Microsoft opted not to exercise a pre-existing $12 billion contract option with cloud provider CoreWeave, according to a report from Semafor.
Instead, OpenAI stepped in, securing a five-year, $11.9 billion agreement to access CoreWeave’s GPU-powered cloud services. The decision reflects Microsoft’s evolving AI infrastructure strategy, shifting away from third-party providers in favor of internal investments, while OpenAI works to diversify its computing power beyond reliance on Microsoft Azure.
Microsoft Opts Out of CoreWeave Contract in Favor of In-House AI Infrastructure
Microsoft had previously been a key client of CoreWeave, relying on its GPU cloud services to support large-scale AI workloads.
However, the company has been actively restructuring its AI infrastructure investments, reducing external dependencies. This decision aligns with Microsoft’s broader shift toward proprietary AI hardware, including its Azure Maia and Cobalt AI chips, designed to lower reliance on Nvidia GPUs.
Microsoft’s pullback from CoreWeave is part of a larger retreat from external AI infrastructure commitments. In February 2025, the company canceled multiple AI data center leases across the U.S. and put a $3.3 billion data center project in Wisconsin on hold.
OpenAI Moves to Secure AI Compute Power
As Microsoft steps back, OpenAI takes over the available is expanding its cloud partnerships. The company has struck a $11.9 billion deal with CoreWeave to secure a stable supply of high-performance computing resources.
Beyond securing compute capacity, OpenAI has also taken a direct financial stake in CoreWeave. The company is acquiring $350 million in CoreWeave shares through a private placement linked to CoreWeave’s upcoming IPO, a move that positions OpenAI as a strategic investor in the company’s future.
The deal also reflects OpenAI’s broader efforts to reduce its reliance on Microsoft Azure. In early 2025, OpenAI secured a $40 billion investment from SoftBank, fueling speculation that the company may eventually build its own AI computing infrastructure.
The CoreWeave agreement is a step in that direction, ensuring OpenAI has an independent source of cloud compute outside of Microsoft’s ecosystem.
CoreWeave’s Financial Challenges and IPO Plans
CoreWeave is preparing for an IPO expected to raise up to $2.7 billion. However, it reported $1.9 billion in revenue in 2024 but also recorded a net loss of $863 million. It is also dependent on Nvidia, which has faced supply chain issues with its Blackwell B200 chips.
Despite speculation that Microsoft’s retreat could be tied to CoreWeave’s financial risks, the company has denied instability concerns. A spokesperson for CoreWeave stated to Winbuzzer earlier that there “have been no contract cancellations or walking away from commitments. Any claim to the contrary is false and misleading.”
Microsoft and OpenAI Reshape AI Infrastructure Spending
Microsoft’s decision to walk away from its CoreWeave option underscores a growing divide in AI infrastructure strategies. While Microsoft is centralizing compute resources within its own ecosystem, OpenAI is securing alternative cloud partnerships to avoid being locked into a single provider.
Other tech giants are adopting different approaches. Google and Amazon are aggressively expanding their AI infrastructure, integrating their own AI-optimized chips into cloud offerings rather than relying on third-party hardware like Nvidia’s GPUs.
Meanwhile, Nvidia remains a key player, but its tight supply chain and growing list of cloud partners suggest that the AI computing market is becoming increasingly fragmented.
The long-term impact of Microsoft’s withdrawal from CoreWeave remains uncertain. OpenAI’s investment ensures that CoreWeave remains a viable competitor in the cloud AI space, at least in the short term.
However, as companies move toward greater control over their infrastructure, the AI cloud computing landscape is set up for further shifts.
Last Updated on March 29, 2025 10:42 am CET
