The New Grok Times

The news. The narrative. The timeline.

Technology

FERC Co-Location Docket Decides Who Pays To Power AI Data Centers

Engineers study a one-line diagram of a plant feeding an adjacent data center.
New Grok Times
TL;DR

X frames AI data centers as national supremacy or grid parasites; FERC's co-location docket asks the colder question of who pays when a data center plugs straight into a power plant.

MSM Perspective

MSM such as Bloomberg and the trade press cover demand growth and blackout risk, less the cost-allocation fight inside the docket.

X Perspective

X frames co-located AI data centers as either the engine of national AI supremacy or a private drain on a public grid.

The fight over powering artificial intelligence has narrowed to a question an engineer can answer: if a data center plugs directly into a power plant, who pays for the grid everyone else still shares?

That question has a docket. FERC opened Docket AD24-11 to examine large loads co-located at generating facilities — the arrangement where a hyperscale data center sits beside a nuclear or gas plant and draws power behind the meter, bypassing the transmission system the rest of the market funds. The commission convened a technical conference on the generic issues the arrangement raises. [1]

The Federal Register notice frames it without the slogans. The conference was called to discuss the reliability, cost, and market implications of co-location — whether a load that leans on the grid for backup and stability should pay toward it, and what happens to other customers if a plant's output is redirected to a single private campus. [2] These are allocation questions, not applause lines.

On X, a co-located data center is either the cornerstone of national AI supremacy or a parasite siphoning public power for private compute. Mainstream coverage — Bloomberg, the trade press — tends to report the demand curve and the blackout warnings. FERC's record asks the prior question both skip: who bears the cost and the reliability risk when a gigawatt of load attaches itself to a single generator.

The venue matters because the money does. The loudest co-location disputes have landed in PJM, the largest U.S. grid operator, where proposals to route plant output straight to data centers have forced regulators to decide what "using the grid" even means when a customer claims not to. [3] The answer sets the precedent for every plant-plus-data-center deal that follows.

This is the divergence the paper keeps. The public argument is about whether AI deserves the power. The institutional question is narrower and more consequential: when a data center co-locates, does it pay its share of the system that keeps it running, or does the cost land on households and factories that never signed up to subsidize a server farm? [1][2]

A docket is a receipt. If a developer wants to argue that co-location is efficient and self-contained, the claim belongs in the record, where staff and rivals can test the load-flow assumptions. If a utility wants to argue that co-location offloads cost onto captive ratepayers, that case belongs there too. [3]

The thread will declare a winner tonight. FERC will decide who pays, and that answer will outlast the thread by a decade. [2]

-- THEO KAPLAN, San Francisco

Sources & X Posts

News Sources
[1] https://www.ferc.gov/news-events/events/commissioner-led-technical-conference-regarding-large-loads-co-located
[2] https://www.federalregister.gov/documents/2024/09/17/2024-21022/large-loads-co-located-at-generating-facilities-second-supplemental-notice-of-commissioner-led
[3] https://www.ferc.gov/industries-data/electric/electric-power-markets/pjm
X Posts
[4] FERC has ordered regional operators to implement large load interconnection reform, fast-tracking data center grid access while forcing tech developers to pay for their own infrastructure upgrades https://x.com/Shalemag/status/2069828920703057967

Get the New Grok Times in your inbox

A weekly digest of the stories shaping the timeline — delivered every edition.

No spam. Unsubscribe anytime.