Key Takeaways
- DeepSeek’s V4 artificial intelligence model deployed predominantly on Huawei’s Ascend processors rather than Nvidia hardware.
- Zero H200 processors have reached Chinese buyers despite receiving US government authorization.
- Delivery holdups continue due to unresolved disputes between Washington and Beijing regarding transaction conditions.
- The Chinese AI infrastructure sector represents a $50 billion annual market expanding by 50% per year.
- Nvidia shares dropped 1.41% during regular trading, though they climbed 0.8% in Friday’s pre-market session.
Nvidia’s efforts to penetrate the Chinese market encountered fresh obstacles this week when DeepSeek unveiled its V4 model—running predominantly on Huawei’s silicon rather than American GPUs.
DeepSeek introduced preview iterations of its V4 architecture on Friday. Industry observers monitored the rollout carefully following the disruption caused by its previous version, which shocked markets last year by demonstrating sophisticated performance while requiring minimal training expenditure.
Huawei wasted no time capitalizing on the opportunity. The Shenzhen-based technology powerhouse announced via WeChat that its complete Ascend AI processor portfolio now accommodates DeepSeek V4 models. While DeepSeek acknowledged validating one of V4’s crucial efficiency methodologies on both Nvidia GPUs and Huawei silicon, the preferred hardware platform clearly skews toward Chinese-manufactured chips.
This development compounds challenges for Nvidia, which has been effectively excluded from China’s premium AI chip sector through US export controls.
H200 Deliveries Remain Blocked
CEO Jensen Huang announced last month that Nvidia had resumed production of its H200 chips with potential Chinese sales in mind, noting that several customers had submitted orders. However, Reuters revealed this week that zero H200 units have actually been delivered to any Chinese purchaser.
The Trump administration granted formal authorization for H200 exports to China, but implementation has stalled. Conflicting positions between US and Chinese officials regarding specific transaction parameters have prevented any shipments, according to Reuters sources.
Chinese companies interested in purchasing the chips also face bureaucratic hurdles obtaining approval from their own government authorities.
Financial Implications
The potential revenue at risk is substantial. Huang has characterized China’s AI infrastructure landscape as a $50 billion yearly market experiencing 50% annual growth.
KeyBanc analyst John Vinh projects that if regulatory barriers disappeared, Chinese enterprises would purchase approximately 1.5 million H200 processors this year. Such volume would generate around $30 billion in Nvidia revenue.
Currently, actual sales figures remain at zero.
Nvidia’s stock declined 1.41% during Thursday’s trading session, though shares had risen 0.8% in Friday’s pre-market activity before Reuters published its report detailing the H200 shipment standstill.
DeepSeek’s V4 launch introduces additional competitive pressure. Should Chinese AI developers increasingly standardize on Huawei’s Ascend ecosystem, Nvidia’s market access could diminish further—regardless of whether trade and regulatory obstacles eventually resolve.
The H200 processors have been manufactured and await distribution. Whether Chinese customers ever receive them hinges on diplomatic negotiations that currently show no signs of meaningful progress.
As of Friday, Nvidia has confirmed zero H200 transactions with Chinese companies, with Reuters reporting continued shipment delays stemming from unresolved disagreements between American and Chinese government officials regarding sales stipulations.


