Nvidia expects CPU revenue to more than double in fiscal 28

Nvidia, the company that built its empire on graphics processors, is coming for the CPU throne. The chipmaker expects its CPU revenue to more than double by fiscal year 2028, a projection that should make Intel and AMD executives reach for the antacids.

The company currently generates roughly $20 billion in CPU revenue for fiscal year 2027, according to CFO Colette Kress during the Q1 FY2027 earnings call. That figure combining standalone Grace and Vera CPUs with those integrated into superchips. The doubling target for fiscal 28 would put the segment somewhere in the neighborhood of $40 billion, which is no longer a side hustle by any definition.

From GPU king to full-stack dominator

Nvidia’s CPU ambitions aren’t about competing for your laptop processor. The company has identified a $200 billion total addressable market focused specifically on agentic AI workloads, a category distinct from traditional general-purpose computing.

Early demand for the Vera CPU is already supporting a trend toward 1:1 CPU-to-GPU ratios in AI deployments. That’s a telling detail. For every GPU Nvidia sells into an AI data center, it could potentially sell a matched CPU alongside it.

CPUs currently account for roughly 3% of Nvidia’s total revenue. Analysts expect that share to climb to approximately 5% by calendar year 2028.

The numbers Wall Street is watching

Wolfe Research has laid out projections that paint a steep growth curve for Nvidia’s agentic CPU revenues: $6.6 billion in 2026, $14 billion in 2027, and $24.6 billion by 2028.

For context, the entire x86 server CPU market sits at around $30 billion. Nvidia doesn’t even make x86 chips. It’s building on Arm architecture instead, effectively trying to carve out a parallel market rather than steal share directly from Intel and AMD’s existing turf.

Partners like Anthropic and OpenAI, companies building the agentic AI systems Nvidia is targeting, stand to benefit from tighter integration between the CPUs and GPUs powering their infrastructure.

What this means for the competitive landscape

AMD has been gaining server CPU market share against Intel with its EPYC line while simultaneously competing with Nvidia in data center GPUs. Nvidia’s CPU push opens a second front that AMD will need to defend.

The key differentiator for Nvidia isn’t raw CPU performance in isolation. It’s the integrated stack. A data center operator building an AI system can now source GPUs, CPUs, networking (via Nvidia’s InfiniBand and Ethernet offerings), and software from a single vendor.

For investors and industry watchers, the metric to track isn’t just whether Nvidia hits its doubling target. It’s whether the 1:1 CPU-to-GPU ratio holds as AI deployments scale.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.