Nvidia’s Worst Nightmare, in Jensen Huang’s Own Words: “Zero Percent” in China

by Raphael Dudler | Jun 27, 2026 | CHN AI NEWS

The CEO of the world's most valuable company admits US export policy "has already largely backfired" — and names the outcome that terrifies him most

In a nutshell

Nvidia's CEO said it himself: "zero percent" in China, and US policy "has already largely backfired." But the market share isn't what scares him most.

Our members-only forecast explains why the CUDA-versus-CANN software battle — not chip speed — will decide China's AI future, predicts when DeepSeek will trigger Huang's named nightmare by going "Huawei-first," and assesses why Nvidia's aggressive lobbying could force a partial reversal of US export policy by 2027.

From 25% to Zero

There is no spin left to apply. Jensen Huang, the chief executive of Nvidia — the most valuable company on earth, the firm whose chips power the global AI revolution — has stated plainly that his company now holds essentially no position in the Chinese market it once dominated. Nvidia's China operations once accounted for up to 25% of its data center business revenue. Huang's current assessment of that market share: "zero percent." stanford
His verdict on the policy that produced this collapse is equally blunt. Jensen Huang has said US export policy "has already largely backfired" — and, in a striking acknowledgement of how completely the competitive landscape has shifted, stated that Nvidia has "largely conceded" China's advanced artificial intelligence chip market to Huawei.

This is not an analyst's prediction or a competitor's boast. It is the leader of the dominant company in the field conceding the world's second-largest AI market to a Chinese rival that, five years ago, US sanctions were designed to cripple.

The Sentence That Defines the Stakes

Buried in a recent interview is the single sentence that captures what actually keeps Nvidia's leadership awake at night — and it has nothing to do with quarterly revenue.

"The day that DeepSeek comes out on Huawei first, that is a horrible outcome for our nation," Huang told podcaster Dwarkesh Patel. "It could lead to a scenario where AI models around the world are developed and they run best on non-American hardware."

This is the strategic nightmare in its purest form. As watchchina.ai has documented, the gamble is already paying off for Huawei, and it began with DeepSeek. While DeepSeek's latest V4 model was trained on Nvidia chips, the company used Huawei's 950PR processor for inference — confirmed last month. If the world's most influential open-source AI models are optimised to run best on Chinese hardware, then every developer, every startup, and every enterprise that builds on those models is drawn, gradually and then suddenly, into a Chinese hardware ecosystem — regardless of US export policy. The chips become irrelevant if the software ecosystem migrates.

Huawei's Clever Flank: Winning the War That Actually Matters

Huawei's strategy is not to beat Nvidia at Nvidia's strongest game. It is to win the game that is growing fastest. Nvidia's chips remain more advanced, but Huawei has improved its chips enough to target a rapidly growing part of the AI market: inference — the computation AI models use to generate answers and perform real-world tasks after training is complete. Huawei believes inference will become the biggest source of AI computing demand as AI assistants and autonomous agents become more common, and since inference workloads are generally less demanding than training massive models, Huawei can remain competitive despite weaker raw chip performance.

The commercial momentum confirms the strategy is working. Huawei expects AI chip revenue to surge to $12 billion in 2026, up from $7.5 billion in 2025 — a 60% increase — based on existing orders for its 950PR chip, which entered mass production last month. Huawei's vice president has publicly committed to a release cadence that directly challenges the pace Nvidia has always controlled: "The Ascend chip is evolving at a rate of one generation per year, doubling the computing power," Huawei VP Chen Lin stated without hedging — a public commitment aggressive enough to redefine how AI chip progress is measured.
Nvidia retains one critical advantage, and it is the reason the war is not yet lost. Nvidia's software ecosystem remains a major moat: while Huawei has been developing its CANN platform as a domestic alternative to Nvidia's widely used CUDA software, developers say it still lags significantly behind in usability and maturity, increasing development complexity and operating costs. The battle for China's AI future, in the end, may be decided not by who makes the fastest chip — but by whose software developers actually choose to build on.

Sources: Tom's Hardware, CNBC, TechRadar, NBC News, Financial Times, Dwarkesh Patel interview

Strategic Analysis — For Members Only

🔒 This analysis is for watchchina.ai Intelligence members only.

→ Become a Member

Already a member? Log in here

"