Highlights: 

  • Beijing has approved Nvidia's H200 chips for sale to Chinese customers 
  • CEO Jensen Huang confirms H200 production has restarted with Chinese purchase orders in hand 
  • Nvidia is adapting its newly acquired Groq AI chip for the Chinese market, with availability expected in May 
  • The Groq chips are not downgraded versions; they will be adapted for system compatibility 
  • China once accounted for 13% of Nvidia's total revenue 

 

In a landmark development for the global AI industry, Nvidia has secured Beijing's long-awaited approval to sell its H200 artificial intelligence chips in China and the company is not stopping there. According to sources, Nvidia is also preparing a China-compatible version of its Groq AI chip, marking a dramatic re-entry into one of the world's most lucrative technology markets. 

H200 Clearance: A Major Breakthrough for Nvidia 

Nvidia has won Beijing's approval to sell its second-most powerful AI chips to China, paving the way for the U.S. chipmaker to resume sales of the H200, a chip that has become a major flashpoint in U.S.-China tech relations.  

Speaking to reporters at Nvidia's GTC developer conference in San Jose, California, CEO Jensen Huang confirmed: "We've been licensed for many customers in China for H200. We have received purchase orders from many customers, and we're in the process of restarting our manufacturing."  

This is significant news for Nvidia, which had been forced to halt H200 production after facing mounting regulatory uncertainty on both sides of the Pacific. Despite strong demand from Chinese firms and U.S. approval for exports, Beijing's hesitation to allow imports had been the main barrier to H200 shipments reaching China.  

In a filing with the U.S. Securities and Exchange Commission late last month, Nvidia confirmed that the U.S. had granted a licence in February allowing small amounts of H200 products to reach specific China-based customers.  

The Groq Chip: Nvidia's Next China Play 

Beyond the H200, Nvidia is making a bolder move by preparing a version of its Groq AI chip for the Chinese market, following its $20 billion licensing deal with the startup late last year, and unveiled a new line of chips from the startup earlier this week at GTC.  

The chips being readied for China are not downgraded versions made specifically for the Chinese market. However, the new variant can be adapted to work with other systems, and the Groq chip is expected to be available in May.  

Nvidia plans to use the Groq chips for inference, the process by which AI systems respond to user prompts, write code, or carry out tasks. This positions the Groq chip as a key inference engine for Chinese hyperscalers at a time when demand for AI compute is surging across the country. 

Why China Matters So Much to Nvidia 

This marks a dramatic shift from the Biden era chip controls, when Nvidia could only sell downgraded H20 chips in the Chinese market, significantly hurting its revenue from China. The H200, by contrast, is almost six times more powerful than the H20 chip previously available to China.  

Chinese tech giants like Alibaba, ByteDance, and Tencent are reportedly preparing to order a combined 400,000 H200 AI GPUs once Beijing's approvals are fully in place. Smaller AI startups are also expected to place orders for chips crucial to training models that compete against OpenAI's ChatGPT.  

The Geopolitical Tightrope 

Beijing's approval comes with strategic caveats. China will continue to incentivise the growth of domestic chipmakers even with controlled access to H200 chips. The Chinese government has been pressing local companies to support homegrown alternatives, particularly those from Huawei. Huawei's Ascend 910C lags behind Nvidia's H200 in performance metrics, and Huawei's current timeline will not see it produce a comparable AI chip until the end of 2027.  

This strategic context explains China's calculated decision: allow Nvidia back in to close an urgent capability gap, while continuing to invest in domestic semiconductor self-sufficiency. 

Frequently Asked Questions 

  1. What is the Nvidia H200 chip?  

The H200 is Nvidia's second-most powerful AI chip currently available for sale, based on the Hopper GPU architecture. It is primarily used for training and running large AI models. 

  1. Why did China block H200 imports previously?  

Beijing had been pushing Chinese tech companies toward domestically produced chips to reduce dependence on foreign semiconductors, even as U.S. approvals for H200 exports were already in place. 

  1. What is the Groq chip and how does it differ from the H200? 

 Groq chips, developed by the AI startup Nvidia struck a $20 billion licensing deal with in late 2025, are specialised for AI inference, the stage where an AI model generates responses or performs tasks. They complement training-focused chips like the H200 

  1. When will Groq chips be available in China?  

According to sources, the China-compatible version of the Groq chip is expected to be available in May 2026. 

  1. Does this mean Nvidia has fully returned to the Chinese market?  

Not entirely. Access remains conditional, with Beijing retaining significant control over licensing. However, the H200 approval and Groq chip preparation represent the most substantial re-entry Nvidia has achieved in several years.