OpenAI is recruiting a power-trading lead for its data centres. The role will own commodity hedging strategy across the company’s power portfolio. It will have no direct reports at first.
Bloomberg’s Julian Hast reported the listing on Monday. OpenAI did not immediately respond to a request for comment.
The company’s own job posting places the role on the Power & Land team, inside datacenter design. It pays $181,000 to $285,000 plus equity, and it is open in Seattle, San Francisco or remote.
What the job actually asks for
OpenAI wants at least 10 years in power trading, commodity risk management or utility strategy. It also wants a deep understanding of US wholesale power markets and the natural gas markets attached to them.
The work is recognisably a trading desk. The hire would quantify exposure across markets and sites, then evaluate fixed-price supply, forwards, swaps and options.
They would also build the frameworks for when to hedge and how much. Governance, risk limits and internal controls sit in the same job. Experience with ISDAs and supply contracts is listed as preferred.
OpenAI put the purpose in a single sentence. The role will “translate large, dynamic electricity and fuel exposures into practical hedging, procurement, and risk-management strategies that protect infrastructure economics while preserving flexibility for growth.”
This is ordinary, and that is the point
Hedging a commodity input is not exotic. Food producers hedge coffee and cocoa, airlines hedge jet fuel, manufacturers trade the metals they consume.
What has changed is the category OpenAI now sits in. Electricity has stopped being a line on a monthly bill. It is a market exposure large enough to need its own risk function.
Meta got there first. It said in November 2025 that it was entering power trading to support its AI data centres.
Microsoft and Google face the same constraint. Securing electricity has become one of the hardest limits on the buildout, and the answer increasingly looks like a trading function rather than a procurement one.
Gas is the exposure underneath
More than 40% of US electricity is generated by burning gas. That is why the posting asks for gas expertise alongside power. Hedging one means understanding the other.
Grid connection queues are long, so many data-centre developers are building their own generation while they wait. Meta is hiring for that too. Its careers site lists an energy manager for onsite generation, confirming that “gas supply, land, water, and permitting” are in place.
The environmental arithmetic follows. Meta left RE100, the corporate clean electricity pledge, as gas moved into its data-centre plans.
Why OpenAI needs this now
The committed footprint is what turns power into a risk-management problem. OpenAI is building a Georgia campus worth around $30bn. It has also been in talks over a 10-gigawatt Ohio campus backstopped by Nvidia.
Energy costs have already changed its plans once. The company paused a UK project over electricity prices and regulation.
Regulation is moving underneath all of it. US regulators have been working on a grid fast lane for large loads, which changes when and how these sites can connect.
A job ad is not a strategy
One caveat is worth keeping in view. This is a listing, not an announcement, and OpenAI has not said what it intends to hedge or at what scale.
The role carries no direct reports initially, which reads as a first hire rather than a desk. Whether it grows into one depends on how large the exposure turns out to be.
The salary band is still the tell. A company willing to pay $285,000 for someone fluent in ISDAs and forward curves has stopped thinking of electricity as a utility bill. It is a position on the books now, and positions have to be managed.