OpenAI has signed a five-year, $11.9 billion agreement with CoreWeave, a rising AI cloud infrastructure provider, as it seeks to secure critical GPU compute capacity for its expanding AI models. The deal, first reported by Reuters, represents one of OpenAI’s largest cloud infrastructure investments to date and signals further diversification of its compute strategy beyond Microsoft’s Azure.
Key Points:
- OpenAI is committing nearly $12 billion to CoreWeave over five years for AI infrastructure.
- The deal grants OpenAI a $350 million stake in CoreWeave, which is preparing for an IPO.
- Microsoft, OpenAI’s biggest backer, is also CoreWeave’s largest customer, creating a strategic shift.
- CoreWeave is backed by NVIDIA and operates 32 AI data centers housing over 250,000 GPUs.
As part of the agreement, OpenAI will receive $350 million in CoreWeave equity, further entrenching its relationship with the cloud provider ahead of CoreWeave’s expected IPO. The deal also strengthens CoreWeave’s position in the market, as it attempts to reduce its reliance on Microsoft, which accounted for 62% of its $1.92 billion revenue in 2024.
This move further complicates the dynamic between OpenAI and Microsoft. Microsoft, which has invested billions into OpenAI and integrated its models into Azure and Copilot, has also been CoreWeave’s largest customer. However, recent reports suggest that Microsoft has scaled back some of its planned spending with CoreWeave, raising questions about the long-term balance of these relationships.
CoreWeave, originally a crypto mining company, has quickly transformed into a major AI infrastructure provider, operating 32 data centers and leveraging over 250,000 Nvidia GPUs. Backed by NVIDIA, CoreWeave has raised over $14.5 billion in debt and equity as it prepares to go public.
Chris McKay
For OpenAI, the deal secures an essential supply of high-performance computing resources at a time when CEO Sam Altman has openly voiced concerns about a shortage of GPUs. It also follows OpenAI’s broader push to secure AI infrastructure through multiple partners, including its Stargate initiative with SoftBank and Oracle.
With AI models growing larger and more expensive to train, securing infrastructure has become a key battleground in the AI industry. This deal not only strengthens OpenAI’s compute capacity but also shifts the power dynamics in an industry where access to GPUs is becoming just as critical as the models themselves.