🧔♂️ A friendly human may check it before it goes live. More news here
Meta brings in ex-OpenAI Stargate team for AI expansion
Meta is set to hire OpenAI infrastructure figures Peter Hoeschele, Shamez Hemani, and Anuj Saharan as it expands AI data center capacity, according to people familiar with the matter.
The three worked on OpenAI’s Stargate program and related computing strategy, with Hoeschele helping lead infrastructure work at the company’s Abilene site in Texas.
The hires add to Meta chief Mark Zuckerberg’s AI buildout after the company projected up to US$135 billion in capital spending this year for AI infrastructure.
The move also follows shifts in OpenAI’s Stargate plans, including pauses in some expansion efforts.
🔗 Source: Bloomberg
🧠 Food for thought
Implications, context, and why it matters.
Meta hires away staff as OpenAI partnerships hit limits
- Talk of serious disruption at OpenAI’s Abilene site goes too far. Oracle said a separate planned 600-megawatt expansion lease will not proceed, while the main campus has two operating buildings and work continues 1.
- OpenAI put its UK Stargate project on hold due to high energy prices plus regulatory hurdles. That pause exposes weak points in a partnership-heavy approach to building global infrastructure 2.
- Meta funds its buildout with its advertising business, which delivered a 41% operating margin. That cash flow supports projected 2026 capital spending of US$115–135 billion 3.
- Meta also plans to curb expenses by building Andromeda AI chips, its in-house processors designed for AI workloads 3.
Power grids and permits now shape the AI race
- The UK stoppage shifts competition away from chip buying and toward access to electricity plus regulatory approvals 2.
- That change raises the stakes for specialized “neocloud” data center builders like Nscale, a startup that provides cloud computing capacity focused on AI. Nscale raised US$2 billion and now faces uncertainty tied to OpenAI’s pause of Stargate UK, where Nscale was partnering alongside Nvidia 4.
- National “sovereign AI” ambitions, efforts by countries to build AI capacity under domestic control, also run into these physical and cost limits. Energy policy can block technology plans 2.
- Interconnection queues, the line of projects waiting to connect to the power grid, for data centers already look “inflated.” Grid operators are expected to cut more speculative proposals, which makes power access a competitive advantage 5.
Stay updated on the go with our mobile app.
Get latest insights with smoother, more personalized experience through TIA mobile app.




