The Stargate Shift: Why OpenAI’s Top Compute Minds are Moving to Meta’s Superintelligence Labs

Meta Platforms has hired three key infrastructure leaders from OpenAI as it intensifies efforts to build the massive computing capacity required for advanced artificial intelligence, ET reported.

The move highlights Meta’s aggressive push to attract talent focused on AI data centre expansion, even as OpenAI’s own infrastructure plans face rising costs and strategic adjustments.

The three hires include Peter Hoeschele, who played a central role in OpenAI’s Stargate initiative, Shamez Hemani, who worked on computing strategy and business development, and Anuj Saharan, a leader within OpenAI’s computing organisation.

Reason Behind Departure

Their departures are notable, as all three were closely involved in OpenAI’s efforts to secure and scale hundreds of billions of dollars in AI data centre capacity.

Meta, led by Mark Zuckerberg, has been investing heavily to build the compute infrastructure needed to support its AI ambitions, including initiatives such as Meta Superintelligence Labs, which is aimed at keeping pace with competitors.

Zuckerberg has indicated that Meta could spend up to $135 billion in capital expenditure this year alone, with a significant portion allocated to AI infrastructure, and has committed to investing hundreds of billions more before the end of the decade.

OpenAI’s Reaction

OpenAI has not publicly commented on the hires, although it has previously expressed appreciation for the departing employees and reiterated its focus on expanding its own infrastructure capabilities.

The company also appointed Sachin Katti, a former Intel executive, in November to lead its industrial compute efforts. OpenAI has maintained that its early investments in computing infrastructure provide it with a competitive advantage over rivals such as Anthropic, citing projects like its site in Abilene as evidence of its head start.

Stargate Project

The Stargate project itself has evolved over time. Originally announced at the White House as a $500 billion initiative involving OpenAI, Oracle, and SoftBank Group, it has since become a broader umbrella for OpenAI’s data centre ambitions.

However, the initiative is now being scaled back in certain areas. OpenAI recently said it is pausing its Stargate infrastructure project in the United Kingdom as it reins in spending ahead of a potential public listing.

Additionally, OpenAI and Oracle have decided not to proceed with an expansion at the Abilene site, signalling that even the most ambitious AI infrastructure plans are being recalibrated as capital requirements continue to rise.