CoreWeave has secured a $11.9bn contract from OpenAI, aimed at delivering crucial AI infrastructure and substantially boosting the ChatGPT developer’s computing capabilities. This five-year collaboration is expected to facilitate the global distribution of OpenAI’s models to millions of users. In conjunction with this agreement, OpenAI will also invest $350m in CoreWeave stock, a transaction that remains independent of the cloud-computing company’s anticipated initial public offering (IPO).
“Partnering with OpenAI on this net new contract underscores CoreWeave’s proven ability to deliver reliable and performant infrastructure services, powering AI Innovations for world-leading AI labs,” said CoreWeave co-founder and CEO Michael Intrator. “We remain a partner-of-choice to meet the bleeding-edge needs of pioneers to unleash AI’s potential to change the world.”
Nvidia-backed CoreWeave manages a cloud service network focused on AI, comprising 32 data centres. By the end of 2024, these facilities housed over 250,000 Nvidia GPUs, including the newer Blackwell GPUs designed for AI reasoning tasks. The addition of OpenAI as a direct customer could help address investor concerns about CoreWeave’s reliance on individual clients as it prepares for its IPO, an issue often viewed cautiously by investors.
Shifts in OpenAI and Microsoft relationship
This development also underscores changes in OpenAI’s relationship with Microsoft, the AI company’s largest investor. Though Microsoft remains a key supporter of OpenAI through revenue-sharing arrangements, it is no longer the exclusive cloud provider for OpenAI after earlier agreements with SoftBank and Oracle. In January, through the $500bn Stargate AI infrastructure initiative involving SoftBank and others, Microsoft ceased being the sole cloud provider for OpenAI. The ChatGPT maker continues to seek additional compute resources, as noted by CEO Sam Altman’s recent comments on GPU shortages. Altman recently stated that OpenAI is “out of GPUs,” indicating the necessity for expanded computing resources.
“Advanced AI systems require reliable compute, and we’re excited to continue scaling with CoreWeave so we can train even more powerful models and offer great services to even more users,” said Altman. “CoreWeave is an important addition to OpenAI’s infrastructure portfolio, complementing our commercial deals with Microsoft and Oracle, and our joint venture with Softbank on Stargate.”
Meanwhile, Microsoft has been developing its own AI models under the MAI brand, heightening competition with OpenAI. Additionally, Microsoft has appointed British AI entrepreneur Mustafa Suleyman to spearhead its AI strategy. Microsoft’s development of models like o1 and o3-mini parallels OpenAI’s efforts in AI reasoning.
Earlier this week, CoreWeave responded to claims reported by the Financial Times regarding Microsoft’s withdrawal from certain contracts. The allegations suggested that Microsoft had retreated from some commitments with CoreWeave, potentially affecting the startup’s planned $35bn IPO next month. Microsoft is a significant contributor to CoreWeave’s expected 2024 revenue of $1.2bn and accounts for 62% of its projected income. Any changes in their relationship could present challenges for CoreWeave, which recently disclosed its acquisition of the AI development platform Weights & Biases. This acquisition will integrate CoreWeave’s infrastructure and managed cloud services with the Weights & Biases platform, facilitating AI model training as well as the evaluation and monitoring of AI applications.
Read more: CoreWeave rejects claims of Microsoft contract withdrawals amid IPO plans
