jan. 1 to 14.5% three years later and 32% by Jan. 1, 2035.

The executive order directs the Pennsylvania DEP to expedite permitting for clean energy and storage facilities on brownfield sites. The DEP must also facilitate the use of advanced reconductoring and other advanced transmission technologies on existing transmission rights-of-way.

In prepared remarks, Shapiro pointed to the speed of data center development in the state, which he said has “seen an unacceptable number of speculative proposals ... many of them led by developers who have no regard for local communities.”

PPL Electric has about 20.7 GW of potential data center load with electric service agreements in Pennsylvania, according to an Aug. 7 investor presentation. FirstEnergy utilities in Pennsylvania have data center contracts totaling nearly 1 GW, the company said in a July 28 presentation.

It is unclear how those data center projects will be affected by the executive order, but slowing data center development could affect utility transmission spending in the state, the Jefferies analysts noted.

Exelon, FirstEnergy and PPL are “materially increasing their transmission investment with data center demand a key driver,” the analysts said. “If the pace of data center growth slows, we see downside pressure to the pace of transmission investments.”

Currently, data center-driven transmission costs are being shared with residential and other utility customers, they noted.

Shapiro’s executive order directed the state Office of Transformation and Opportunity to remove any existing data center project from the PA Permit Fast Track Program, established in 2024, and rescinded their eligibility for the program. It also barred state agencies from entering into confidential agreements with data center developers.

The PJM Interconnection is considering bolstering its interconnection reliability requirements for “computational loads,” such as data centers and crypto-mining facilities, the grid operator said August 11.

The move to consider “ride-through” standards and other requirements comes after about 3,800 MW of data center load tripped offline in northern Virginia on July 22, the latest and largest such event in PJM’s history, the grid operator said.

The event — triggered by a fault on a 230-kV line in Dominion Energy’s zone — caused a high imbalance between generation and load, and large swings in voltage and frequency, PJM said.

During the event, which occurred in two waves, PJM’s overall load fell about 3.8% to 96,205 MW from 99,984 MW, according to the grid operator, which runs the electric system and wholesale power markets in 13 Mid-Atlantic and Midwest states and the District of Columbia.

PJM said it recovered the Balancing Authority Area Control Error Limit — a measure of the imbalance between sources and uses of power — within nine minutes, compared with North American Electric Reliability Corp. standard of 30 minutes, and dispatched reactive power resources to lower system voltage.

Two similar events in Dominion’s zone in northern Virginia — home to the largest data center cluster in the world — caused sudden load transfers of about 1,500 MW each in 2025 and 2024, PJM said.

The issue was discussed at an Aug. 6 PJM Operating Committee meeting where PJM staff said data centers may be too sensitive to grid conditions.

“This was a normally cleared fault,” Operating Committee Chair Emanuel Bernabeu said at the meeting, per PJM. “They should not disconnect from the grid. These data centers are too sensitive to the kind of voltage that they like, and we feel they are disconnecting too early.” 

As a result, PJM is evaluating a possible expansion of its reliability requirements, “with consideration of existing and future industry ride-through standards and practices,” Matthew Wharton, PJM manager of reliability engineering, said.

PJM’s effort comes as the NERC is moving to set reliability requirements for computational loads.

Under a Federal Energy Regulatory Commission directive, NERC is set to finalize by Dec. 31 registry criteria and initial reliability standards for adding computational loads to the bulk power system. NERC will also develop a plan by March for setting additional reliability standards for computational loads.

However, NERC will not address minimum voltage or frequency ride-through requirements until some time next year, according to PJM.

“You really need these requirements in place before the load comes on the system,” David Souder, PJM executive director of system operations, said at the meeting.

The Texas Public Utility Commission on July 9 unanimously approved rules that will require large computational loads, like data centers and crypto-mining facilities, within the Electric Reliability Council of Texas footprint to stay stable and connected to the grid through disruptions.

Modern computational loads, Kenteel Engineering said in a June blog, are “engineered to protect extremely sensitive and expensive equipment,” and during a voltage dip are programmed to disconnect or enter momentary cessation.

However, this presents a reliability problem, Kenteel Engineering said, as when “several hundred — or several thousand — megawatts of computational load all detect the same sag and drop simultaneously, the grid experiences a sudden, large loss of demand.”

“As LCLs increase on the ERCOT System, similar events would be expected to increase in magnitude and frequency, leading to frequency instability and other reliability problems absent frequency and voltage ride-through requirements,” the Texas PUC rules state. 

The rules don’t immediately penalize facilities that fail to ride through a qualifying event, Kenteel Engineering noted, but instead put them “on the clock” to investigate and report the root cause within 90 days of ERCOT’s request, “develop a corrective plan within 90 days of completing that investigation, and implement the approved plan within 180 days unless ERCOT grants more time.”

“Overriding all of that, if ERCOT judges that continued operation poses an imminent risk to local or system reliability, it can order the [large electric load] — and keep it disconnected — until the Customer demonstrates compliance to ERCOT's satisfaction,” Kenteel Engineering said.

In comments, the Data Center Coalition argued that the PUC lacks the statutory authority to “impose the binding and ongoing operational requirements contained in [the rules] directly on retail customers — a category of entity that the Legislature deliberately excluded from ERCOT’s authority.” 

Texas Industrial Energy Consumers made similar comments, writing, “Unlike Market Participants who must agree to comply with and be bound by all ERCOT Protocols as a condition of participating in the wholesale market, pure retail loads have made no such commitment and have no such obligation.”

TIEC also argued that ERCOT lacks the expertise to “develop reasonable operational requirements for complex, costly manufacturing equipment. It is completely inappropriate to give ERCOT the ability to directly regulate businesses who are not participating in the wholesale market and are not otherwise regulated entities.”

