What's happened
Reflection AI will pay SpaceXAI $150 million per month through 2029 for access to Nvidia GB300 chips at Colossus 2, enabling open-weight AI model training. The deal mirrors similar arrangements with frontier labs and signals a push toward open-source AI infrastructure as the race for compute intensifies.
What's behind the headline?
What this means for the open-source AI path
- Reflection has secured a steady compute pipeline, signaling that access to hardware is now a primary bottleneck alongside model development.
- The deal situates SpaceX as a critical compute conduit, potentially reshaping who controls access to frontline AI resources.
- This advance could accelerate the pace at which open models scale, narrowing the gap with closed systems.
What the industry stance suggests
- Backers view open-weight models as offering transparency and control, which governments and enterprises increasingly demand.
- The economics of compute contracts will influence which startups can compete with larger frontier labs.
What to watch next
- Whether Reflection expands its model offerings or enters new partnerships to widen compute access.
- How regulatory environments respond to increased use of external compute for sensitive workloads.
How we got here
Reflection AI, founded in 2024 by former Google DeepMind researchers, is pursuing open-source AI models as a counterpart to closed frontier labs. SpaceX operates Colossus 2, a data center built to support internal AI efforts and now rented to outside labs. Nvidia chips underpin the compute needed to train and run large models, making such partnerships pivotal for startups competing with OpenAI and Anthropic.
Our analysis
TechCrunch reports that Reflection will pay SpaceXAI $150 million monthly through 2029 for Nvidia GB300 chips at Colossus 2. Axios notes the same figure, framing it as part of a broader trend of frontier-lab compute deals. CNBC confirms the terms and highlights SpaceX’s strategy to monetize Colossus infrastructure. All sources describe the deal as a strategic move toward open-source AI models and compute-enabled competition with major labs.
Go deeper
- What does Reflection hope to gain by emphasizing open-source models?
- Will SpaceX extend Colossus 2's external-user access beyond Reflection?
- How will this affect prices or availability of compute for other startups?
More on these topics
-
Google - Technology company
Google LLC is an American multinational technology company that specializes in Internet-related services and products, which include online advertising technologies, a search engine, cloud computing, software, and hardware.
-
Nvidia - Computer game company
Nvidia Corporation is an American multinational technology company incorporated in Delaware and based in Santa Clara, California.
-
Anthropic - Artificial intelligence company
Anthropic PBC is a U.S.-based artificial intelligence startup public-benefit company, founded in 2021. It researches and develops AI to "study their safety properties at the technological frontier" and use this research to deploy safe, reliable models for
-
SpaceX - Aerospace company
Space Exploration Technologies Corp., trading as SpaceX, is an American aerospace manufacturer and space transportation services company headquartered in Hawthorne, California.
-
Elon Musk - CEO of SpaceX
Elon Reeve Musk FRS is an engineer, industrial designer, technology entrepreneur and philanthropist. He is the founder, CEO, CTO and chief designer of SpaceX; early investor, CEO and product architect of Tesla, Inc.; founder of The Boring Company; co-foun