In a significant move within the artificial intelligence sector, SpaceX has entered into a substantial compute agreement with Reflection AI, an open-source AI startup. This partnership grants Reflection AI immediate access to Nvidia’s latest GB300 AI chips and supporting hardware housed in SpaceX’s Colossus 2 data center near Memphis, Tennessee. The agreement, commencing on July 1, 2026, involves monthly payments of $150 million from Reflection AI to SpaceX, totaling up to $6.3 billion through 2029. Notably, either party retains the option to terminate the contract with 90 days’ notice after the initial three-month period.
This collaboration underscores the escalating demand for advanced computational resources among AI companies striving to develop and deploy sophisticated models. Reflection AI, established in 2024 by former Google DeepMind researchers, positions itself as a proponent of open-weight AI models—those that publicly release their trained parameters. This approach offers an alternative to the closed models developed by leading AI labs, addressing growing concerns over the risks and costs associated with reliance on proprietary systems.
The Colossus 2 data center, originally constructed by xAI—a company founded by Elon Musk and now integrated into SpaceX—was initially intended to support internal AI initiatives. However, as SpaceX’s internal AI projects faced challenges, the company has leveraged its substantial AI infrastructure by offering compute resources to external AI labs. This strategy has led to significant agreements with other major players in the AI field, including Anthropic and Google, which have secured compute deals valued at $1.25 billion and $920 million per month, respectively.
Reflection AI’s partnership with SpaceX is particularly noteworthy in the context of recent developments in the AI landscape. The U.S. government’s ban on Anthropic’s closed models, Fable and Mythos, has intensified interest in open-source AI solutions. Reflection AI’s commitment to open-weight models aligns with a broader industry trend toward transparency and accessibility in AI development.
As the AI sector continues to evolve rapidly, the collaboration between SpaceX and Reflection AI highlights the critical role of computational resources in driving innovation. This partnership not only provides Reflection AI with the necessary infrastructure to advance its open-source initiatives but also positions SpaceX as a pivotal player in the AI compute market. Observers will be keenly watching how this alliance influences the competitive dynamics among AI labs and the broader adoption of open-weight models in the industry.
