AI startup Reflection signs computing power deal with SpaceX
Summarized and contextualized by DistantNews.
At a glance
- AI startup Reflection has signed a deal with SpaceX for access to computing power at its Colossus 2 data center.
- The agreement grants Reflection immediate access to Nvidia AI chips and requires monthly payments to SpaceX starting July 2026.
- This deal follows similar commercial agreements SpaceX has made with other tech giants like Google and Anthropic.
AI startup Reflection has secured a significant deal with Elon Musk's SpaceX, gaining access to substantial computing power at the company's Colossus 2 data center. The agreement, reported by CNBC, will provide Reflection with immediate access to Nvidia GB300s, advanced AI chips crucial for training and running sophisticated AI models.
Under the terms of the deal, Reflection is set to pay SpaceX $150 million per month. These payments are scheduled to begin on July 1, 2026, and continue through 2029. If the contract runs its full course, the total payments could amount to approximately $6.3 billion. Both companies have the option to terminate the contract with 90 days' notice after an initial three-month period.
Reflection, an open-source AI startup backed by Nvidia, stated that the increased compute capacity will allow it to "push the frontier on open models." This partnership underscores the growing demand for high-performance computing resources needed to develop advanced AI technologies. The deal also highlights SpaceX's expansion into providing data center services, leveraging its infrastructure to support the burgeoning AI industry.
The Reflection deal is the latest in a series of commercial agreements for SpaceX, which has recently struck similar deals with tech giants Google and AI startup Anthropic. Earlier in June, SpaceX announced a deal with Google, under which the search giant will pay $920 million per month from October 2024 to June 2029 for computing capacity. These agreements demonstrate SpaceX's strategy to diversify its revenue streams beyond space launches by offering its considerable computing infrastructure to leading AI developers.
More compute gives us more room to push the frontier on open models.
Originally published by CNA. Summarized and contextualized by our editorial team with added local perspective. Read our editorial standards.