In a staff memo from the PUC’s R. Floyd Walker, senior counsel with the commission’s market analysis division, Walker dismissed concerns over the PUC’s authority. Those commenters “seem to be working under the assumption that explicitly statutory authority is required,” he wrote. “Staff respectfully submits that delegated authority is sufficient.”

“There is no debate that voltage and frequency excursions on the transmission network create reliability concerns, which increase with the interconnection of each new large computational load,” Walker said. “Accordingly, if approved by the Commission, the provisions of [the rules] would be within ERCOT’s authority by virtue of that approval.”

Comments from the Texas Blockchain Council argued that proposed mitigation approaches for LCLs, such as the installation of dedicated battery storage, “are neither practical nor economical at scale.”

“While battery solutions have been suggested, we are not aware of any that have been successfully tested or deployed at the scale required for [LCLs],” the group wrote. “Even if pursued, mandating dedicated batteries for each facility would be unlikely to fully resolve the underlying technical challenges and would impose substantial costs, currently estimated at more than $1.6 million per MW, making such an approach economically prohibitive for most operations.”

ERCOT staff approved of the rules, writing in a market impact statement that they provide “necessary requirements to reduce the reliability risk posed by LCLs unexpectedly tripping or transferring to backup generation when frequency and voltage excursions within a specified range occur.”

In official comments, ERCOT wrote that LCL loss wasn’t a hypothetical, and ERCOT “has experienced 28 events involving LCL trips of at least 100 MW due to voltage and frequency excursions since the beginning of 2023. This risk will increase exponentially with the significant growth of LCLs expected in the ERCOT Region.”

Developers have requested studies for more than 438 GW of large load projects within ERCOT’s footprint, and “even if only a small fraction of these projects materialize, this will significantly increase the risk that cascading outages could occur due to LCL failures to ride through typical voltage or frequency disturbances,” ERCOT said.

With the U.S. facing an electricity affordability crisis and demand rising quickly for the first time in decades, experts say the Federal Energy Regulatory Commission’s June 18 order to system operators to provide transmission for flexible large loads underscores the urgent need to implement data center load flexibility.

Just a 1% to 2% reduction in data center peak demand can reduce electricity rates 0.5% to 2.8% and protect reliability, according to a 2026 Duke University Nicholas Institute study. And pilots and analysis led by the Electric Power Research Institute, or EPRI, show how flexibility is within reach that can also get data centers interconnected faster.

Electricity demand by artificial intelligence, or AI, data centers is driving a global urgency, Nat Bullard, chief strategy officer of research group Halcyon, reported in May. From Q1 2025 to Q1 2026, Amazon Web Services’ cloud business grew 28%, Microsoft Azure grew 40% and Google Cloud revenues increased 63%, he reported.

“Growth rates this high in already-mature businesses mean total revenue doubles in two years (or less),” Bullard wrote. “That revenue can only be serviced with compute, and that compute can only serve when energized.”

Experts told Utility Dive that public opposition to data centers, focused on local electricity costs, can be addressed by flexibility. Data center flexibility can reduce electricity demand, which reduces the costs to ratepayers of investments to protect reliability, the North American Electric Reliability Corp. acknowledged in its 2026 large load risk mitigation guidelines.

Flexibility can get a data center interconnected faster, make it a benefit to system reliability and increase system utilization which lowers rates, utilities, researchers and analysts agree. The challenge remaining is to resolve control issues between risk-averse utilities and impatient data center operators, they also agree.

The flexibility framework

Power demand from U.S. data centers will reach 66 GW in 2027, up from 31 GW in 2025, with summer peaks growing to 8.5% in 2027 from 4.1% in 2025, Goldman Sachs reported May 20. Though AI proliferation is uncertain and growth will vary widely by state, data centers could consume as much as 17% of U.S. electricity usage by 2030, EPRI found.

Flexibility adds “headroom” to a power system by allowing system operators to add large loads and maintain reliability while minimizing rate raising infrastructure investments, EPRI’s FlexMosaic framework concluded. That is why data centers willing to be flexible can also maximize their “speed to power,” the institute’s analysis found.

There are five flexibility “classes,” according to EPRI. Class A supports power systems during “infrequent extreme stresses” and Class B manages “daily or weekly” demand peaks. Class C meets long-lasting energy shortages, Class D protects in sudden supply or demand swings and Class E provides frequency stabilizing.

Key determinants of the flexibility class of a data center are the notification time it needs, the duration and frequency of its response, and the depth and quickness of its flexibility, FlexMosaic said.

FlexMosaic’s objective is to align incentives of data centers and utilities by linking faster and bigger data center interconnections with contractual agreements that protect utilities. Data centers with class D and E flexibilities that mitigate the local thermal overloads or voltage drops “unlock” the most system value, EPRI added.

To be flexible, data centers can combine three flexibility pillars — managed workloads, reduced AI plant energy consumption and back-up power, said Anuja Ratnayake, the emerging technologies executive leading EPRI’s DCFlex Initiative. FlexMosaic’s aim is to standardize data center designs and utility programs that include flexibility, she added.

EPRI’s work contradicts the assertion that data centers cannot be flexible, said Ann Rendahl, president of the National Association of Regulatory Utility Commissioners and a Washington state utilities commissioner.

The findings from EPRI’s pilots and analysis will allow state regulators to require data centers to take the potential for flexibility seriously, Rendahl added.

The flexibility breakthrough

Using onsite or stored power during system emergencies and reducing building electricity use at operator requests is not new for data centers and other large loads. But FlexMosaic’s third pillar, data center workload flexibility, is a new opportunity to protect reliability and affordability that is now within reach.

