Nvidia’s License to Lose: The $4 Billion China Opportunity That Won’t Materialise

Nvidia’s H200 China export licenses yield zero revenue; market share collapsed to zero. Stock flat at €181, yet analysts target 43% upside, ignoring $3.5-4B potential from a reopening.

Nvidia’s stock is treading water at €181.00, barely 3% above its 50-day moving average of €177.93. The RSI of 49.4 screams neutrality. On the surface, this is a market waiting for a catalyst. Below the surface, the waiting has a name: China — and it has become the most expensive no-show in the semiconductor industry.

The numbers tell a clear story. The H200 export licence from Washington is, on paper, a green light. Ten Chinese companies — Alibaba, Tencent, ByteDance, JD.com, and distributors Lenovo and Foxconn among them — have been approved to buy up to 75,000 chips each. Yet after months of approvals, Nvidia’s CFO confirmed on the latest earnings call that the company has booked exactly zero dollars in revenue from these licences. “We do not know whether any imports will be allowed into China,” he said. That is not a supply chain problem; it is a policy vacuum.

A Deal That Exists Only on Paper

The mechanism behind the stalemate is as convoluted as it is effective. The Trump administration’s framework requires 25% of chip revenues to flow back to the US, a condition Beijing views as a security threat. China’s state-directed push to domestic semiconductor independence has only accelerated. Huawei, Alibaba and Baidu are now developing their own AI processors. Cities like Hangzhou subsidise companies that run models on Chinese-made chips. The financial incentive to avoid American technology is now baked into China’s industrial policy.

Jensen Huang acknowledged in April 2026 that Nvidia’s market share in China has effectively collapsed to zero. That is a remarkable fall from the 95% share the company once commanded in the advanced AI chip market, and a far cry from the 13% of total revenue China contributed as recently as fiscal 2025. Nvidia’s own quarterly filing with the SEC is blunt: the company cannot currently offer a competitive product for Chinese data centres that satisfies both Washington and Beijing simultaneously.

The Trillion-Dollar Blind Spot

Wall Street’s consensus price target of €257.45 — roughly 43% above current levels — is built on the assumption that the AI build-out is unstoppable. And it may be. Nvidia still holds an estimated 85-92% of the AI accelerator market. AMD and Intel together command only single-digit share in training. The Vera Rubin platform, already deployed at Microsoft Azure in its first rack, promises inference at up to ten times lower cost per token than Blackwell and four times fewer GPUs for Mixture-of-Experts models. Full production deliveries are scheduled for the second half of 2026.

But that rosy outlook ignores a gap worth billions. Nvidia’s revenue forecasts assume zero H200 sales to China. If the export channel ever actually opened, the company would regain $3.5-4 billion in annual revenue. That optionality is currently priced at exactly zero by the market. Analysts’ targets do not fully account for the asymmetry — either the upside of a reopening or the downside of a permanent lockout.

Waiting for a Decision

The stock has gained roughly 45% in the past 12 months and a little over 12% year to date. That performance has been fuelled by relentless AI demand, not by any resolution of the China standoff. The 30-day annualised volatility of 42% shows how quickly geopolitical headlines can move the share price. Huang himself travelled to China alongside Trump for a summit with Xi Jinping in May 2026. No breakthrough emerged. The market remains in limbo, with the stock hovering just shy of the level where a China reopening would provide genuine momentum.

Meanwhile, the structural dynamics are shifting. Hyperscalers like Amazon, Google and Microsoft are building their own custom silicon, with their combined market share expected to rise from 21% in 2025 to 28% in 2026. Yet these same companies remain Nvidia’s largest customers — a contradiction that has so far always resolved in the chip giant’s favour.

For a company of Nvidia’s size and technological dominance, China is not an existential threat. It is a persistent drag. A market that once contributed 13% of sales, now effectively zero, and that could theoretically reopen at any moment. That optionality is unpriced, unresolved — and the most interesting part of the stock right now. The next big move will come when Washington or Beijing decides to end the waiting game. Until then, Nvidia is a license to lose $4 billion a year.