After establishing dominance in the AI chip market with its GPUs, NVIDIA is now setting its expansion sights on a previously untapped domain—the CPU market. NVIDIA's CFO, Colette Kress, stated during the company's earnings call this Wednesday that its first standalone AI-focused CPU product, Vera, has opened up a "completely new" potential market worth $200 billion. She indicated that NVIDIA already has nearly $20 billion in revenue visibility for its CPU business this year, with Vera "propelling us to become a leading global CPU supplier." This announcement quickly captured Wall Street's attention. Multiple analysts noted that if the projected $20 billion scale materializes, it would position NVIDIA directly among the leaders in the CPU market, surpassing AMD's revenue in the cloud CPU segment. Simultaneously, this move signifies NVIDIA's expansion into a market long dominated by Intel, AMD, and Arm. For competitors like Arm and Qualcomm, which rely on independent CPU architectures, NVIDIA's entry could strain already tight production capacity and pose a substantial challenge.
Vera Unlocks a $200 Billion New Market
Officially launched in March, Vera is NVIDIA's first-ever standalone CPU product, specifically designed for AI inference and agentic workloads and built on the Arm architecture. Kress mentioned during the earnings call that hyperscale cloud customers and AI system manufacturers have begun deploying the chip, noting that the CPU market is one NVIDIA had "never previously entered." According to Kress, NVIDIA has nearly $20 billion in revenue visibility for its CPU business this year. If realized, this figure would exceed UBS analyst Timothy Arcuri's forecast of approximately $16 billion in CPU revenue for AMD this year—AMD currently holds over a 50% share in the cloud CPU market. In a research note on Thursday, Bernstein analyst Stacy Rasgon wrote that NVIDIA "appears poised to become a king in the CPU space, with a business scale potentially matching or even surpassing more traditional competitors in that market."
Wall Street: The $20 Billion Figure Requires Interpretation
Analysts generally agree that the $20 billion figure warrants careful interpretation and should not be directly compared with traditional CPU revenue metrics. Arcuri pointed out that Vera can be used either as a standalone chip paired with non-NVIDIA AI chips or sold as a complete system solution including CPU and memory components. Since expensive memory chips are included in the latter bundled offering, the $20 billion projection is not "as straightforward as it seems." He estimates Vera's value constitutes about 30% of the entire system package and projects NVIDIA's actual CPU revenue for this year to be in the range of $5 billion to $7 billion. Rasgon suggested the high $20 billion figure might stem from NVIDIA pricing its CPUs by the rack rather than per individual chip. Morgan Stanley analyst Joseph Moore shared a similar view, stating the number "likely includes CPUs shipped alongside GPU cards, which counts but is relatively easier to achieve." Nevertheless, Moore noted that a $20 billion target would place NVIDIA "at the threshold of market leadership," while his revenue modeling for Intel's data center business (including some networking products) remains slightly above NVIDIA's.
Inference Wave Reignites CPU Demand
NVIDIA's entry into the CPU market coincides with the accelerated shift in AI applications from model training to inference and agentic AI paradigms. Unlike the training phase, which heavily relies on GPUs, the inference stage sees a significant resurgence in demand for server CPUs. This trend has already shown signs of recovery for previously struggling Intel, and Arm Holdings earlier this year announced its first-ever self-developed CPU chip, specifically targeting inference and agentic AI workloads. Moore stated that the Arm-architecture CPU segment NVIDIA is entering competes in a different dimension compared to Intel's x86-architecture CPUs, implying "no additional negative impact" for x86-based chips. In the competitive landscape beyond NVIDIA, Arcuri warned that NVIDIA's aggressive push into the CPU field would further consume already constrained wafer and memory capacity, "posing negative implications" for the market expansion of companies like Arm and Qualcomm. Arcuri particularly noted that Arm's partnerships with foundries and memory suppliers are relatively less mature compared to NVIDIA's, potentially making the impact more direct. Moore, however, believes that NVIDIA's years of accumulated supply chain bargaining power and procurement priority will maintain its strong competitiveness even in a capacity-constrained environment.
