Nvidia has informed its Chinese customers that shipments of its H200 AI chips could resume by mid-February, according to industry sources, in what would represent the first significant reopening of advanced AI chip exports to China since restrictions tightened under the Biden and Trump administrations.

The chipmaker plans to ship up to 10,000 chip modules—equivalent to approximately 80,000 individual H200 AI chips—to Chinese customers in the initial wave. The announcement comes after months of uncertainty about whether U.S. export controls would permanently lock Nvidia out of the world's second-largest AI market.

The New Trade Framework

The path to resumed exports emerged from the Trump administration's decision to allow H200 sales to China subject to a 25% tariff surcharge. This represents a notable shift from the previous policy of outright export bans on advanced AI chips, though the tariff adds significant cost to Chinese buyers.

For Nvidia, the financial stakes are substantial. Despite existing export restrictions, the company generated $17.1 billion in revenue from China in calendar 2024, though it estimates the controls cost an additional $8 billion in lost sales. Resuming H200 exports could help recapture some of that lost business.

"The China market remains important to us. The H200 is designed to comply with export regulations while still delivering significant AI performance for Chinese customers."

— Nvidia spokesperson discussing the potential export resumption

Beijing's Complicated Response

However, the path to resumed sales isn't straightforward. Chinese government officials have instructed domestic technology companies not to purchase H200 chips "unless necessary," effectively creating a soft barrier to imports even where U.S. law permits sales.

Beijing reportedly will only allow H200 imports under special circumstances, such as university research or development laboratories. The restriction appears aimed at reducing dependence on American semiconductors while domestic chip programs accelerate.

This creates an unusual dynamic: the U.S. is offering to sell, but China is reluctant to buy. The posture reverses the typical export control dynamic and reflects Beijing's determination to achieve semiconductor self-sufficiency regardless of near-term costs.

The Competitive Landscape Shifts

China's hesitance to embrace H200 imports also reflects the rapid advancement of domestic AI chip capabilities. Companies including Huawei have developed AI accelerators that, while not matching Nvidia's performance, provide acceptable alternatives for many applications. The longer export controls persist, the more entrenched these domestic alternatives become.

Meanwhile, Nvidia faces intensifying competition in markets where it can freely sell. AMD has gained ground with its MI300 series, and custom chips developed by cloud giants like Google and Amazon are capturing workloads that would have previously gone to Nvidia hardware.

The $275 Billion Backlog

Despite the China uncertainty, Nvidia's overall business remains extraordinarily strong. The company has accumulated a data center chip backlog of approximately $275 billion, representing orders that will take years to fulfill. Demand from U.S. hyperscalers—Microsoft, Amazon, Google, and Meta—shows no signs of slowing.

Key metrics underscore Nvidia's dominance:

  • Market share: Approximately 80% of AI training chips globally
  • Quarterly revenue: Over $30 billion in data center sales alone
  • Production capacity: Sold out through 2026 for most products
  • Analyst targets: 69 analysts tracking the stock with median price target of $250

Strategic Implications

The potential resumption of China sales carries strategic implications beyond Nvidia's bottom line. For the U.S. government, allowing exports—even with tariffs—represents a calculation that completely cutting off China may accelerate domestic chip development without significantly degrading Chinese AI capabilities.

For China, the decision about whether to purchase H200 chips involves trade-offs between near-term AI development and long-term technology independence. Buying American chips provides immediate capability but perpetuates dependence on foreign technology.

For investors, the China situation represents both opportunity and risk. If exports resume smoothly, Nvidia could see a meaningful revenue boost. But if Beijing's restrictions tighten or U.S. policy reverses, the company could face renewed revenue pressure.

What Happens Next

The mid-February timeline suggests Nvidia is confident in the regulatory pathway, but multiple variables remain uncertain. Chinese customs enforcement of Beijing's purchasing guidelines, U.S. Commerce Department oversight of the tariff mechanism, and broader geopolitical developments could all affect whether chips actually flow.

Investors should expect volatility around any headlines related to China chip policy. The H200 export situation exemplifies the broader challenge facing technology companies navigating great power competition: business opportunities exist in China, but accessing them requires threading an increasingly narrow needle between two governments with conflicting objectives.