←News detail

Sina TechUpdated Nov 2 23:00

OpenAI pledges that its self-built U.S. data centers can cut power load during extreme weather

AI summary

OpenAI announced the launch of the "Grid Guardian Plan," under which data center campuses it directly builds in the United States will proactively reduce electricity consumption when extreme weather or other emergencies put stress on the grid. The plan extends to the entire country a commitment already made for a planned campus in Effingham County, Georgia; OpenAI pledges to cut up to 1 gigawatt of power load when the grid needs relief, roughly equivalent to the electricity consumption of 700,000 households. John McCarrick, OpenAI's head of energy policy, said this reflects a shift in industry thinking and technology, as well as adjustments made after incorporating community feedback.

Why read this

AI score 56

OpenAI is extending its data center power-curtailment pledge from a single campus to the entire United States, giving readers a window into the tug-of-war between AI compute expansion and grid regulation.

Topics

Tags

Body · AI translation

Related topics top start Related topics top end Futures promotion begin Futures promotion end Miaopai begin Miaopai end Video player start Video player end Market chart begin Market chart end

OpenAI has promised that at data center campuses it directly builds in the United States, it will proactively reduce electricity consumption when extreme weather or other emergencies cause pressure on the power grid.

Just as the high power consumption of AI computing centers triggers opposition from the public and political circles and drags down industry expansion plans, OpenAI is advancing its own data center construction.

News background: This initiative, called the "Grid Protectors Program," intends to extend to the whole United States the commitments already made for the planned data center campus in Effingham County, Georgia.

For this project, OpenAI has promised that when the grid needs relief, it can cut up to 1 gigawatt of electricity load, roughly equivalent to the electricity consumption of 700,000 households.

OpenAI will say in a blog post later today: "Our goal is to make this kind of load adjustment capability a core component of the construction and operation of future data center campuses."

The blog post says: "This means... we will negotiate with power companies to determine the scale of load that can be cut, the response speed, and the duration."

Specific plans vary from campus to campus, and details will be announced separately once finalized.

Related statement: John McCarrick, OpenAI's head of energy policy, said the move reflects a shift in industry thinking and technology, as well as adjustments made after incorporating community feedback.

He said: "About a year ago, I met with another major cloud provider, and they were still clinging to the five-nines reliability standard, but now no one insists on that anymore." "Five nines" means data centers must achieve 99.999% continuous uptime.

AI and technology giants, including OpenAI, have successively made commitments to be grid-friendly users and avoid driving up electricity bills for ordinary residential users.

OpenAI frames Monday's announcement as moving from a visionary commitment to a set of actionable operating rules.

OpenAI also argues that data centers can actually benefit ordinary electricity users.

Follow-up observation directions: Policymakers have pushed in multiple ways for data centers to have electricity demand flexibility.

The U.S. Federal Energy Regulatory Commission (FERC) requires grid operators to explain or reform the interconnection process for large customers, and the relevant provisions encourage this kind of flexible adjustment.

The Senate permitting bill explicitly allows states to make "power restrictions during peak hours" a precondition for data centers to connect to the grid. news_keyword_pub,usstock, iframe below main text iframe below main text General statement General statement QR code at end start QR code at end start

QR code at end start QR code at end start Editor name and work code

Editor name and work code

Source

Sina Tech