SpaceX inks compute deal with Reflection AI, an open source AI lab | TechCrunch

Open original

Reflection AI will pay $150 million monthly from July 2026 through 2029 for Nvidia GB300 chips and hardware at SpaceX’s Colossus 2 data center, in a contract worth up to $6.3 billion and cancellable after three months with 90 days’ notice. The startup says this deal underpins its open-weight AI strategy, positioning it against closed labs like Anthropic and OpenAI.

  • Reflection AI will pay SpaceX up to $6.3 billion ($150M/month, 2026-2029) for GB300 chip access at Colossus 2, but either party can cancel with 90 days' notice after the first quarter.
  • Reflection is the first open-weight AI lab to land a compute deal in the same tier as Anthropic ($1.25B/month) and Google ($920M/month) with SpaceX.
  • The Colossus facility was originally built for xAI, and its underused capacity is now being sold to outside labs since xAI's own projects have lagged.

Reflection AI has locked in a compute deal with SpaceX for $150 million a month starting July 1, 2026 and running through 2029. That adds up to as much as $6.3 billion, though either side can walk away on 90 days’ notice after the first quarter. The startup will tap Nvidia’s latest GB300 AI chips housed in SpaceX’s Colossus 2 data center near Memphis, Tennessee. SpaceX already leases chip capacity to Anthropic and Google under much larger agreements—$1.25 billion and $920 million per month respectively—but Reflection is the first open-source AI lab in this tier. Founded in 2024 by two ex-Google DeepMind researchers, the company bets on “open-weight” models—ones that publish their parameters—for transparency and broader adoption. Reflection says the deal is a sign that open-source strategies can secure frontier-class infrastructure alongside closed labs. The Colossus facility started life under xAI, Elon Musk’s internal AI offshoot, before being rolled into SpaceX. With xAI’s own projects lagging, those idle Nvidia chips became available to outside teams. Reflection’s contract highlights a shift: big compute pools once reserved for in-house efforts are now up for grabs, and open-source players can claim a share.

Ask about this article