Nvidia CEO Jensen Huang made his most direct public statements yet about the Chinese market during his visit to Taipei for Computex 2026, confirming that the company’s $200 billion CPU market forecast explicitly includes China and expressing a clear desire to deliver its advanced H200 AI accelerators to Chinese customers. The comments, reported by Reuters and CNBC, are being closely watched by investors and policymakers alike as a signal of Nvidia’s strategic priorities amid intense geopolitical pressure.
The $200 Billion Forecast and China’s Role
Huang’s $200 billion CPU market forecast is one of the most ambitious revenue projections in the semiconductor industry. The figure encompasses the total addressable market for data center CPUs as AI workloads drive a fundamental shift in computing architecture. By explicitly confirming that China is included in this forecast, Huang acknowledges that any realistic assessment of the global AI chip market must account for Chinese demand, which, by most estimates, accounts for 15-20% of global AI infrastructure spending.
The inclusion of China in the forecast is commercially significant. Nvidia has faced sustained pressure from U.S. export controls that have progressively restricted the chips it can sell in China, culminating in the effective prohibition of its most advanced products including the H100 and H200. As EastFrontier reported in May, Nvidia’s market share in China’s AI chip segment has fallen to near zero for high-end accelerators, with Huawei filling the gap. Huang’s forecast language suggests Nvidia still views China as a recoverable market rather than a permanent loss.
“Would Be Terrific”: The H200 Statement
The most striking element of Huang’s Taipei remarks was his comment that delivering H200 chips to China “would be terrific.” The H200 is Nvidia’s current flagship AI accelerator, offering substantially higher memory bandwidth than the H100 and designed specifically for the demands of large language model inference. Under current U.S. export control regulations, the H200 cannot be sold to Chinese entities.
Huang’s phrasing is carefully chosen, he is expressing a commercial preference without making a policy commitment or suggesting Nvidia would circumvent regulations. But the statement is nonetheless notable for its directness. Previous Nvidia communications on the China chip restriction have been more circumspect, emphasizing compliance and the development of China-specific products that fall below export control thresholds. By saying H200 deliveries “would be terrific,” Huang is publicly signaling to the U.S. government that the current restrictions are costing Nvidia real revenue and market position.
The Vera Rubin Announcement and China’s Implications
Huang’s Taipei visit also featured the announcement of Nvidia’s next-generation Vera Rubin GPU architecture, which will succeed the current Blackwell generation. The Vera Rubin chip is expected to deliver another significant performance leap for AI training and inference workloads. However, given the trajectory of U.S. export controls, there is no certainty that Chinese customers will be able to access Vera Rubin hardware upon shipment.
This creates a strategic dilemma for Nvidia. Each new generation of chips that Chinese customers cannot access accelerates the development of domestic Chinese alternatives. Huawei’s Ascend series has already demonstrated that China can produce competitive AI accelerators at scale. As EastFrontier has noted, Beijing has actively used regulatory pressure to push Chinese tech giants toward domestic chips, a trend that benefits Huawei and disadvantages Nvidia regardless of what export controls allow.
Navigating the Geopolitical Tightrope
Huang’s comments in Taipei reflect the impossible position that Nvidia, and many other U.S. technology companies, occupies in the current geopolitical environment. The company is simultaneously dependent on Taiwan’s TSMC for manufacturing, reliant on Chinese customers for a significant portion of its historical revenue, and subject to U.S. government restrictions that limit what it can sell and where.
The $200 billion forecast including China is not merely an investor relations statement — it is a signal to Washington that the current export control regime has costs that extend beyond Nvidia’s balance sheet to the broader question of whether American companies can maintain their leading position in the global AI hardware market. If Chinese customers are permanently redirected to Huawei and other domestic suppliers, the long-term consequences for U.S. semiconductor leadership could be significant.
Huang’s Computex 2026 visit also highlighted the degree to which Nvidia’s fortunes are intertwined with Taiwan’s semiconductor ecosystem. TSMC remains the sole manufacturer capable of producing Nvidia’s most advanced chips at the required scale and yield, and the Taiwan supply chain, encompassing not just TSMC but packaging specialists, substrate manufacturers, and advanced testing facilities, is a critical dependency that Nvidia cannot easily replicate elsewhere. The $10 billion investment that AMD announced in Taiwan the same week, as EastFrontier reported, underscores how deeply the world’s leading AI chip companies are committed to the Taiwan manufacturing ecosystem. For Nvidia, maintaining its relationship with Taiwan while navigating U.S. export controls on China is a geopolitical balancing act that will define the company’s strategic options for the foreseeable future.