New $4 Billion Loan Secured for AI Cloud Infrastructure Expansion; Executives Continue to Sell Shares

CoreWeave, Inc. (Nasdaq: CRWV), a U.S. AI‐cloud infrastructure provider, on May 15 closed a new $3.1 billion delayed‐draw term‐loan facility through a special purpose vehicle sponsored by its parent company. The financing will fund data‐center infrastructure acquisitions, bolster liquidity and cover transaction costs. The loan is secured by a first lien on the assets and equity of the borrower and its guarantor subsidiary, includes broad financial and operating covenants, and carries a make‐whole premium if repaid during the first three years.

In late April and early May, CEO Michael N. Intrater and Chief Strategy Officer Brian M. Venturo, under their 10b5-1 trading plans, converted Class B shares into Class A shares and sold several hundred billion Korean won worth of stock to streamline certain corporate and trust‐held interests. Both executives, however, continue to hold substantial indirect stakes through multiple trusts and family‐owned shares.

Recently, CoreWeave launched new offerings on its AI cloud—including integrated agentic AI features and a serverless reinforcement‐learning service—to support large‐scale autonomous agent training and inference.

In its first-quarter earnings report, the company said it maintained strong growth driven by AI‐cloud demand. CoreWeave highlighted a robust pipeline of major customers—including a $21 billion commitment from Meta, $6 billion from Jane Street and a multi-year AI‐cloud contract with Anthropic—and noted it has secured $8.5 billion in GPU‐backed financing to expand its infrastructure investments.

Headquartered in New Jersey, CoreWeave is building out over 5 gigawatts of AI “factories” and data centers by 2030 in collaboration with NVIDIA, supplying large-scale GPU compute capacity. Amid the generative AI boom, demand for AI-dedicated “neo-cloud” providers has surged, and CoreWeave continues to raise significant capital and invest in facilities to secure high-performance GPUs, power and network infrastructure as it competes with major hyperscalers.