Nvidia Stays Clear of China Amid Tight U.S. Export Controls

TLDRs;

• Nvidia CEO Jensen Huang says there are no plans to resume AI chip sales to China.
• U.S. export curbs continue to block Nvidia’s high-end GPU shipments to Chinese firms.
• China accounted for 13% of Nvidia’s 2025 revenue before bans crushed H100 and H20 chip sales.
• Despite setbacks, Nvidia’s $5 trillion market value underscores global AI chip dominance.

Nvidia CEO Jensen Huang has ruled out any near-term resumption of AI chip sales to China, emphasizing the company’s adherence to U.S. export controls that restrict advanced semiconductor shipments to Chinese firms.

Speaking in Tainan, Taiwan, Thursday Huang dismissed circulating rumors suggesting Nvidia might return to the Chinese market through its next-generation Blackwell chips.

Nvidia, the world’s most valuable company with a market capitalization exceeding $5 trillion, remains at the epicenter of the global AI boom. However, U.S. trade policy continues to limit its ability to engage with China, once one of its most lucrative markets.

U.S.-China Tensions Freeze Advanced Chip Trade

Nvidia’s absence from the Chinese market stems from stringent U.S. government export restrictions targeting high-performance GPUs used in artificial intelligence and data center applications. These curbs, initially implemented in 2022 and later expanded, effectively block the sale of Nvidia’s flagship chips such as the H100, A100, and H20.

A recent round of trade discussions between Washington and Beijing brought no relief to these constraints. U.S. officials signaled that no regulatory changes are expected soon, leaving Nvidia and other American chipmakers in a holding pattern.

The impact has been substantial. Nvidia’s China-related revenue plunged from $17 billion in fiscal 2025, representing 13% of its total $130.5 billion, to near-zero in 2026 for its most advanced products. The company even recorded $4.5 billion in charges linked to the halted H20 GPU shipments earlier this year.

Strategic Focus Shifts to Hyperscalers and AI Factories

As geopolitical friction sidelines China, Nvidia is doubling down on its relationships with hyperscalers (massive cloud providers like AWS, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure) as well as specialized GPU cloud companies such as CoreWeave.

Its latest innovation, the GB300 NVL72 system, connects 72 Blackwell Ultra GPUs with 36 Grace CPUs, promising 1.5x faster performance than its predecessor and a 50x boost in potential revenue for large-scale AI clusters. Orders for these systems are already projected to surpass $500 billion through 2026, according to industry sources.

CoreWeave is among the first to roll out Nvidia’s GB200 NVL72 chips for major clients like IBM, Mistral AI, and Cohere, signaling that the U.S. and European AI ecosystem remains Nvidia’s most profitable frontier.

Nvidia’s Global Tour Strengthens Supply Chain Ties

Huang’s stop in Taiwan, home to Taiwan Semiconductor Manufacturing Co. (TSMC), Nvidia’s critical manufacturing partner, is part of a broader global tour spanning Washington and South Korea. The meetings underscore Nvidia’s push to secure its supply chain amid growing geopolitical uncertainty and rising competition from domestic Chinese chipmakers.

With the AI race accelerating, Nvidia’s long-term strategy appears clear: focus on compliant markets, lead in innovation, and expand AI infrastructure globally. While China remains a $50 billion AI market opportunity, Huang’s remarks suggest Nvidia is prepared to wait out the political storm.

The post Nvidia Stays Clear of China Amid Tight U.S. Export Controls appeared first on CoinCentral.