CoreWeave Plans $23B AI Infrastructure Push Despite Short-Term Market Jitters

CoreWeave, the Nvidia-backed AI cloud services provider, announced plans to invest between $20 billion and $23 billion in AI infrastructure and data centers in 2025. This aggressive capital expenditure strategy aims to support soaring demand from major clients like Microsoft and OpenAI.

Despite reporting first-quarter revenue of $981.6 million—well above analysts’ forecast of $852.9 million—CoreWeave’s shares dropped 5%, reflecting investor concerns over its high spending. The company expects Q2 capex to reach $3 billion to $3.5 billion, far exceeding projected quarterly revenue of $1.06 billion to $1.1 billion.

CEO Mike Intrator explained the company’s unconventional financial model, which involves fully recovering infrastructure costs while maintaining profitability. “You pay back your infrastructure fully loaded and you have significant profit,” he told Reuters.

CoreWeave’s long-term outlook remains strong, with a revenue backlog of $25.9 billion as of March 31. A recent five-year agreement with OpenAI added $11.2 billion to the backlog, positioning CoreWeave as a key AI infrastructure provider. The deal also gives OpenAI a stake in the company.

However, analysts caution that CoreWeave’s role as overflow capacity for Microsoft may limit future demand. The company is also navigating U.S.-China trade tensions by diversifying its supply chains to reduce tariff exposure.

CoreWeave forecasts annual revenue between $4.9 billion and $5.1 billion, surpassing LSEG’s consensus estimate of $4.61 billion. While the AI boom fuels optimism, the company’s capital-intensive model and future dependency on top-tier clients pose long-term challenges.

As the race to build generative AI accelerates, CoreWeave’s bold bet on infrastructure could secure a competitive edge—if it balances growth with financial discipline.