Today’s AI data centers can “be designed with flexibility as a core operational principle,” Boston University researchers reported in June. Training and inference workloads “can offer between 18% and 55% flexibility relative to their average power consumption” and still meet quality of service requirements, modeling found.

Emerald AI “operates as an orchestration and optimization layer between utilities and data centers” and “does not directly control utility or data center operations,” Emerald AI head of product Mansi Shah said. “Utilities retain full dispatch authority,” she added.

Emerald AI “enables utilities to issue curtailment or flexibility requests to participating data centers through a secure software interface,” Shah continued. The platform “translates grid requirements into operationally feasible dispatch targets and provides telemetry, verification, and event compliance reporting back to the utility operator,” she said.

In short, Emerald AI coordinates “approved flexibility actions,” Shah said. The actions address “compute, cooling, energy storage, backup generation, and power infrastructure,” but always respect “predefined operational guardrails,” she added.

Every utility-data center interaction “is governed by operational parameters agreed upon by both parties in advance,” Shah said.

The utility defines the parameters of an event requiring flexibility, including “maximum magnitude, minimum notice period, frequency limits, and event duration,” Shah continued. But data centers set the “hard floors that protect critical workloads and infrastructure under all circumstances,” she said.

These terms give utilities “confidence that contracted flexibility will perform reliably,” and assure data centers “that operational and [technical] boundaries will never be violated,” Shah added.

EPRI’s demonstrations have delivered flexibility of up to 40%, but future AI infrastructure with designed-in flexibility can “materially” expand that, Shah said.

EPRI, NVIDIA and utility demonstrations with Emerald AI software have shown that training workloads “can be paused or slowed,” and inference queries “could be redirected” to a data center on a less stressed system, NVIDIA said in a July 2025 blog post.

In a peer-reviewed test of data center workload flexibility, Emerald AI’s platform gradually ramped an AI workload at electric utility Salt River Project 25% during a three-hour peak, the release said.

Another Emerald AI test shifted a data center workload from Virginia to Illinois without a loss of service quality.

A UK test showed Emerald AI software using NVIDIA hardware could reduce an AI load over a third in under a minute while protecting critical compute.

Emerald AI sent Portland General Electric’s signals to an Oregon data center and reduced power 20% in simulated weather emergency scenarios, Emerald AI reported in March.

Emerald AI has now proven “temporal flexibility” by “slowing or pausing AI workloads” with “customer-designated flexibility on completion time,” the company reported in March. It has also proven “spatial flexibility” by “seamlessly rerouting latency-sensitive loads” from a power-constrained data center “halfway across the country” to where power is available, it added.

In late 2026, EPRI, NVIDIA, Emerald AI and their partners plan to bring the 96-MW Aurora AI Factory online in Manassas, Virginia. It will validate workload flexibility at scale in a data center “designed for flexibility,” EPRI’s Ratnayake said.

The needed agreement

The missing piece in EPRI's DCFlex plans is a standardized binding agreement between utilities and data centers, stakeholders said.

A key parameter of that agreement will be explored in the next EPRI-led pilot. Emerald AI will flex an NVIDIA data center served by Silicon Valley Power, or SVP, which serves 58 data centers in its 20 square mile territory, Chris Karwick, its chief operating officer, said.

For phase one of the pilot, Emerald AI is developing “a bidirectional communication platform with the data center,” Karwick said. SVP “will see the NVIDIA data center load in real time, send a signal to reduce the load in response to simulated events like a forecasted heat wave or a sudden lightening strike, and see the load reduced,” he added.

SVP must also have “the safety net of 100% control of the loadside breaker if the data center wants to have the faster interconnection and additional capacity,” Karwick said. “That is non-negotiable,” and would be part of the interconnection agreement SVP makes with the data center, Karwick said.

The Portland General Electric, or PGE, large load study process already identifies each new data center’s flexibility capabilities, said Isaac Barrow, the utility’s senior manager of data centers and growth. But traditional bill credit incentives are too small to guarantee AI data centers will reduce their lucrative workloads when the utility sends a signal through Emerald AI, he added.

An interconnection agreement offering the right incentive structure will, however, “unlock the next wave of flexibility technologies,” Barrow said. For PGE, an agreement providing accelerated interconnection for a data center must guarantee “visibility” and “dispatchability” of the data center’s load, he added.

New utility regulations could simplify the agreement process by requiring data centers to have flexibility capabilities, said Joe Reele, vice president, solution architects, for energy technology company Schneider Electric. But regulations could require consent to a single digital platform by the power and data center ecosystems, he acknowledged.

With 50 state regulatory jurisdictions and many other local jurisdictions, “it is not an easy solution” with so many proprietary software platforms in use, Reele said.

Will they agree?

The value of a definitive flexibility agreement between data centers and utilities is clear.

“Power availability has become the gating issue for AI,” said Mona Dajani, c-chair of multinational law firm Cooley’s infrastructure, energy and real estate group. In constrained power markets, projects that have “credible, measurable, and contractually defined flexibility may increasingly receive more favorable treatment,” she added.

But the “pain point” remains utilities’ need for control and data centers’ need to choose how to respond to a utility signal for a load reduction, said EPRI’s Ratnayake.

A control mechanism like the one SVP’s Karwick described “is a really hard concept for data centers to embrace because they lose control,” Ratnayake said. Data centers need a more gradual load reduction because their highly sensitive hardware would be at risk and it is probably worth much more than many power system assets, she added.

Utilities need control because their experience with demand response programs shows they cannot always depend on voluntarily load reductions, Ratnayake continued. And “if a GW scale data center goes offline instantaneously it would create a whole host of other system issues,” she said.

