The wholesale electricity market has a new price signal, and it's coming from the interconnection queue, not the trading floor.
PJM Interconnection — the grid operator covering 65 million people from New Jersey to Illinois — filed a proposal with FERC that reframes the fundamental relationship between hyperscale data centers and the grid. The framework, filed August 13 and targeting an October 12, 2026 effective date, tells any new load of 50 MW or more the same thing: bring qualifying new capacity to cover your peak demand, or face curtailment before everyone else when the grid gets tight. PJM calls it "Bring Your Own New Capacity" — BYONC — and it's the most consequential structural change to wholesale market design in years.
The implications for power purchase agreements and wholesale pricing are enormous. And they're already moving.
The BYONC Mechanism Restructures the Demand Side of the Market
Here's the mechanism. Under PJM's proposal, any "New Large Load" — defined as a facility with cumulative peak demand of at least 50 MW at a single electrical site — must designate qualifying new capacity equal to or greater than its peak demand. Qualifying resources include new generation, storage, certain repowered assets, fuel conversions, and demand resource aggregations. A load that covers its full registered peak faces no curtailment exposure. A load that doesn't gets cut first, before PJM even calls pre-emergency demand response.
This is a capacity adequacy rule dressed up as a reliability framework — and it's brilliant market design. By making curtailment risk the penalty for not procuring new capacity, PJM has created a direct financial incentive for hyperscalers to go out and contract for generation. That means more PPAs, more behind-the-meter generation deals, and more pressure on the wholesale capacity market to price the scarcity that data centers are creating.
The FERC chairman agrees the grid can't keep up. FERC Chair Laura Swett told CNBC on September 14 that the U.S. power grid still lags surging AI data center demand and that urgent regulatory reforms are needed. When the top federal grid regulator is calling the situation urgent, the market is already behind.
The Demand Signal Is Real — and Being Revised in Real Time
The scale of what's coming at the grid is staggering. Data Center Dynamics cites BloombergNEF projections that U.S. data center electricity demand could reach 106 GW by 2035, up from roughly 25 GW in 2024. The IEA expects global data center electricity use to more than double to roughly 945 TWh by 2030. McKinsey pegs U.S. data center consumption at 606 TWh by 2030, up from about 147 TWh in 2023.
But the signal is getting noisier, not cleaner. Bloomberg reported that the EIA's 2027 electricity sales growth forecast was cut from 2.9% to 1.8% after Texas paused new data center development. I covered the Texas situation in the September 2 issue — the 474 GW fiction collapsing under grid physics. The point stands: announced data center capacity and actual connected load are very different numbers, and the market is starting to price that distinction.
This is exactly why BYONC matters. It forces the distinction into the open. A data center that wants to be in PJM's territory and draw 200 MW has to go find 200 MW of qualifying new generation — not announce it, not put it in a queue, but actually contract for it. That's a fundamentally different demand signal than a letter of intent.
Virtual Power Plants Are the Emerging Third Path
There's a quieter structural shift happening alongside BYONC that deserves attention. RMI published analysis on September 10 showing that new state laws, commission decisions, and utility proposals are creating pathways for large electricity customers — including data centers — to fund distributed capacity: demand flexibility, distributed generation, and energy efficiency aggregated into virtual power plants. Several states have moved since RMI's November 2025 report on the subject.
The design of these pathways matters enormously for wholesale pricing. If a hyperscaler can satisfy a capacity requirement through a VPP aggregation rather than a dedicated gas peaker or nuclear PPA, the marginal cost of compliance drops — and so does the upward pressure on wholesale capacity prices. But if VPP pathways remain narrow or administratively complex, the pressure concentrates on conventional generation procurement, which is exactly where supply is tightest.
Watch for FERC's response to PJM's October 12 filing deadline. If FERC approves BYONC largely as filed, expect a wave of new PPA announcements from hyperscalers in PJM territory through Q1 2027 — and a corresponding move in wholesale capacity auction clearing prices. The data centers that move first on qualifying capacity lock in better economics. The ones that wait face a market that's already priced their delay.
The future is electric. But PJM just made clear that the future belongs to whoever shows up with electrons in hand.
