Three OpenAI Stargate Executives Move to Meta Amid the AI Race

Three senior OpenAI executives who previously worked on Stargate, a program to build AI data centers costing hundreds of billions of dollars, have joined Meta Platforms (Meta is designated an extremist organization and its activities are banned in Russia). The move coincides with a sharp increase in Meta's AI infrastructure spending and with a review of some Stargate projects inside OpenAI itself, 3DNews reports.

What is the Stargate project?

Stargate is OpenAI's large-scale infrastructure program to build data centers designed specifically for training and running frontier AI models, with a stated budget of hundreds of billions of dollars. It is projects like this, rather than the models themselves, that now determine which company can train next-generation models faster than its rivals: access to computing power has become as strategic a resource as access to data once was. Executives who delivered such a program from inside OpenAI have rare expertise, from negotiating with energy and hardware suppliers to managing multibillion-dollar capital spending at specific construction sites.

Several significant departures from a single program rarely happen by chance at a company of OpenAI's size. Such waves usually point either to dissatisfaction with the pace or terms of specific projects, or to a deliberate poaching effort by a competitor willing to offer better terms to specialists who have already proven themselves on a task of comparable complexity.

In recent years the AI industry has often been described as living through an unprecedented talent race: the largest tech companies offer specialists with experience in leading AI projects compensation packages that were previously the norm only for top executives of public companies. The move of infrastructure executives, not just model researchers, shows that this race has spread to those responsible for the physical side of the AI industry: power supply, construction and operation of data centers.

• Who: three OpenAI executives involved in the Stargate project.
• New employer: Meta Platforms.
• Stargate: a program to build AI data centers costing hundreds of billions of dollars.
• Coinciding events: a rise in Meta's AI infrastructure spending and a review of some Stargate projects inside OpenAI.
• Source: 3DNews.

What is behind the reshuffle?

That infrastructure executives moved, rather than model researchers, is telling: it suggests competition among the largest AI companies is shifting further toward physical infrastructure, meaning the construction and operation of data centers, and not just algorithm development. The timing alongside the review of some Stargate projects inside OpenAI can be read two ways: either the company itself adjusted its priorities for certain sites, or the loss of these specialists was one reason some plans had to be revisited.

The article does not disclose which positions the executives took or which specific Stargate projects were reviewed. Still, the simultaneous outflow of talent and revision of infrastructure plans is already a talking point in an industry where every such executive move is traditionally treated as a signal of a company's internal priorities.

What does this mean for the AI infrastructure race?

Heavy investment in their own data centers has become a way for the largest AI companies to hedge against a shortage of computing capacity, which in recent years has slowed model training for every player in the market. Hiring experienced leaders of such programs from a direct competitor lets the new employer cut the time needed to build its own expertise from scratch: instead of learning from its own mistakes, it gets a ready-made team that has already walked this path inside one of the industry's most ambitious infrastructure projects. For OpenAI, losing these people just as Stargate plans are being reviewed may mean extra difficulty in delivering the rest of the program.

For the AI infrastructure labor market, such moves also have a demonstration effect: they signal to other specialists that experience on megaprojects like Stargate converts into offers from direct competitors, which further raises the already high price of this expertise across the labor market.