Ultimately, the data center and the utility need a program structure defined in a contractual interconnection agreement, Ratnayake said.

The April memorandum of understanding between EPRI and the Open Compute Project Foundation, or OCP, is a major step toward that agreement and those protocols, Ratnayake said. It creates a channel for EPRI utility members and OCP’s tech community to collaborate on standards and protocols that work for both, she added.

Validating technologies, regulatory structures and market certainty will take time, but “it will unlock the power needed to keep pace with needs of AI,” Ratnayake said

Meanwhile, “the reality today is if you want grid power at scale, you cannot have 100% of the hours,” said James Kacergis, senior vice president of corporate and business development for Terawulf, which is building the 750-MW Mariner Lake AI data center for anchor tenants Core42 and Google-backed Fluidstack.

“A hyperscaler’s primary incentive is getting the interconnection to scale the business, and if it is not willing to work with the utility or system operator, a competitor will,” Kacergis said. But “it is not yet clear what the amount of load reduction and duration is that will attract data center customers because it's an emerging market,” he added.

Data centers are capable of slowing, capping or shifting workloads, “but it's questionable whether they want to do it,” added Steven Carlini, chief advocate of AI and data centers for Schneider Electric. If, though, the choice is “having 80% of the data center working or none of it, they would take 80%,” he added.

“Data centers want power as soon as possible and they want that power to be reliable, and utilities want to bring them online as quickly as possible and protect reliability,” said NARUC’s Rendahl. “Flexibility is a way to do what both want.”

The pathway for adding data centers to the U.S. transmission system became clearer when the Federal Energy Regulatory Commission on June 18, 2026, found that major grid operators’ rules for large load interconnections appeared to be inadequate.

“We take historic action to push our country’s electric markets and economy into the future — a future of fair cost allocation, unprecedented transparency for the American ratepayer, respect for states’ rights, efficient markets and speed to power,” FERC Chairman Laura Swett said in a concurring statement with the six orders the agency issued.

In its “show cause” orders, FERC detailed five issues it wants regional transmission organizations and independent system operators to address in their large load interconnection rules:

• efficient transmission service application and study processes, including the consideration of alternative transmission technologies;
• prevention cost shifts and transparency into transmission costs;
• co-location and behind-the-meter generation;
• new transmission services for flexible large loads; and,
• a study process for generating facilities that serve “electrically proximate” large loads and co-located loads.

FERC’s decision was driven by a massive wave of data center development that started about two years ago, after years of generally flat electric demand growth in the United States. The development has sparked a race for data center developers to get power to their facilities as quickly as possible and created political backlash over rising electricity bills.

In October, the U.S. Department of Energy directed FERC to establish rules for enabling data center interconnection to the transmission system. FERC’s response to the DOE’s “advanced notice of proposed rulemaking” drew more than 3,500 pages of comments and was completed in roughly eight months.

“FERC's action is far more substantively ambitious than the ANOPR, and it enables region-specific investigation pathways that should be more effective than a uniform rulemaking,” Devin Hartman, a senior fellow at the R Street Institute, said in an email to Utility Dive.

Ultimately, new interconnection rules could provide greater cost clarity for electric utilities, according to Morningstar DBRS.

“For regulated utilities, the key benefit is a clear framework for determining the need for large-load-related infrastructure, who pays for it, and how long the large-load customer must remain financially responsible for the investment and to an extent mitigate customer affordability concerns," Zujian Li, Morningstar DBRS’ vice president of energy and natural resources ratings, said Monday.

For data center developers, FERC’s orders indicate that the agency is “committed to prioritizing projects that can prove they are real, financeable, operationally flexible, and capable of integrating with the grid without imposing unjustified costs on other customers,” Jane Rueger, Perkins Coie data center co-chair, said in an email to Utility Dive.

FERC’s orders were issued to: the California Independent System Operator; ISO New England; the Midcontinent Independent System Operator; New York Independent System Operator; the PJM Interconnection; and, the Southwest Power Pool.

Here are six takeaways from FERC’s landmark decision.

FERC opts for regional approach

FERC directed the RTOs and ISOs to develop their own data center interconnection rules that can best meet their specific needs.

“The era of one national standard for data center interconnection is over before it began,” Mona Dajani, a partner and the co-chair of Cooley's infrastructure, energy and real estate practice, said in an email to Utility Dive. “What replaces it is six regional answers to the same question, decided on six different timelines.”

As a result, data center developers face new diligence needs for near-term deals, according to Dajani.

“A term sheet signed against today’s PJM tariff carries different risk than one signed against today’s MISO or CAISO tariff, because each region’s show cause filing will land on its own schedule with its own substantive answer on cost allocation, co-location, and flexible load service,” she said.

FERC was smart to avoid a “one-size-fits-all approach,” according to Gretchen Kershaw, COO and vice president of strategy at Grid Strategies, a transmission-focused consulting firm.

“FERC did an admirable job of explaining the nuances of each region in the orders,” Kershaw said in an email to Utility Dive.

However, Kershaw said she would like FERC to take some action outside of RTOs as well. “The framework of vertically integrated utilities and bundled retail sales is distinct, but there is no part of the country where there is less transparency than in these regions,” she said.

Hartman echoed those comments. “I wish FERC had included non-RTO areas, which typically suffer from the worst transmission and interconnection practices,” he said.

About a third of Americans live outside of RTOs, former FERC Commissioner Allison Clements noted in an email. “By punting on the tougher jurisdictional questions, FERC has left these customers without the cost and transparency protections that will be made available to families in RTO regions," said Clements, a partner with digital infrastructure advisory firm ASG and principal of 804 Advisory.

