CoreWeave chief dismisses talk of circular AI funding

CoreWeave CEO Michael Intrator slammed “flawed” circular investment claims. He said the significant AI partnerships represent actual infrastructure expansion rather than Big Tech money that has been recycled.

Intrator dismissed concerns that the company’s partnerships with other major players in artificial intelligence are circular investments, claiming that this line of reasoning is “fundamentally flawed.” 

CoreWeave expands AI infrastructure amid circular investment claims

Intrator said that major firms, including Meta, Microsoft, Amazon, and Google, are investing in infrastructure to meet increasing customer demand. CoreWeave CEO also emphasized that these expenditures are driven by real market demands rather than circular financial activity.

Wall Street analysts are concerned that the agreements are too circular. Intrator responded that the agreements represent a “fundamental infrastructure buildout.” He added that this dynamic occurs in other markets and that “it is not unusual to see partnerships as people try to serve infrastructure to the consumers” during such a massive infrastructure buildout.

On March 27, CoreWeave reported that it had raised $1.5 billion to go public, making it the largest U.S. tech IPO since 2021. Since CoreWeave’s Wall Street debut, the company’s stock has increased by more than 200%, driven by robust investor demand for AI and data center expansion.

On September 25, CoreWeave announced a $6.5 billion expansion of its current agreement with OpenAI, bringing its total contracts with the company to $22.4 billion. A few days later, CoreWeave and Meta inked a $14.2 billion agreement. CoreWeave and Nvidia announced a $6.3 billion order earlier in September.

CoreWeave strengthens market position with new tools and acquisitions

NASDAQ: CRWV increased 8.7% on Wednesday as the artificial intelligence cloud provider announced new tools to help programmers develop AI agents. The announcement follows its $1 billion acquisition of Weights & Biases, a firm that provides developers with AI model training and evaluation software, which was completed five months prior. The agreement aimed to supplement CoreWeave’s current business model, which involves renting out Nvidia graphics processing units to businesses that require infrastructure. 

The tech firm confirmed that with the new serverless reinforcement learning service, there is no need to worry about increasing or deleting processing resources.

Tests showed that developers could train models quickly and at 40% less expense by using CoreWeave’s new service instead of operating Nvidia H100 graphics processing units locally, “with no impact on model quality.” 

Salanki also stated that CoreWeave is distinguishing itself from its competitors by integrating infrastructure, reinforcement learning frameworks, and developer tools to enable businesses of all sizes to harness the full potential of AI agents.

CoreWeave reported that Serverless RL is already generating significant interest from clients, including Fortune 500 companies and AI-native businesses. The tech firm added that Serverless RL will be utilized by SquadStack.ai, an AI-powered contact center platform that delivers highly customized experiences for top consumer brands, to enhance customer engagement. QA Wolf, a hybrid platform that helps tech teams produce better software more quickly, will also utilize the Serverless RL.

According to Intrator, companies have been rushing to secure GPUs to implement AI projects in the cloud. He added that CoreWeave competes with leading providers such as Amazon Web Services. Intrator claimed that some companies may prefer to keep GPUs in their own data centers.

If you're reading this, you’re already ahead. Stay there with our newsletter.