NVIDIA is further solidifying its dominance in the AI infrastructure market through a significant strategic move into the server CPU sector. CEO Jensen Huang confirmed the upcoming launch of a new CPU codenamed "Vera," which will be marketed for the first time as a standalone product, aiming to eliminate computational bottlenecks in the AI supply chain and fully support the development of Agentic AI. On January 26, it was reported that NVIDIA has committed to an additional $2 billion investment in the emerging cloud service provider CoreWeave, which is set to become the first customer to deploy the Vera CPU as an independent infrastructure option. In an interview, Jensen Huang described Vera as a "revolutionary" CPU. Analysis suggests this partnership not only continues their long-standing relationship but also marks a significant strategic shift for NVIDIA at the computing stack level. This move by NVIDIA is reportedly aimed at addressing the issue of server CPUs becoming a major bottleneck in the AI supply chain. By offering the standalone Vera CPU, NVIDIA provides customers with a more cost-effective alternative for high-end computing, enabling them to run their computing stack not only on NVIDIA GPUs but also process workloads on NVIDIA CPUs, which is expected to have a direct impact on the existing server processor market landscape. As technical details of the Vera CPU are revealed and production plans advance, NVIDIA is demonstrating its ambition in the "post-Blackwell era." This strategy not only targets high-end rack-scale solutions but is also beginning to permeate the broader general-purpose computing market, in an attempt to build a comprehensive, full-stack computing ecosystem. CoreWeave plays a pivotal role in this strategic deployment. As a key partner of NVIDIA, CoreWeave has ambitious plans to build a 5-gigawatt (GW) AI factory by 2030. According to official disclosures, NVIDIA will purchase Class A common stock of CoreWeave at a price of $87.20 per share; this substantial additional investment of $2 billion will directly support CoreWeave's infrastructure expansion. Jensen Huang revealed that CoreWeave would have to "race" to become the first customer to deploy the Vera CPU. Unlike previous sales strategies where CPUs were typically part of an overall superchip or rack system, NVIDIA is now explicitly offering the Vera CPU as a "standalone part" of the infrastructure. This grants customers greater flexibility when building their computing stack, without being forced to purchase an entire rack-scale solution. Huang hinted that while no other CPU design wins have been formally announced, more customers are expected to join subsequently. From a technical specifications perspective, the Vera CPU represents a significant leap forward in NVIDIA's processor design. Based on a next-generation custom Arm architecture codenamed "Olympus," it is designed to handle the most demanding AI and computational workloads. Data shows the Vera CPU is equipped with 88 cores and 176 threads, and introduces what NVIDIA terms "Spatial Multi-Threading" technology. In terms of memory performance, Vera features 1.5 TB of system memory, triple that of the previous-generation Grace CPU; it also utilizes SOCAMM LPDDR5X technology to achieve a memory bandwidth of 1.2 TB/s. Furthermore, via NVLink-C2C interconnect technology, its coherent memory interconnect speed reaches up to 1.8 TB/s, double that of Grace. Compared to the Grace CPU, Vera also features a substantial upgrade in cache design. Each core possesses 2MB of L2 cache (double that of Grace), along with 162MB of shared L3 cache (a 42% increase). This enables Vera to share data with paired GPUs more rapidly, significantly enhancing overall system efficiency. NVIDIA stated that Vera will be one of its most powerful CPU products to date and holds a leading advantage in rack-scale confidential computing. It is noteworthy that the Vera CPU does not exist in isolation; it is a core component of NVIDIA's next-generation Vera-Rubin platform. Dion Harris, Senior Director of High-Performance Computing and AI Factory Solutions at NVIDIA, revealed that the core chips for the Vera-Rubin NVL72 platform have returned from TSMC and are currently undergoing debugging before being sent to key partners, with mass production expected to begin in the second half of 2026. This platform combines the Vera CPU with the next-generation GPU codenamed "Rubin." The Rubin GPU is expected to utilize HBM4 memory with a bandwidth of up to 22 TB/s, which is 2.75 times that of the Blackwell GPU. The driving force behind NVIDIA's aggressive pursuit of an independent CPU strategy is a deep understanding of AI supply chain bottlenecks. As Agentic AI applications accelerate in adoption, server CPUs are increasingly becoming a key factor constraining overall system performance. NVIDIA recognizes that relying solely on the growth of GPU computing power is no longer sufficient to meet market demand; a CPU platform with matching performance must be provided. Analysis indicates that by offering the standalone Vera CPU, NVIDIA can not only provide customers with a "workaround" to bypass bottlenecks associated with traditional x86 architectures but also attract budget-conscious customers seeking high-end CPU capabilities through a lower-cost option. Beyond CoreWeave, the early customer list for NVIDIA's Vera-Rubin system includes cloud giants such as Amazon AWS, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure, as well as emerging players like Lambda, Nebius, and Nscale. Analysis points out that even as some cloud providers are developing their own accelerators, NVIDIA is making a heavy bet on the server CPU ecosystem, attempting to maintain its unshakeable core position in the competition for next-generation AI infrastructure.