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

2026-06-22 · Source: AI News & Artificial Intelligence | TechCrunch · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure · Depth: Intermediate, quick

Summary

SpaceX has finalized a compute deal with Reflection AI, an open source AI startup, for immediate access to Nvidia's latest GB300 AI chips and supporting hardware. Beginning July 1, 2026, and extending through 2029, Reflection AI will pay \$150 million monthly, totaling up to \$6.3 billion, for resources at SpaceX's Colossus 2 data center near Memphis, Tennessee. This agreement, while smaller than SpaceX's prior contracts with Anthropic (\$1.25 billion/month) and Google (\$920 million/month), represents one of the largest announced open AI infrastructure commitments. Reflection AI emphasizes its open-weight AI strategy, positioning itself as a key alternative to closed frontier labs, especially following the U.S. government's ban on certain closed models. The Colossus data center, initially built by xAI, now part of SpaceX, is being utilized by SpaceX to rent out its AI chip holdings as its internal AI pursuits have shifted.

Key takeaway

For AI Product Managers evaluating compute strategies, this deal signals a growing trend where open-weight AI models are gaining substantial infrastructure backing. You should assess the long-term cost-effectiveness and strategic advantages of open-weight models, especially given the increasing regulatory scrutiny and dependency risks associated with closed-source alternatives. Consider securing dedicated, long-term compute resources to support ambitious open-model development, potentially exploring partnerships with infrastructure providers with excess capacity.

Key insights

Open-weight AI labs are securing significant compute infrastructure, challenging closed models and leveraging available data center capacity.

Principles

• Open-weight AI models gain traction due to perceived risks and costs of closed models.
• Excess compute capacity from internal AI efforts can be monetized by infrastructure providers.

In practice

• Secure long-term compute access for open-weight model development.
• Monetize underutilized AI chip infrastructure by leasing to other AI labs.

Topics

• SpaceX
• Reflection AI
• AI Compute Deals
• Open-Weight AI
• NVIDIA GB300
• Colossus Data Center

Best for: Director of AI/ML, AI Product Manager, Investor

Related on AIssential

• SpaceX: The AI Orbital Hyperscaler — A massive compute lease deal between Anthropic and SpaceX highlights critical AI infrastructure supply-demand imbalances.
• 640: Nvidia Invades the PC, SpaceX is Really a Neocloud, Frontier Token Math, OpenAI = Dropbox?, RAM, Meta AI Insta Hack, Google's 64 Million Mosquitoes, and Led Zeppelin — Frontier AI models deliver disproportionate value for critical tasks where quality improvements are non-linear.
• Seven Mental Models to Understand the AI Compute Era — The true AI arms race is in physical infrastructure and compute capacity, not just model development.
• Google will pay SpaceX $920M per month for compute — Major tech firms are securing massive, long-term compute capacity to meet surging AI product demand.
• Cursor officially joins the SpaceX AI machine — SpaceX's \$60B acquisition of Cursor and Z AI's GLM-5.2 release highlight rapid advancements in AI coding and open-weights models.
• SpaceX Inks Multibillion-Dollar Computing Deal With Reflection AI — The provided content, titled "SpaceX Inks Multibillion-Dollar Computing Deal With Reflection AI," is a paywall prompt from Bloomberg.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI News & Artificial Intelligence | TechCrunch.