US AI Infrastructure Company Secures $3 Billion Loan, Executives Sell Hundreds of Millions

CoreWeave, Inc. (NASDAQ: CRWV) agreed on May 15 to a new delayed-draw term loan (DDTL) of up to $3.1 billion (approximately ₩4 trillion) under a subsidiary’s name. The financing, which includes collateral pledges and various financial and restrictive covenants, will fund AI infrastructure asset acquisitions and the establishment of a dedicated liquidity account.

Separately, in early May, Chief Strategy Officer Brian Venturo converted several hundred thousand Class B shares into Class A shares through entities and trusts he controls and sold them as planned, generating approximately $10.5 million (around ₩14 billion) in proceeds. He still retains a significant indirect stake through multiple trusts and his spouse.

In March, CoreWeave secured an investment-grade, GPU-collateralized DDTL of $8.5 billion (about ₩11 trillion). In April— including a $1 billion private investment in public equity (PIPE)—the company raised over $20 billion (roughly ₩27 trillion) in debt and equity year-to-date to accelerate its AI cloud infrastructure expansion.

NVIDIA expanded its stake in January by acquiring an additional $2 billion (approximately ₩2.6 trillion) of CoreWeave common stock. Meanwhile, multibillion-dollar, long-term GPU cloud contracts with Meta, OpenAI, Anthropic, Jane Street and others have followed, further boosting the company’s data center investment plans and borrowing capacity.

Headquartered in New Jersey, CoreWeave is a high-performance AI cloud infrastructure provider powered by NVIDIA GPUs. It has seen rapid growth by supplying customized computing resources to large AI model developers and financial institutions.

Given the surging demand for AI compute and the need to secure expensive GPUs and data center capacity in advance, CoreWeave has rapidly scaled its operations through an aggressive financing structure—combining GPU-backed loans, corporate bonds, convertible notes and equity investments—all anchored by long-term usage contracts.