The mechanism. On December 18, 2025, FERC issued a unanimous order finding PJM's tariff "unjust and unreasonable" on two fronts that matter enormously to merchant nuclear: how co-located load gets priced, and how behind-the-meter generation (BTMG) is treated in resource adequacy. FERC didn't ban co-location — Commissioner Rosner's concurrence explicitly frames it as clearing a path for deals that "lower grid costs." But the order also directs that a generator can't pull capacity off the grid to serve a data center until all transmission upgrades needed to preserve reliability for everyone else are built — with 100% of that cost allocated to the generator, not spread to ratepayers. That single allocation rule is the whole ballgame. The Amazon-Talen Susquehanna arrangement drew a formal challenge from Exelon and AEP alleging roughly $140 million a year in transmission cost-shifting onto other PJM customers. FERC's compliance directive to PJM (filings due through February 2026) is designed to close exactly that gap. If nuclear operators must now fund their own interconnection upgrades to keep serving legacy grid customers while also carving out behind-the-meter capacity, the arbitrage that made these deals so lucrative — sell clean, firm power to a hyperscaler at a premium, dodge market/capacity pricing — gets a lot less clean.

Who cashes in. GE Vernova (GEV) wins regardless of how the cost allocation shakes out: someone has to build the grid-scale transformers, switchgear, and transmission upgrades that FERC is now mandating generators pay for directly, and GEV's backlog only grows as compliance deadlines force utilities to act. Vistra (VST) is comparatively insulated versus pure-nuclear peers — its fleet is diversified across gas, nuclear, and renewables/storage, so a nuclear-specific BTMG squeeze dents a smaller share of earnings. BWXT benefits on the supply side almost independent of the ruling — it makes reactor components and naval/medical isotopes, not merchant power, so it's a picks-and-shovels beneficiary of the broader nuclear buildout without co-location price risk.