Members of America’s Largest Power Grid Can’t Agree on How to Power Data Centers

After months of wrangling over how to handle surging power demand from AI-driven data centers and the resulting spike in electricity prices for millions, members of PJM Interconnection, the nation’s largest grid operator, could not agree on a solution.

Their disagreement played out last week in an “advisory vote,” part of the final phase of PJM’s expedited rulemaking track known as the Critical Issue Fast Path, or CIFP.

CIFP is aimed at figuring out how PJM can reliably serve the region as data center demand climbs toward a staggering 30 gigawatts by 2030—roughly the power use of 20 million homes. Early stages of that growth have already driven double-digit percentage increases in electricity prices across states, notably in New Jersey, where the state saw a 20 percent bill increase in the summer that became a hot-button issue in its recent gubernatorial race.

The Garden State saw that frustration show up in polling, with a recent statewide survey finding large majorities of voters in both parties saying data centers should shoulder more of the grid costs instead of ordinary households.

Meanwhile, energy projects that could add the needed gigawatts of energy onto the grid wait for years in the PJM interconnection queue.

During the meeting, PJM members—hundreds of representatives, most from the energy industry—reviewed a final slate of a dozen proposals and cast advisory votes on each of them. The options ranged from requiring or nudging data centers to bring their own generation, to creating a new fast track for connecting energy projects to the grid, to temporarily halting new data-center hookups altogether until PJM can reliably serve them.

Any plan that won at least two-thirds approval was deemed endorsed, a non-binding label the PJM Board can point to in making the final decision on a package of rule changes it will file with the Federal Energy Regulatory Commission (FERC).

None of the proposals reached the two-thirds threshold.

It also handed PJM’s 10-member board of managers virtually a blank slate to craft a policy of their own or pick among the proposals. PJM Interconnection operates a competitive market for wholesale electricity in all or parts of Delaware, Illinois, Indiana, Kentucky, Maryland, Michigan, New Jersey, North Carolina, Ohio, Pennsylvania, Tennessee, Virginia, West Virginia and the District of Columbia.

Mark C. Christie, a recent chairman at FERC, characterized PJM members as representative mostly of the interests of an energy industry that gets to “decide which rules they want to live under” rather than average bill-paying consumers.

“[I]t is really just an insiders’ game and always has been. Consumers, especially residential, have no serious representation or influence in this ‘inclusive’ process,” Christie wrote on LinkedIn.

Tougher proposals could cap the revenue of many energy companies and could push more risk onto them instead of ratepayers.

Independent energy market monitor Joseph Bowring has put forward a notably strict but straightforward idea: Data centers shouldn’t be allowed to build until PJM can guarantee reliable power for them.

“I’m quite sure the data centers do not want to be added quickly and unreliably,” said Bowring, the president of Monitoring Analytics, an independent monitoring unit that reports on PJM’s electric power market.

Another notable proposal came from a coalition of governors from Pennsylvania, Maryland, New Jersey and Virginia, joined by the Data Center Coalition (DCC), whose members include tech giants such as Google, Microsoft and Amazon. They urge states to entice data centers to supply their own power in exchange for priority in permitting and siting—the main bottlenecks for these projects.

A recent proposal from bipartisan lawmakers across the PJM region nodded to the DCC governors proposal but included a policy that in times when power would be spread thin, like during a heatwave or a hurricane, data centers would have to forego power unless they brought their own generators into the system to increase reliability for everyone. When their power is cut, data centers usually fire up diesel backup generators that pollute the air in nearby neighborhoods.

In all these proposals, powering data centers would need rapid development of energy generators, most likely natural gas plants, according to energy experts. But Julia Kortrey, a deputy director of state policy at Evergreen Action, a climate advocacy group, urged PJM members to look at Texas as a model for swift deployment of renewables.

“We should be getting as much energy as we possibly can right now, and that would still include renewables,” said Kortrey.

The PJM board is expected to spell out the details of its own solution in December. It seeks to sign off a package of rule changes to file with FERC before the end of the year. FERC will then decide whether the proposal is just, reasonable and not unduly discriminatory under federal law.

4 Likes

Why aren’t we charging data centers more for the electricity they want? We know who they are (or will be) and we can let them know prior to construction!

