Skip to content
Friday, October 23, 2026 · Today’s edition, 6:30 a.m. ETFri, Oct 23 · 6:30 a.m. editionSign inMy Desk 0EditionsSign in

Mistral hires Google energy lead for 1GW push

The veteran PPA closer joins Mistral Compute just as its €3 billion war chest meets Europe's grid bottleneck.

At a glance

  1. On LinkedIn, Oman framed the move as a fight over who shapes Europe's AI future: "European tech sovereignty isn't just a goal, it's a necessity."
  2. The move comes the same week Mistral raised €3 billion ($3.48 billion) in Series D financing and said it would spend a large portion on compute.
  3. The hiring marks the point at which power procurement moved from a back-office chore to a boardroom concern inside AI companies.

Marc Oman, the Google executive who ran energy sourcing strategy for the company's European data centers for almost 12 years, is joining Mistral Compute, the newly formed division of French AI company Mistral, and will be based at the firm's new Paris headquarters to help execute the plan it laid out in August: deploy more than 1GW of compute capacity in Europe by 2030.

On LinkedIn, Oman framed the move as a fight over who shapes Europe's AI future: "European tech sovereignty isn't just a goal, it's a necessity." He said he personally led Google's closing of more than 2GW of solar and wind power purchase agreements across Europe, and that he was honored by the trust Timothee Lacroix and Mistral's leadership team had placed in him.

The move comes the same week Mistral raised €3 billion ($3.48 billion) in Series D financing and said it would spend a large portion on compute. Mistral Compute is the new unit charged with providing private high-performance computing resources for training and running AI models.

The hiring marks the point at which power procurement moved from a back-office chore to a boardroom concern inside AI companies. Oman spent his Google years matching data center load to renewables PPAs and navigating grid access, exactly the work Mistral will need as it deploys that capacity; he built that skill set at companies like Google rather than at model labs, and it is the kind of experience that turns a funding round into megawatts.

It is also the latest high-level exit from hyperscaler infrastructure ranks: OpenAI's data center chief exited in late August as its $750 billion buildout churns. The two moves pull in different directions—one out, one in—but they point to the same conclusion: the people who can connect AI to the grid are now as contested as the people who train it.

The new hire is the purest form of a developing market rule: grid access is becoming a customer class, not a queue. Google has already put a price on duration with its West Virginia storage deal; Mistral is now buying the people who can execute on that same logic. The €3 billion will be easier to raise than the megawatts will be to connect—and the new hire is the firm's answer to that bottleneck.

Reported by the Data centers & fiber desk from Data Center Dynamics, PID archive: OpenAI's data center chief exits, PID archive: Google puts a price on duration and PID archive: Grid access becomes a customer class. Published . Corrections

Related

Infrastructure

OpenAI hires a power screen for its buildout

A single remote hire, capped at $385,000, will decide which power options and developer pitches survive contact with OpenAI's construction schedule.

Infrastructure

Anthropic hires Google's data center design chief to lead EMEA capacity

John M. Wilson spent six years at Google and previously ran European design and construction at Amazon.

Infrastructure

Crusoe files for a $4.8bn data center campus in Jayton, Texas

The two buildings are slated to go live in 2029, and the filings name no tenant, landowner or power source.

Infrastructure

O3 Group plans Benghazi data center for March 2027

The Libyan operator gave no capital figure, capacity or named customer; international connectivity would use the Medusa cable.

Infrastructure

Meta has converted Ohio Facebook site into first phase of AI campus

The New Albany site is planned to scale to a gigawatt with roughly one million GPUs; it has already trained Meta's Muse models.

Private Infrastructure Dailyfrom Private Wealth Daily

Energy transition, digital infrastructure and core-plus funds. Weekdays at 7:25 a.m. ET.

Free. Weekday mornings. Unsubscribe from any edition in one click.

From other sections