Nvidia CEO Jensen Huang just told the world that the company’s $200 billion total addressable market forecast for CPUs includes demand from China. In the same breath, Nvidia’s own guidance for next quarter excludes any data center compute revenue from China entirely.

The big number and the big asterisk

Speaking in Taipei on May 23 following Nvidia’s Q1 FY2027 earnings release, Huang laid out the company’s vision for its Vera CPU platform. The new chip line is built specifically for agentic AI workloads, and Nvidia sees nearly $20 billion in revenue visibility for Vera CPUs in 2026 alone.

The broader claim is even more ambitious. Nvidia is sizing the entire CPU market opportunity at $200 billion, a figure that explicitly factors in Chinese demand. This is notable because Nvidia has historically been a GPU-first company, dominating the AI accelerator market before making this aggressive push into CPUs, a space long controlled by established players like Intel and AMD.

While Huang is happy to include China in the long-term addressable market, Nvidia’s Q2 guidance tells a different story. The company has stripped out any projections for data center compute revenue from China, citing US export controls that continue to restrict sales of advanced AI chips to Chinese buyers.

Record earnings meet geopolitical reality

Nvidia just posted Q1 FY2027 revenue of $81.6 billion, a record quarter. The company has noted that H200 AI chip licenses have been approved but are still awaiting further clearance from China’s side.

Huang was deliberate in distinguishing between short-term earnings visibility and long-term growth potential. Nvidia is positioning itself to own the full AI infrastructure stack, from GPUs to CPUs to networking.

What this means for investors

The bullish case is straightforward. Nvidia posted record revenue of $81.6 billion in a single quarter, is launching a new product line with $20 billion in near-term visibility, and sees a $200 billion addressable market that includes the world’s second-largest economy.

The cautious case is equally compelling. Including China in a long-term TAM estimate while explicitly excluding Chinese revenue from next-quarter guidance is a tension that won’t resolve itself quickly. The H200 licensing situation, approved on one end but waiting for clearance on the other, illustrates how bilateral friction can create indefinite delays.