OpenAI secures nearly $12 billion in computing power from AI infrastructure provider Coreweave while taking an ownership stake in the company.
OpenAI is forging a strategic partnership with AI infrastructure specialist Coreweave. The companies announced a deal worth up to $11.9 billion that gives OpenAI dedicated computing capacity for training and running AI models. The agreement goes beyond just purchasing compute power - OpenAI is investing $350 million to become a Coreweave shareholder.
"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," OpenAI CEO Sam Altman said in the announcement.
Building a multi-provider compute network
The Coreweave partnership adds to OpenAI's existing computing arrangements with Microsoft and Oracle, as well as its Stargate joint venture with Softbank. Coreweave, which calls itself an "AI hyperscaler," runs an expanding network of data centers across the US and Europe. The company has built its cloud platform specifically for compute-intensive AI workloads.
The massive investment shows OpenAI's continued belief that more computing power leads to better AI. This likely stems from recent test-time scaling findings - giving AI models more time and computing power to "think" appears to boost their performance, a principle OpenAI's newest o-models embrace.
In recent statements, Microsoft CEO Satya Nadella expressed a more cautious view on massive computing investments. He warns against a supply-first approach, emphasizing that building capacity alone won't guarantee success without proven customer demand. Looking ahead, Nadella predicts data center costs will decrease after 2027 and sees AI models becoming commoditized, with OpenAI transitioning primarily into a product company.
Building a complete AI development platform
Coreweave recently announced its acquisition of the AI development platform Weights & Biases. Founded in 2017, Weights & Biases provides tools for AI developers and serves over one million AI engineers, including teams at OpenAI, Meta, NVIDIA, and Toyota.
The acquisition aims to create an end-to-end platform helping AI labs and companies launch new AI innovations faster. The deal should close in the first half of 2025.