FERC encouraged transmission owners outside of RTOs — such as much of the West and Southeast  — to propose data center interconnection rules.

FERC ‘stays in its lane’

With its orders, FERC appears to have avoided conflicts over federal and state jurisdiction, a major concern that arose when the DOE ordered the agency to develop large load interconnection rules.

“We recognize that states will continue to regulate: (1) the specific terms of retail sales to large load; (2) which entities may make retail sales within their borders, including which entities are legally permitted to provide electricity to retail large load customers; and (3) any siting decisions and construction associated with the large load project,” FERC said in its PJM order. “Nothing in this order is intended to intrude upon state authority over retail service to large loads.”

FERC did a “great job” staying in its jurisdictional ambit, according to Kershaw.

“As a long-time FERC lawyer, I think FERC has more jurisdiction than it has asserted, but I recognize that fast and durable action is easier when litigation risks are reduced — and that is no more true than when sticking to existing jurisdictional lanes,” Kershaw said.

Protecting ratepayers from cost-shifts

In its orders, FERC said it was concerned that there appears to be a lack of transparency in how RTOs and ISOs determine how network upgrades are assigned and paid for. Also, grid operators appear to lack pro forma cost recovery agreements for network upgrades that can be required to safely interconnect new loads to the grid. Those agreements can help ensure that large load customers bear the risk and pay for their transmission service, including the cost of network upgrades, FERC said.

FERC, for example, directed PJM to explain whether its tariff remains just and reasonable without providing “robust, accurate, and systematic” and easily searchable data on its website on network upgrade costs.

FERC directed RTOs and ISOs to have “cost recovery agreements” to help make sure that large loads pay their share of the costs incurred to serve them, FERC Commissioner David Rosner noted in a concurring statement.

“We target speculative projects by establishing escalating readiness requirements for distinct phases of the study process to deter duplicative or speculative requests for transmission service,” Rosner said.

Options coming for flexible loads

“Load flexibility can avoid inefficient and costly transmission system build-out,” FERC said. “In addition, transmission services that reflect that flexible large loads are willing and able to limit their withdrawals from the transmission system under certain conditions could help timely interconnect flexible large loads.”

FERC’s orders reflect a growing recognition that data centers can be more flexible than traditional loads, Robert Montejo, a data center partner at the Duane Morris law firm. “The ability to curtail or operate flexibly could become one of the most valuable tools for getting new projects connected faster,” he said in an email.

FERC’s orders preliminarily find that the agency should extend the transmission services developed in PJM’s co-location proceeding — interim network integration transmission service, firm contract demand and non-firm contract demand — to new types of load and new regions, FERC Commissioner Judy Chang said in a concurring statement.

“While these services have the potential to facilitate more efficient use and build out of the transmission system, they also contemplate running the system ‘tighter’ than we have done in the past, potentially with more loads on the system served by co-located or behind-the-meter generation, and potentially more use of batteries, load control systems, and backup resources to manage demand during system peaks or other stressed conditions,” Chang said.

FERC calls for ATT evaluations

In its orders, FERC called on grid operators to consider alternative transmission technologies — such as advanced power flow control devices, synchronous condensers, advanced conductors and dynamic line ratings — when conducting transmission service request studies for data centers and other large loads.

If a grid operator decides traditional network upgrades are needed, they should clearly show why ATTs are infeasible or would not lower costs or offer a faster timeline for interconnecting a large load, according to FERC.

“The goal is to respect transmission providers’ engineering judgments while protecting against upgrade costs when ATTs could solve transmission needs faster and with a lower bill,” FERC Commissioner Lindsay See said in a concurring statement. 

ATTs can “unlock every megawatt of existing capacity from our current transmission system,” FERC Commissioner David LaCerte said. “We have the technology. We should use it now to enable faster interconnection of large loads, lower costs to customers, and help prevent cost shifting.”

Reform timeline could be challenging

FERC gave the RTOs and ISOs 60 days to respond to their show cause orders. However, within 45 days, they can request 90-days extensions to be added to the initial 45 days.

That timeline may result in “relatively cursory reports, not detailed reform proposals,” Hartman said in a social media post. “Regional stakeholders have to rapidly diagnose deficiencies and identify reforms for issues that often take multiple years to develop remedies.”

The timeline is “really aggressive” and there may be governance barriers to meeting it, according to Kershaw.

“I expect we'll see some requests for abeyance,” she said. “I will also be watching the PJM governance technical conference carefully for how FERC is thinking about reforms that can result in expedited filings like those envisioned by the show cause orders.”

LaCerte warned that the agency expects solid responses from the RTOs and ISOs, saying the agency used “considerable restraint” through its show cause orders.

“The commission has very broad jurisdiction over transmission that we will not hesitate to utilize as necessary to ensure that we meet our objectives in all show cause proceedings,” he said in a concurring statement to the orders.

If the grid operators fail to adequately address the large-load-related concerns identified by FERC, the agency will dictate the solutions, according to LaCerte.

“I say this not as a threat, but as a statement of duty,” LaCerte said. “While you may not like our remedies — I have often said that the federal government generally does not come up with the best solutions — this is an outcome we are prepared to pursue given the gravity of the moment and our statutory obligation.”

The North American Electric Reliability Corp. on May 4 issued a rare Level 3 alert — the watchdog’s highest level — following instances of data centers unexpectedly dropping load or oscillating demand rapidly, creating reliability concerns. The alert includes seven actions grid entities “must implement to address immediate risks posed by computational loads,” NERC said.

Transmission planners and operators, system planners and balancing authorities are among the entities that must act. The required actions address the modeling, study, operation, protection and control of computational loads, including artificial intelligence training and cryptocurrency mining.

