Nvidia H200 China Sales Remain Stalled Despite US Approval

Nvidia has secured approval from the United States to sell its H200 artificial intelligence chips to around 10 Chinese companies. However, no deliveries have taken place so far, leaving major technology deals frozen as chief executive Jensen Huang seeks progress during a visit to Beijing.

The stalled sales underline the growing strain created by the technology rivalry between the United States and China. Although Washington approved the transactions, uncertainty on both sides has delayed implementation.

Huang Joins Trump Delegation

Huang joined a delegation accompanying Donald Trump to talks with Xi Jinping after receiving an invitation from the US president. According to a source, Trump picked Huang up in Alaska while travelling to the summit.

The visit has raised expectations that negotiations in Beijing could help unlock the blocked H200 sales. Speaking to Chinese state broadcaster CCTV on Thursday, Huang expressed hope that Trump and Xi would strengthen relations during their discussions.

Chinese Tech Firms Receive Approval

The US Commerce Department has reportedly approved sales of Nvidia’s H200 chips to several leading Chinese technology firms. Approved companies include Alibaba, Tencent, ByteDance and JD.com.

In addition, distributors such as Lenovo and Foxconn have also received approval to handle the chips in China. Sources said each approved buyer could purchase up to 75,000 chips under the licensing terms.

Lenovo confirmed that it was among the companies authorised to sell the H200 in China as part of Nvidia’s export licence arrangements.

Beijing Hesitates Over Imports

Despite receiving approval from Washington, Chinese firms have reportedly delayed purchases following guidance from Beijing. Sources said concerns increased after changes on the US side, although the precise nature of those changes remains unclear.

Chinese officials are reportedly worried that dependence on imported Nvidia chips could weaken efforts to build a domestic semiconductor industry. As a result, Beijing has intensified support for locally developed AI processors.

Chinese firms increasingly promote domestic alternatives, including chips produced by Huawei. Meanwhile, companies such as DeepSeek have highlighted their use of Chinese-made hardware to reduce reliance on foreign suppliers.

Export Controls Create Obstacles

The sales process has become more complicated because of strict requirements imposed by both governments. US export rules introduced in January require Chinese buyers to prove they have strong security procedures and will not use the chips for military purposes.

Nvidia must also certify that sufficient inventory remains within the United States. Furthermore, Trump negotiated an arrangement allowing the US government to receive 25% of the revenue generated from the chip sales.

That structure requires the chips to pass through US territory before shipment to China. Consequently, some officials in Beijing have expressed concerns about potential tampering or hidden vulnerabilities in the supply chain.

China has also increased scrutiny of foreign technology following new supply chain security regulations issued by the State Council. According to sources, the measures aim to reduce foreign dependence in critical technology infrastructure.

Nvidia Faces Pressure From Both Sides

Before tighter export controls took effect, Nvidia controlled roughly 95% of China’s advanced AI chip market. China previously accounted for 13% of the company’s revenue, while Huang estimated the country’s AI market alone could reach $50 billion this year.

However, Nvidia’s position in China has weakened significantly. Huang has warned that US export controls have effectively reduced the company’s share of China’s AI accelerator market to zero.

The continuing delays have also strengthened criticism from China-focused policy hawks in Washington. Chris McGuire of the Council on Foreign Relations argued that allowing Nvidia to sell more chips to China could narrow the United States’ lead in artificial intelligence technology.

With inputs from Reuters