SpaceX fuels AI ambitions with $6.3 billion Reflection compute deal
SpaceX signed a $6.3 billion AI compute agreement with Reflection AI, starting July 2026 and running through 2029. The deal adds to SpaceX's Colossus infrastructure strategy, alongside existing contracts with Anthropic and Alphabet Inc. valued at approximately $45 billion and $30 billion respectively.

*this image is generated using AI for illustrative purposes only.
SpaceX has signed a $6.3 billion AI compute agreement with open-source startup Reflection AI, extending Elon Musk’s Colossus infrastructure strategy and highlighting the rise of AI compute as a standalone business line. The deal positions SpaceX to monetize its high-performance computing capacity amid surging demand for AI training power.
The Reflection contract will run about $150 million per month starting July 1, 2026, and extend through the end of 2029, for a total near $6.3 billion if fully executed. Reflection will tap Nvidia-powered clusters in xAI data centers, making it the latest external tenant on Colossus.
Strategic Partnerships
Reflection joins Alphabet Inc. and Anthropic as major external tenants on the Colossus infrastructure. Previous agreements from Google and Anthropic to use SpaceX’s compute are valued at roughly $30 billion and $45 billion, respectively, and extend to around mid-2029.
These arrangements sit inside Musk’s broader strategy to sell excess Colossus capacity to rival model labs rather than rely solely on internal workloads. By selling high-margin GPU capacity to customers like Anthropic, Google, and now Reflection, SpaceX is turning scarce compute into a core strategic asset.
Contract Details
The following table outlines the key financial terms for SpaceX's major AI compute contracts:
| Customer | Value | Monthly Rate | Duration |
|---|---|---|---|
| Reflection AI | $6.3 billion | $150 million | July 2026 – Dec 2029 |
| Anthropic | ~$45 billion | ~$1.25 billion | Through May 2029 |
| Alphabet Inc. | ~$30 billion | ~$920 million | 32 months (starting later this year) |
In May, filings showed Anthropic agreed to buy roughly 300 megawatts of AI infrastructure from xAI, effectively taking the entire output of the Colossus 1 site near Memphis through May 2029. Google has separately agreed to pay SpaceX about $920 million per month for compute over 32 months to access roughly 110,000 Nvidia GPUs plus supporting infrastructure in xAI facilities.
How will SpaceX manage potential capacity constraints if internal xAI training needs surge alongside these external contracts?
Will this trend of major AI labs renting rival infrastructure lead to increased regulatory scrutiny regarding market consolidation?
Does the 2026 start date for the Reflection AI contract indicate current capacity is fully reserved by Anthropic and Alphabet?
































