Nvidia doubles down on CoreWeave with $2B Class A stock buy as part of $3.8B AI bet
What's the deal? Nvidia has invested $3.8B across two AI-related stocks, with the bulk going to CoreWeave, an AI data centre provider. The chipmaker put $2B into CoreWeave Class A common stock at $87.20 per share, lifting its total stake to $3.65B — a 95% increase from prior holdings. It also opened a new position in Coherent, an optical networking company.
CoreWeave buys Nvidia's GPUs and rents compute capacity to companies building AI solutions. The two firms plan to deepen integration of CoreWeave's software with Nvidia's computing platforms, including the Rubin, Vera, and BlueField systems.
Why now? Enterprise demand for AI computing is surging, and the bottleneck is increasingly infrastructure — not chips. CoreWeave is targeting 5 gigawatts of AI data centre capacity by 2030, and Nvidia's capital will help it secure land, power, and the physical infrastructure needed to get there.
For Nvidia, the investment is strategic as much as financial. By backing a major buyer of its GPUs, it locks in demand while ensuring the buildout of data centres that run its hardware.
What could go wrong? The arrangement raises questions about circular economics. Nvidia is effectively investing in a company whose business depends on buying Nvidia products. If AI demand softens or CoreWeave struggles with its ambitious expansion targets, Nvidia would face losses both as a supplier and as an investor.
CoreWeave also carries significant debt from its rapid buildout. Any slowdown in the AI infrastructure cycle could put pressure on its finances and, by extension, on Nvidia's stake.
The signal: Nvidia's dual bet pairs a late-growth GPU cloud provider in CoreWeave with a mature optical-components maker in Coherent, signalling that the corporate investor sees constraints shifting from raw compute to the physical fabric — power, land, and networking — that connects it. With CoreWeave still in its late-growth stage and carrying heavy buildout debt, Nvidia is effectively underwriting its own demand pipeline, a vertical integration play that could pay off handsomely if AI infrastructure spending holds but would compound its exposure if the cycle turns.
Read more: ainvest.com