“The grid faces unprecedented challenges from a surge in large power consumers,” NERC said in a statement. Summer peak demand across the bulk power system is expected to rise 24% in the next 10 years, with data centers accounting for most of the increase, the organization said in its most recent Long Term Reliability Assessment, published in January 2026.

In the face of rising data center loads — and instances where they unexpectedly disconnected from the grid — NERC issued a Level 2 warning last year that elicited alarming responses from grid stakeholders, the reliability watchdog said.

“Entities generally did not have sufficient processes, procedures, or methods to address emerging computational loads,” NERC said in its warning. The essential actions NERC calls for include the following:

• Transmission planners and planning coordinators should develop a detailed list of modeling data, settings, and parameters needed from computational loads and distribute this to transmission operators in their footprint. Transmission operators should reflect this information in their facility interconnection requirements.
• Transmission planners and planning coordinators should collect data from computational loads such as the expected minimum and maximum consumption in megawatts and the percentage of IT load vs. non-IT load (cooling, for instance), at various load levels.
• Planning coordinators should revise definitions that trigger a review of local area protections, stability limits and other reliability studies to account for computational loads.
• Transmission operators should establish a “commissioning process” for computational loads.
• Where possible, the commissioning process transmission operators develop should include testing facilities at full load and at no load, and, if possible, with at least a 10% change from nominal voltage.
• Transmission operators should install and utilize dynamic fault recording devices to understand computational load facility electrical performance during system disturbances.

Registered NERC entities must acknowledge receipt by May 11 and must respond by Aug. 3.

Level 3 alerts, which are rare, call for grid operators to take specific steps in response to an immediate issue. NERC last issued a Level 3 alert a year ago regarding possible grid disruptions related to inverter-based resources. Solar, wind and battery resources have been known to trip offline during grid disturbances. 

Grid instability issues caused by data centers "could become quite severe — to the the point of creating widespread blackouts," Ben Inskeep, program director for the Citizens Action Coalition, told Utility Dive in an email.

"It is critical that we have standards in place to protect all ratepayers from negative impacts to grid stability, reliability, and resiliency that are caused by data centers, especially given the rapid proliferation of these mega users across the country," Inskeep said.

NERC’s efforts to better align utility and data center operations may take longer than anticipated, Digital Power Optimization CEO and founder Andrew Webber told Utility Dive. The company builds and manages data centers.

“It will take years of coordinated effort in the drafting of new regulations, understanding limitations and opportunities related to physical equipment, understanding limitations and opportunities related to software and control systems, re-prioritizing various loads all throughout society, etc.,” Webber said in an email.

He pointed to recent comments by Laura Swett, chair of the Federal Energy Regulatory Commission, about a lack of communication between data centers and energy regulators.

“Given the absolutely critical need [for data center developers] to source power for their developments, I found it quite interesting that FERC isn’t seeing more direct engagement from the data center industry,” Webber said. “I think it exemplifies the challenges NERC is also going to face in getting genuine buy-in and wholehearted acceptance by the data center industry as it pertains to grid reliability.”

“New evaluation paradigms will need to be created and worked through in partnership with the data center industry,” he said.

State policymakers are increasingly looking to large load tariffs as a way to shield ratepayers from the surge in interconnection requests from artificial intelligence data centers, with 77 such tariffs pending or in place across 36 states, according to the Smart Electric Power Alliance, which recently updated its Database of Emerging Large Load Tariffs. 

In 2025, state regulators approved 29 large load tariffs, compared to 14 between 2018 and 2024, SEPA said. The database, which is jointly maintained with the North Carolina Clean Energy Technology Center, has added 12 entries and 3 additional states since its last update in November 2025

Experts say it’s too soon to know whether the tariffs are working as intended, and broader reforms to utility planning and power markets are likely needed. “We’ve seen the tariffs that set the initial agreements and the kind of frameworks, and they have these really important safeguards, but in a lot of states, we haven’t yet gotten to that process of cost allocation, which is complicated,” Louisa Eberle, a senior associate with the nonprofit Regulatory Assistance Project, told Utility Dive. “There are several reasons why rates might still go up for other customers if cost allocation isn’t done effectively.”

The growth of large load tariffs comes amid rising retail power costs that are increasing pressure on elected officials, regulators, utilities and tech companies to justify an estimated trillion-dollar grid buildout to meet demand growth projections from data centers and, to a lesser extent, manufacturing and electrification.

“I think the properties of that growth at a regional level are what drives the increased look at how we could build upon the longtime commercial and industrial tariff structures and principles to shore up guardrails for this new growth era that we’re in,” Ann Collier, SEPA’s director of grid strategy, said in an interview. “The growth is coming at a much faster pace than we have seen in decades, and it’s much denser.”

In particular, the definition of “large load” has evolved as the size of projects have grown. A few years ago, utilities defined “large loads” using a threshold of five, 10 or 25 MW, Collier said, but more recently-approved tariffs are defining large loads at 50 MW and above.

Other common features of these tariffs include asking loads to pay more of the upfront costs for engineering studies or power delivery infrastructure, minimum contract terms, minimum load guarantees, defined load ramp schedules, exit fees and financial security in case a project falls through. Some have also introduced incentives, such as reduced load guarantees or faster interconnection, for customers that bring their own capacity or agree to flex their power demand during times of grid stress.

Collier said while utilities are still originating many of their own large load proposals, she’s noticed that states are taking a more active role in regulating large loads, either through regulatory bodies or legislation. Oregon, Minnesota, Indiana, Texas and Virginia are among the states that have passed laws governing how large load customers are treated by utilities. State regulators in Pennsylvania are developing a model large load tariff that could be applied to all utilities there. 

