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

SpaceX has entered into a significant computing agreement with Reflection AI, an open-source artificial intelligence laboratory, marking another milestone in the rapidly evolving landscape of AI infrastructure. According to reports, the deal involves Reflection AI paying SpaceX $150 million per month beginning July 1, 2026, through 2029. This arrangement grants Reflection immediate access to Nvidia's latest GB300 AI chips and related hardware located in SpaceX's Colossus 2 data center near Memphis, Tennessee. The total value of the contract could reach up to $6.3 billion over the three-year period, with both parties retaining the right to terminate the agreement with 90 days' notice after the initial three-month period.

This new partnership comes amid growing interest in open-source AI initiatives, particularly in light of recent regulatory actions affecting closed AI models. Reflection AI, established in 2024 by two former Google DeepMind researchers, has positioned itself as an alternative to major closed AI laboratories such as Anthropic and OpenAI. The company emphasizes its commitment to open-weight AI models, where trained parameters are made publicly available. This approach has gained traction following the U.S. government's restrictions on Anthropic's closed models, including Fable and Mythos.

The significance of this deal lies not only in its financial magnitude but also in its implications for the broader AI ecosystem. It represents one of the largest open AI infrastructure commitments announced to date. A spokesperson for Reflection AI highlighted the increasing recognition among nations and enterprises of the risks and costs associated with relying solely on closed models. They noted that the agreement with SpaceX underscores Reflection's strategic role in the frontier AI space, providing additional computational resources necessary for developing high-quality open models at scale.

The Colossus 2 data center, initially constructed by xAI—a company founded by Elon Musk that is now integrated into SpaceX—was originally intended for internal AI projects. However, as SpaceX's own AI endeavors have faced challenges, the organization has turned to monetizing its surplus AI chip capacity by offering them to leading AI research institutions. This shift reflects a broader trend in the industry where companies with excess computational resources are exploring opportunities to generate revenue through partnerships with external entities.

The deal between SpaceX and Reflection AI follows similar agreements with other prominent players in the AI field. For instance, SpaceX had previously inked contracts with Anthropic and Google, valued at $1.25 billion and $920 million per month, respectively. These agreements also extend through July 2029, though Elon Musk has often emphasized the flexibility of these arrangements, noting that they can be terminated at any time without long-term obligations.

As the AI sector continues to evolve, the collaboration between SpaceX and Reflection AI exemplifies the dynamic interplay between private enterprise and emerging technologies. With the potential to reshape the landscape of AI development, this partnership highlights the growing importance of open-source strategies in fostering innovation and ensuring accessibility across global markets. The ongoing developments in this area suggest that further collaborations and advancements are likely to emerge as stakeholders navigate the complexities of AI governance and technological progress.