A 15-year guarantee? Inside the "Emergency" Capacity Auction

The "Power Game" is shifting. Here’s what you need to know from this week’s Energy Story:

• Courts 3, White House 0: Three different federal judges have now lifted the "stop-work" orders on massive offshore wind projects, ruling that the administration failed to prove any urgent "national security" risk.
• A "Social Good" vs. A Commodity: PJM took the rare step of siding against the administration, arguing that stopping these projects causes "irreparable harm" to the reliability of the grid for 67 million people.
• The 15-Year Hook: A new bipartisan proposal suggests an "Emergency Capacity Auction" specifically for data centers. It would offer developers 15-year guaranteed revenue—a massive shift from the current (and often "useless") 1-year auction cycles.
• The "Parallel Grid" Risk: We explore the danger of creating two markets: a highly lucrative one for AI developers and a "starved" one for existing ratepayers.
• The 2028 Bottleneck: Even with guaranteed money, the world is running out of hardware. GE reports gas turbine availability is limited until late 2028, and new transmission capacity is essentially non-existent.

Chapters

• 0:03 Judges Reverse Offshore Wind Halt
• 0:31 PJM Steps In And Why It Matters
• 1:40 White House Push For Data Center Auction
• 2:26 Supply Constraints And Grid Risks
• 3:38 Two Smart Ideas In A Flawed Plan
• 4:31 Capacity Signals, History, And Next Steps
• 5:26 Politics, PJM, And What’s Next

Transcript

SPEAKER_00: 0:03

I've got your energy stories for this, the fourth week of January 2026. And just like that, three different federal judges overruled the Trump administration's offshore windstop work order, allowing that work to resume. In the latest decision last Friday, January 16th, a judge granted Dominion's two-plus Gigawatt offshore project a preliminary injunction that lets it continue work while a lawsuit challenging the federal order proceeds. The federal action from the Trump administration has cited an unspecified national security risk, but the three judges in their decisions pushed back, essentially saying the same thing, that the government had failed to demonstrate that any national security risk was so urgent that construction should cease, and that the project's economics were being harmed as a consequence of these government actions. Particularly fascinating in the Dominion case is that Mid-Atlantic regional grid operator PJM took the highly unusual step of filing an amicus brief, citing the need for the project's output and describing it as an integral component of needed new generation that PGM has been relying upon that can help mitigate the region's capacity constraints. The grid operator also noted that with long lead times for alternative new projects, the extended delay of the offshore wind project would cause, quote, irreparable harm to the 67 million Americans served by PJM. It also noted that there are national security benefits in the form of a stronger and more reliable electric grid, so take that. The grid operator's actions are laudable, especially given the pressure it is under these days, combined with the political risk of antagonizing the White House. Last week at a White House meeting, the administration and a bipartisan group of governors met and they called on PGM to schedule a one-time emergency capacity auction that would dedicate supply resources for 15 years solely to new data centers, this in an effort to minimize costs affecting other ratepayers after we've seen these amazingly high three recent capacity auctions. The White House event didn't include any data company officials, nor did it include PJM. The administration claimed that this new approach would bring$15 billion worth of new power projects online and thus inoculate other ratepayers from cost pressures. However, it noted it can't compel PJM to take this action, which is probably a good thing if you step back for one minute and think about it. This approach would now create two parallel auction structures, one out-of-market approach that would be highly lucrative for supply asset developers with guaranteed 15-year take or pay revenues, and the other one pretty much status quo. But there's a problem. Supply is constrained and finite. For a GE's most recent investor presentation, gas turbine availability is limited until after 2028. Meanwhile, interconnection locations on the transmission grid are finite, and almost no new transmission capacities being built. So you may find developers rushing to provision new data-related load while the existing system gets starved for additional supply since it's less desirable to serve it. And who potentially gets hurt then? Everybody else. That's who, both in terms of costs and overall grid reliability. The challenge with our grid is that it's a system of systems interwoven with our regional economy and with global supply chains. Plus, electrons have their own unfortunate pesky way of being dictated by physics, so power flows can also be impacted, and the unintended consequences of these actions could result in additional effects on grid infrastructure costs and reliability. That said, there are two features I like a lot in the proposed approach. First, it focuses on new resources. ISO New England used to have a separate price tranche for new resources, and it always made sense to me that one should have a price to entice new entrants versus a price to keep on paying for existing steel on the ground, where a single-year marginal cost signal otherwise isn't going to do much. Second, this approach proposes a capacity price signal of 15 years. So investors can actually determine if a new investment is going to pay off. Will it pencil? By contrast, today's annual prices that vary year to year are virtually useless. Back when I was heading up Constellation's demand response group, we couldn't plan on revenues more than one year out. Some years the auction would bring us a windfall and nice bonuses, and the next year we get badly torched and miss our commitments, and of course, bonuses could be ugly those years. The design was and is in my opinion ill-advised. If the point of capacity markets is to pay for reliability, which in itself is more of a social good than a commodity, then perhaps it's time to send a price signal that matters and helps guarantee reliability. That would allow developers of new power plants as well as virtual power plants to get much more visibility into the potential payoff of their investments. ISO New England had such a capacity lock-in price for years, paying up to seven years of known prices. But the FERC directed them to eliminate that approach in late 2020. At that time, analysts suggested the elimination of this approach could hurt developers of new gas generation, and they were probably right. It's time to reconsider this approach more broadly. The Federal Energy Regulatory Commission will still need to bless this new emergency measure coming out of the White House and these governors, and nothing's going to happen anytime in the immediate future. In the interim, though, one can expect the PGM and its capacity issue will remain right in the middle of the political crosshairs, and that won't change anytime soon. Well, thanks for watching, and we'll see you again soon.