Collier said “time will tell” if the tariffs and other rules for large loads are effective, and the needs of each territory are different. But some early evidence appears to show that the tariffs can weed out speculative requests from the large load pipeline. 

AEP Ohio cut its large load forecast by more than 50%, from 30 GW to 13 GW, after regulators approved a new large load tariff last year, though manufacturers in the state say data center demand is still inflated.

“AEP has designed its large‑load tariffs, both those already approved and those pending across our jurisdictions, to ensure that large‑load customers pay their fair share of the costs to serve them,” Scott Blake, a spokesperson for AEP, said in a statement to Utility Dive.

“Our approach is focused on protecting existing customers from bearing the costs of grid upgrades or infrastructure improvements needed to accommodate these large new loads,” he added. “Each of our state jurisdictions has unique regulatory frameworks, but across all of them, our priority is ensuring fairness, transparency, cost causation, and long‑term system reliability as demand from large‑load customers continues to grow.”

Eberle said utilities should prepare for more scrutiny of their requests as regulators contend with a growing affordability crisis. 

“In some of these cases, the utilities are really asking regulators to approve things that, you know, they’re big resource commitments,” she said. “If a utility can also come in and say, ‘We’ve looked at grid enhancing technologies; we’ve looked at energy efficiency; we’ve looked at demand response; we’ve looked at all of these measures that can make this cost effective, and we’ve included them,’ I think that can go a long way.”

The data centers being planned and built across the U.S. need a massive amount of electricity, and utilities are racing to build new generation and grid infrastructure to meet unprecedented demand growth — driven in significant part by these large load customers — while coming up with contract structures that protect their other customers. 

But several factors are working against utilities as they seek to manage risk, experts and industry sources say, potentially leaving them and their ratepayers or members vulnerable to stranded investments.

The electric power industry works on planning horizons of decades, while most of the data center demand surge associated with artificial intelligence arrived in November 2022 with the public release of ChatGPT, which was followed by rapid technological advancements in the field.

The load profile of these AI data centers is fundamentally different from earlier data centers and other industrial or commercial loads. They would be hard to replace once their presence has been written into utility planning. At the same time, the ever-evolving nature of both AI and the hardware it runs on makes it difficult to predict how much power data centers may need in the future.

Many have compared the possibility of an AI bubble to the dot-com bubble that burst around the year 2000. 

However, the legacy of that crash — millions of miles of abandoned fiber-optic cable, or “dark fiber,” later tapped to power data-intensive applications like AI — has no good analogue to the current moment, said Advait Arun, senior associate for energy finance at the nonprofit Center for Public Enterprise.

“It's still not exactly clear what you could do with [data center] assets in, like, a dark GPU situation,” Arun said. “With fiber optic, the Internet was still a thing, and it was the infrastructure on which you could build newer websites or new service models. But it's not really clear what the future of the inference services industry looks like, simply because we have all of these competing companies with fairly identical services.” 

“This is a business model that hasn't exactly proved it can generate revenue or have a stable core to its demand,” he added.

Arun examined the issue of GPU depreciation in a November paper for CPI, in which he wrote that neocloud companies like CoreWeave, which often provide GPU capacity to hyperscalers like Google, are “disproportionately exposed to the risk of falling asset values — but this risk is a threat to the whole sector.”

In the event of any crash or market correction that reduces demand for inference services, neoclouds “are definitely most at risk,” Arun said, along with any data centers that they're going to be building.

“These companies will almost certainly have a much worse time than the leading cloud service providers,” such as Google, Microsoft and Amazon, he said.

One of the biggest risks of a significant increase in large load customers is simply the credit risk they pose, said Scott Engstrom, chief customer officer at GridX.

“Who’s on the other side of signing up for, say, $100 million a year of minimum payments?” he said. “You have to be confident that the counterparty is going to be around for the period of time that they are committed to recover that infrastructure investment. That’s a way that this can go wrong — somebody goes into bankruptcy, they go out of business, they lose funding, whatever it is.”

Engstrom said that when it comes to hyperscalers like Google, Meta or Amazon, “you certainly feel very good that they are going to have a lot of money 10, 15 years from now. Could the world change? Obviously.” 

The increasingly circular nature of the inference services industry also poses challenges for utilities that are trying to get a clear picture of demand and manage risk. 

There are “reasonable concerns” about the interconnected nature of some of the investments, said Daniel Farris, a law partner at Foley & Lardner who works on data center and energy contracts.

“The chip manufacturers are investing in neoclouds so that they can go secure data centers that are backstopped by those chip makers, so that they can turn around and spend money that they receive from the chip makers to buy chips from the chip makers,” Farris said. “Is there the potential for market correction? Yes, absolutely.”

In the CPI paper, Arun noted that CoreWeave’s growth trajectory “may not be sustainable if its two key revenue sources — Microsoft and NVIDIA, over 70% of CoreWeave’s revenues — do not meet their payment commitments,” and that CoreWeave is “on the hook for over $56 billion in data center lease payments, which will last around 10 years.”

‘We don't really have anything else that could take up that space’

Gas and nuclear are two of the most sought-after solutions for data center energy demand, but a supply chain crunch is stretching out the timeline for new gas generation to at least several years, and new nuclear can take even longer.

“Data center demand is hard to project over the next few years,” Arun said. “In a market correction, it's very possible the data centers that have promised to pay for these [combined-cycle gas turbine plants], that can no longer pay for them, will end up crashing out of their tariff arrangements, and the utilities will cut these gas plants from their [integrated resource plans].”

In the event of a market correction that leaves data centers “unbuilt but in possession of those rights for the land interconnection, that's basically hundreds of megawatts worth of interconnection that is now not being used — maybe up to a gigawatt, if that gigawatt data center gets built,” he said. “And we don't really have anything else that could take up that space.”