3 Likes

@jaagu copying more than a few small snips from an article is a copyright violation. That’s why I only post a link, the author, and at most 2 or 3 paragraphs.

If your post is a large copy TMF could remove the entire post. They’ve done it to me.

Wendy

2 Likes

Missouri just passed legislation requiring data centers agree to pay premium prices for power.

CME (Chicago Merchantile Exchange) just reported an extended outage due to cooling failure in their data center. They had sold the data center they built to private equity who failed to maintain it. So now switching to cloud based data center.

4 Likes

With all due respect Wendy have you ever seen how much you place in your posts?

1 Like

I never copy a complete article from start to finish. I try to take the most significant snippets from within the articles. Usually 3 or 4. Never the entire article from start to finish. This takes a lot of time. Try it sometimes – it’s not easy.

Wendy

5 Likes

Had to rec that one. I do love Wendy’s snark (even when directed at me!).

Pete

1 Like

How about posting the AI version below:

Members of the PJM Interconnection, the largest US grid operator, have failed to reach an agreement on how to manage the massive surge in power demand from AI-driven data centers, a situation that has already led to double-digit electricity price increases for millions of customers. This disagreement, which played out in advisory votes, means the PJM Board must now unilaterally propose a solution to federal regulators.

The Core Problem: Demand Outpacing Supply

PJM’s region, which includes the “Data Center Alley” in Northern Virginia, is experiencing unprecedented demand growth, projected to reach 30 gigawatts by 2030 (enough to power 20 million homes). The issue is that data centers can be built in 18 to 24 months, while new power plants often take many years (sometimes eight years for new generation to come online in PJM). This mismatch has led to:

  • Higher Bills: Skyrocketing capacity prices in recent auctions, with the total capacity bill for ratepayers jumping from $2.2 billion to over $16 billion in two years, largely due to data center forecasts.

  • Reliability Concerns: The grid is facing tightening supply conditions and a potential risk of falling below reliability standards as early as June 2027 if the situation isn’t addressed.

  • Interconnection Bottlenecks: A backlog of projects waiting to connect to the grid, slowing the deployment of new power sources.

Key Areas of Disagreement

During the Critical Issue Fast Path (CIFP) stakeholder process, no single proposal garnered the two-thirds support needed for adoption. The main points of contention include:

  • Who Pays for Upgrades: How the costs of building new transmission lines and system upgrades should be shared among data center companies, utilities, and retail customers.

  • Reliability Requirements: A key proposal from PJM’s independent market monitor, Monitoring Analytics, argues that data centers should only be allowed to connect if they can be reliably served, potentially requiring them to wait in a queue or bring their own power.

  • Backup Power: A suggestion that data centers without their own firm power supply be the first to face blackouts during emergencies. Data center operators like Meta have opposed mandatory curtailment, citing operational requirements.

  • Energy Sources: Debate over whether new power sources should be natural gas plants (suggested by some energy experts as the fastest option) or whether PJM should prioritize a swift deployment of renewables.

Next Steps

The PJM Board of Managers is expected to develop its own reform package based on areas of consensus and file it with the Federal Energy Regulatory Commission (FERC) by the end of December. Monitoring Analytics has already filed a formal complaint with FERC, asking the commission to rule that PJM can only add new loads when they can be reliably served. FERC will ultimately decide whether PJM’s final plan is “just, reasonable and not unduly discriminatory” under federal law.

Yes they have done it to me. Mostly because DB2 flags my posts that he does not like.

An uneasy relationship

The artificial-intelligence boom’s promise of runaway electricity demand has jolted shares of US power companies to all-time highs. But those generators and utilities are now learning that the hype comes with an edge: Investors won’t wait forever for results.

Companies that recently hit record valuations are returning to earth as investors realize the massive data-center deals they’d banked on are actually smaller, or slower, than expected.

Constellation Energy Corp. saw shares tumble 11% from an October high after a third-quarter earnings call that yielded no details on new power generation. The headline of one Jefferies analyst note read: “No Data Center Deals.” Similarly, Vistra Corp. has fallen 16% since mid-October as analysts noted a slower pace of data center announcements than they expected.

The S&P 500 Utilities Index is now set to post the worst monthly performance since August, after reaching an all-time high in October.

