CoreWeave secures US$21 billion Meta AI cloud deal and shares rise 24%

The article reports CoreWeave expanded a long-term agreement to supply Meta with AI cloud capacity through December 2032 worth about US$21 billion, and that the company also announced related financings and a multi-year Anthropic cloud deal.

CoreWeave announced an expanded long-term agreement to supply Meta with AI cloud capacity through December 2032 worth about US$21 billion. The company also reported a separate multi-year cloud infrastructure deal with Anthropic and multiple financings to fund its data center build-out. Together, the contracts and financings are described as deepening CoreWeave's role as an AI infrastructure provider while materially increasing its committed revenue base and balance sheet obligations. The article examines how the Meta agreement reshapes CoreWeave's investment narrative around growth, risk, and scale. Key details: - The expanded Meta agreement covers AI cloud capacity through December 2032 and is valued at about US$21 billion, according to the article. - CoreWeave announced a multi-year cloud infrastructure deal with Anthropic. - Financings noted in the article include an upsized US$3.5 billion 1.75% convertible note due 2032, US$1.75 billion 9.75% senior notes, and an US$8.5 billion term loan to fund the data center build-out. - The article says these moves deepen CoreWeave's role as AI infrastructure for leading model developers while increasing fixed obligations and potential dilution for shareholders. - The article mentions CoreWeave's shares rose about 24% on the news and cites company forecasts projecting roughly US$26.9 billion revenue and US$1.6 billion earnings by 2028, figures that imply very rapid annual growth. Summary: The agreements and funding package reinforce a near-term catalyst through backlog conversion and capacity ramp while also magnifying execution and leverage risks if build-out costs or AI demand differ from expectations. The article highlights that rising leverage could quickly matter if workloads grow more slowly than forecasted. Undetermined at this time.