Nvidia’s $200 Billion CPU Pitch Still Counts on China

Nvidia’s latest growth story is starting to look less like a pure GPU narrative and more like a fight over who controls the full AI system. The key new signal came on May 23, when Chief Executive Jensen Huang said the company’s forecast for a new $200 billion CPU market includes China. That is a small sentence with large implications. It suggests Nvidia is not writing off the Chinese market even after export controls and political friction have made its advanced AI chip business there highly uncertain. For investors, that turns the Vera CPU from a product extension into a strategic hedge on how much of the next AI spending wave Nvidia can capture outside its traditional accelerator lane.

The timing matters because Nvidia has enough momentum to make a new market claim financially relevant. In first-quarter fiscal 2027 results released on May 20, the company reported record revenue of $81.6 billion, up 85% year over year, and record Data Center revenue of $75.2 billion, up 92%. It also approved an additional $80 billion share repurchase authorization and lifted its quarterly dividend to $0.25 from $0.01. Those are not the moves of a company trying to defend a mature business. They are the moves of a company signaling confidence that the AI build-out still has room to widen. Vera is part of that widening story because it lets Nvidia attach more silicon and more system value to each AI factory deployment rather than relying only on GPU unit growth.

China is the real complication inside the bullish CPU pitch. Reuters reported on May 23 that Huang said he would expect China to be part of the $200 billion opportunity. But the same report also underlined the present constraint: Nvidia has U.S. licenses to sell H200 chips into China, yet deliveries have not started because Chinese authorities have not moved those purchases forward. That split matters for valuation. Investors can reasonably count China as a long-term demand pool for general-purpose and AI-adjacent compute, but they cannot assume near-term conversion. In other words, the addressable market may still be there, while the monetizable market remains narrower and politically conditional.

Vera’s customer list explains why Nvidia wants that optionality. In its March launch materials, Nvidia said customers collaborating to deploy Vera include Alibaba, ByteDance, Meta and Oracle Cloud Infrastructure, with additional planned deployments from Cloudflare, CoreWeave, Crusoe, Lambda, Nebius, Together.AI and Vultr. That is a broader mix than the classic GPU story. It spans Chinese internet groups, U.S. hyperscale and AI-cloud providers, and the system makers assembling the hardware. Economically, that broadens Nvidia’s reach from accelerator supplier to platform landlord. If GPU demand remains strong but buyers increasingly want tighter control over data movement, orchestration, inference environments and rack-level efficiency, owning the host CPU becomes another way to capture wallet share.

That does not mean investors should treat the CPU push as a margin-free upside option. CPUs are a more contested market than frontier GPUs, with entrenched competition from AMD and Intel and with more room for customers to resist a single-vendor stack. Nvidia’s own language around Vera emphasizes efficiency, throughput and AI orchestration rather than a simple x86 replacement story, which is sensible. The stronger thesis is not that Nvidia suddenly conquers the whole CPU market on CPU economics alone. It is that bundling Vera into Rubin-era systems raises Nvidia’s revenue per deployment, protects its platform position, and makes its networking and software layers harder to dislodge.

The China angle adds a second-order effect that investors should not ignore. If Nvidia is still including China in its CPU market math, then management is effectively saying the company sees a path to participate in Chinese compute demand even if the premium GPU lane stays politically constrained. That preserves optionality, but it also sets up a tougher test: whether Nvidia can sell enough compliant or CPU-led infrastructure into China before domestic alternatives become sticky enough to close the door. For now, the company is keeping that market in the model. The burden of proof is on future shipments and disclosed customer wins.

Why investors are paying attention

Nvidia already dominates the AI infrastructure narrative, so the next valuation question is how much more of the stack it can own. Vera matters because it expands the company’s claim from GPU acceleration into the control plane of AI factories. The China detail matters because it shows management still sees one of the world’s largest compute markets as strategically relevant, even if current sales are blocked or delayed. That combination supports the bull case for a larger long-run Nvidia revenue pool, while also sharpening the risk that geopolitical friction or slower-than-expected CPU adoption could limit how quickly that pool turns into booked revenue.

What to watch next

The next checkpoints are practical ones. Investors should watch for evidence that Vera revenue is being disclosed separately or discussed more specifically on future calls, for named hyperscaler or AI-cloud deployments beyond launch-partner lists, and for any update on whether Chinese customers actually begin receiving licensed Nvidia products. Gross margin commentary will matter too, because a CPU-led expansion that deepens platform control is attractive only if it does not dilute the economics that made Nvidia’s GPU story so powerful. If Vera becomes a standard companion to Rubin-era systems, Nvidia’s platform value expands again. If China remains closed and CPU attach rates stay ambiguous, the $200 billion market claim will look more aspirational than investable.