“These are not your father’s and mother’s utility stocks,” said Mark Malek, chief investment officer at Muriel Siebert & Co. “People are starting to question, ‘Can these companies scale as fast as they want to, are they throwing capital at these projects that’ll never get done?’”

1 Like

Wrongly credited to Mark Twain (*it was actually said by another guy in the 17th century)

“I didn’t have time to write it shorter.”

I know the feeling.

1 Like

How about posting the human intelligence version instead?

DB2

How ironic (maybe intentional humor?). jaagu uses AI to point out the unsustainable path of AI. The great AI power crunch is ironic. AI is generally associated with problem-solving. One might expect the technology’s development to be smoothly integrated and managed with foresight. Instead, the rapid physical infrastructure demands of AI has created a massive, immediate problem (a “massive surge in power demand”) that human decision-makers (“Members of the PJM Interconnection”) are failing to solve, leading to real-world negative consequences for millions of people (“double-digit electricity price increases”).

In the long term, AI data centers could potentially decrease electricity prices for the average customer by spreading fixed costs over a larger consumer base. And the AI winner might be determined by who builds the largest energy supply.

“Residential retail electricity prices in September were up 7.4%, to about 18 cents per kilowatt hour… The basic reason for rising prices: Electricity demand — including actual and forecasted demand — is outstripping new supply. Data centers are expected to consume anywhere from 6.7% to 12% of total U.S. electricity by 2028, up from 4.4% in 2023”

Average Price: Electricity per Kilowatt-Hour in U.S. City Average

“[Buffett] warns that AI’s exponential appetite for power could soon outstrip available capacity, suggesting that the tipping point might arrive as early as mid-2026… [Zuckerberg] said: “Before we hit capital constraints, we’ll run into energy constraints.” … “Every new model demands exponentially more compute, and exponentially more electricity,” [Jassy] told investors, hinting at what could become one of the largest infrastructure challenges of the decade. … Huang told Bloomberg that the industry is entering a “power-limited” era. “The world needs to build energy infrastructure as fast as it builds AI infrastructure,” he said. “Every single data centre in the future will be power-limited. Your revenues will associate with the power you have to work with.””

1 Like

The real price of electricity dropped about 30% between 1985 and 2000, and has been in a +/-10% range since 2000. The real price is now near the top of the range, with a breakout higher expected because of the great AI power surge.

1 Like

Already did that in the OP. I guess you missed the OP.

Didn’t miss it (and the discussion about copyright and intellectual property).

DB2

Trump’s Plan For AI Dominance Threatened by His Own Attacks on Solar, Wind Power

  • The Trump administration is moving to fast-track the construction of power-hungry data centers as a matter of national security, while adding roadblocks for new solar and wind farms.

  • Hindering renewable energy projects risks slowing the AI boom and could exacerbate rising electricity prices, according to data.

  • Renewable energy remains the fastest and cheapest option to add power to the grid, with nearly 80% of planned power plant capacity tied to renewable sources, according to filings with federal regulators and grid operators.

The Trump administration is moving to fast-track the construction of power-hungry data centers as a matter of national security. At the same time, it’s adding roadblocks for new solar and wind farms.

But the two policies could be at odds: Hindering renewable energy projects risks slowing the AI boom — and could exacerbate rising electricity prices, a slew of data suggests.

“It’s an all-hands-on-deck moment right now to get the power to supply this,” said Robert Whaley, director of North American power at Wood Mackenzie, an energy consultancy. “In the next 10 years, there’s really nothing to replace renewables.”

Renewable energy so far remains the fastest and cheapest option to add power to the grid. Nearly 80% of the planned power plant capacity in the pipeline is tied to renewable sources, according to filings with federal regulators and grid operators compiled by Cleanview.co, an energy data company.

Planned Battery, Solar and Wind Power Plants

More than 90% of power plants in the pipeline are tied to renewable fuel sources.

Planned Coal, Gas and Nuclear Power Plants

Less than one in 10 power plants in the pipeline are tied to traditional fuel sources.

https://www.bloomberg.com/news/articles/2025-12-04/trump-s-ai-dominance-plans-threatened-by-his-own-attacks-on-solar-wind?cmpid=BBD120425_GREENDAILY&utm_medium=email&utm_source=newsletter&utm_term=251204&utm_campaign=greendaily&embedded-checkout=true

4 Likes