NVIDIA’s CEO, Jensen Huang, reaffirmed the company’s strong support for CoreWeave, announcing that the neocloud will be the first to utilize the new Vera CPUs as an independent offering.
CoreWeave Unveils Vera CPUs as Standalone Option: A Bold Move by NVIDIA in Agentic AI
In a significant financial maneuver, NVIDIA has returned to the capital markets, fresh off lucrative partnerships with major players like Groq and Intel. The tech giant has pledged an additional $2 billion investment in CoreWeave, underscoring their commitment to long-term collaboration. As detailed in a recent blog post by the company, NVIDIA plans to acquire Class A common stock priced at $87.20 per share. This investment is pivotal for CoreWeave’s ambitious target of establishing 5 gigawatts of AI factories by 2030. During an interview with Bloomberg’s Ed Ludlow, Huang disclosed that CoreWeave will be the inaugural user of the Vera CPUs under this new arrangement.
“For the very first time, we’re going to be offering Vera CPUs. Vera is such an incredible CPU. We’re going to offer Vera CPUs as a standalone part of the infrastructure. And so not only can you run your computing stack on NVIDIA GPUs, you can now also run your computing stack, wherever their CPU workload, run on NVIDIA CPUs…”
“Vera is completely revolutionary…CoreWeave is going to have to race if CoreWeave’s going to be the first to stand up Vera CPUs. We haven’t announced any of our CPU design wins, but there are going to be many.”
– NVIDIA’s Jensen Huang via Ed Ludlow
This agreement illustrates two main narratives emerging within the industry. First, NVIDIA has acknowledged that server CPUs are increasingly becoming a bottleneck in the AI supply chain, especially as demands for agentic AI applications continue to escalate. A robust platform is now more crucial than ever. Second, by offering Vera CPUs as standalone options, NVIDIA is providing clients with an alternative for high-end CPU capabilities—potentially a cost-effective solution compared to acquiring a full rack-scale system.
The Vera CPUs promise to be among NVIDIA’s most advanced offerings, significantly outpacing the existing Grace Blackwell models. Built on next-generation custom ARM architecture known as Olympus, this impressive CPU boasts 88 cores, 176 threads (incorporating NVIDIA Spatial Multi-Threading), and an astonishing 1.8 TB/s NVLink-C2C coherent memory interconnect. Additional specifications include 1.5 TB of system memory (three times that of Grace) and a memory bandwidth of 1.2 TB/s with SOCAMM LPDDR5X, facilitating rack-scale confidential computing.
This strategic direction signals NVIDIA’s intent to make a substantial impact on the server CPU market. Moreover, when Jensen references “CPU design wins, ”he might also be hinting at forthcoming ARM-based N1/N1X SoCs, slated for consumer applications tailored for AI-driven PC workloads.
Related Articles:
Microsoft Introduces Xbox Mode for All Windows 11 PCs, Merging Console and Desktop Gaming Experiences
20:22
Samsung Ends LPDDR4 Memory Production Amid 50x Profit Surge, Predicts Increasing Memory Shortages by 2027
19:07
Ultimate Guide to Windrose PC Performance Analysis and Tuning for Optimal PC Experience
18:18
Leave a Reply