Farris noted that “everybody in the industry expects there to be something of a power cliff” — in other words, a lot of the available dispatchable power has largely already been “acquired or secured by most of the hyperscalers, neoclouds, data center operators, and so renewable energy is still a source people are looking to.”

A big factor in determining power sources “is going to be the consistency of the power to keep up with these AI loads, which are still somewhat chaotic,” he said. “They're not as well-balanced as more traditional CPU kind of loads.”

GPUs, a more advanced type of processor than a CPU, are essential to AI workloads, and can individually draw up to 700W. They also have a tendency to generate unforeseeable energy spikes when running those workloads, sometimes called AI power bursts. Data centers can house tens to hundreds of thousands of GPUs. 

Not all data centers contain GPUs, but all of them contain servers, which themselves “use a lot of electricity just because of the sheer number,” said Christopher Tozzi, a technology analyst and senior lecturer at the Rensselaer Polytechnic Institute in Troy, New York. “I don't really see a way to mitigate that issue. There are ways of trying to make server components a little more energy efficient, but they're already pretty energy efficient.”

Servers, with or without GPUs, also use a lot of electricity for cooling. More energy efficient technologies for cooling exist, Tozzi said, but they are more expensive, and he doesn’t see pressures in the market that are likely to drive down the cost.

“My overall sense is that the data center industry right now is more focused on the idea that energy itself will become cheaper and more abundant and that will solve their energy problems, as opposed to trying to find ways to make data centers more energy efficient,” Tozzi said.

Before the public release of ChatGPT, Farris noted, hyperscalers like Microsoft, Amazon, Google and Meta were very focused on improving efficiency and sustainability in their facilities.

“Since we've had the last three-ish plus years of the arms race for securing power and creating these much larger [high-performance computing] data centers, some of that has gone by the wayside as everyone's just trying to secure power,” he said.

“But I think you'll have a return to that strategy over time,” Farris added. “We're seeing public opposition to data centers, and that’s for a variety of reasons, one of which is certainly sustainability. And there are definitely folks pushing for cleaner power, so that helps to support renewables.”

Arun said he anticipates that solar and storage, which are significantly cheaper to build than new gas plants, “will likely stay on IRPs and will stay in the queue, regardless of what happens to data centers.”

“Not just because electrification and demand growth is still happening at a lower rate even without data centers, but also, these are just cheaper and better for the ratepayers’ balance sheet,” he said. “I think utilities will want to avoid, at all costs, burdening ratepayers with assets that a large load customer is no longer able to pay for.”

Large load tariffs, bring-your-own-capacity models gain traction

Utilities are increasingly using methods like large load tariffs and long-term contracts, sometimes with take-or-pay clauses, to manage the risk of connecting data centers to the grid. A December analysis from Enverus Intelligence Research found that three dozen utilities have adopted new-large load tariffs, with several geared specifically toward data centers, and the group expects that trend to continue.

In July, American Electric Power’s Ohio utility introduced a load study tariff and began to charge between $10,000 and $100,000 to look at large load proposals, which it said slashed its large load forecast from 30 GW to 13 GW — though the Ohio Manufacturers’ Association says the utility is still inflating its forecast.

Dominion Energy, which serves the Northern Virginia area of high-density development called Data Center Alley, got approval from the State Corporation Commission in November to create a new GS-5 rate class, which starting on Jan. 1, 2027, will require data centers and other customers over 25 MW to sign 14-year contracts and pay a minimum of 85% of contracted distribution and transmission demand, along with 60% of generation demand.

“A company like Dominion is actually much better poised than a lot of other utilities, because they're in a vertically integrated territory where they've always been handling almost all parts of the grid process,” said Arun. “And in the process, they’ve developed the full stack of experience for understanding interconnection requests, for dealing with all different kinds of load classes.”

Smaller co-op utilities, Arun noted, generally lack this kind of experience. Those utilities — along with municipally owned utilities — have a different risk profile, Engstrom said, “because their customers are the shareholders.”

“If we think about the worst-case outcomes, the utility signs up with one of these customers who has a lot of money today, and they agree to pay these terms, and then any of these bad events happen, and the counterparty’s unable to pay,” Engstrom said. “Let's say that the gap there is $100 million or $500 million — is that the responsibility of the other customers, or is that the shareholders for the utility?”

Some small utilities “have talked about how the volume of what they're delivering to the data centers might literally double their size," he added.

One such utility is the Northern Virginia Electric Cooperative, which operates in parts of Data Center Alley and doesn’t own generation but purchases wholesale electricity through the PJM Interconnection.

In a January 2025 article in the Prince William Times, Gilbert Jaramillo, the co-op’s vice president for power supply, told the newspaper that by 2032, data center customers are expected to account for more than 95% of NOVEC’s energy sales. Jaramillo described this as “very concerning” but a “great opportunity for the rest of the membership as well.”

Since then, Dominion has entered talks to purchase NOVEC, Bloomberg reported in November. Dominion and NOVEC are already intertwined, as Dominion's transmission services are tied to NOVEC's substations dedicated to data centers.

The success of the bring-your-own-generation model for meeting data center demand is also influenced by the utility’s structure, Engstrom said.

“In some states, that works well where the utilities don't own the generation, at least in terms of negotiating with the utility,” he said. “In the states where they're still vertically integrated, that can create some conflict.”

In some ways, utilities hold significant power, Arun said, because “the growth of the power sector is the ultimate constraint on the actual realization of this bubble.”

“Utilities are able to impose lockout fees and create large load tariff structures for these interconnections,” he said. “And I think that will help rationalize for the rest of the system the kind of demand that we're actually expecting.”