"Nvidia ( NASDAQ:NVDA ) first invested approximately $100 million in CoreWeave ( NASDAQ:CRWV ) in 2023 and added to its stake around CoreWeave's IPO last March. In 2025, the close relationship between the two companies drew sharp criticism over potential "circular financing," with skeptics arguing Nvidia was effectively financing demand for its own GPUs through equity investments in major customers. CoreWeave turned out to be one of Nvidia's worst-performing investments through most of 2025, but despite these concerns, Nvidia is now making a significantly larger bet on the specialized AI cloud provider."
"This morning, Nvidia announced a $2 billion investment in CoreWeave Class A common stock priced at $87.20 per share. The deal significantly deepens the strategic partnership between the two companies. Under the expanded collaboration, CoreWeave will adopt Nvidia's CPU and storage platforms, deploy multiple generations of Nvidia infrastructure - including the upcoming Rubin platform and Vera CPUs - and accelerate the buildout of AI factories. The companies aim to bring more than 5 gigawatts of AI infrastructure online by 2030. CoreWeave will also test and validate its AI-native software stack for potential inclusion in Nvidia's reference architectures."
"The investment benefits both parties. For Nvidia, it secures a major, high-growth customer committed to heavy GPU purchases while expanding its ecosystem. For CoreWeave, the $2 billion cash infusion strengthens its balance sheet and helps fund aggressive expansion of land, power, and data center capacity needed to meet surging AI demand."
Nvidia initially invested about $100 million in CoreWeave in 2023 and increased its stake around CoreWeave's IPO. The relationship drew criticism in 2025 over alleged "circular financing," with skeptics saying Nvidia was indirectly financing demand for its own GPUs. CoreWeave underperformed through much of 2025, yet Nvidia has now committed $2 billion at $87.20 per share. The expanded deal requires CoreWeave to adopt Nvidia CPUs and storage, deploy multiple generations of Nvidia infrastructure including Rubin and Vera, and target over 5 gigawatts of AI capacity by 2030 while testing AI-native software for Nvidia reference architectures.
Read at 24/7 Wall St.
Unable to calculate read time
Collection
[
|
...
]