This is not investment advice. The author has no position in any of the stocks mentioned. Wccftech.com has a disclosure and ethics policy.
After NVIDIA revealed yesterday that it expects the Trump administration to grant export licenses for the H20 GPU sales to China, research firm TrendForce believes that domestic Chinese chip designers and semiconductor firms should continue developing chips in parallel. US restrictions on the H20 were levied earlier this year and saw a sharp response from NVIDIA CEO Jensen Huang, who asserted that chip restrictions on China would only spur the local ecosystem's development. TrendForce reports that it now expects China's procurement ratio for foreign AI chips to sit at 49% for a seven percentage point jump over the earlier 42%.
NVIDIA's H20 Chip Will Experience Significant Pent Up Demand In China, Says Report
TrendForce agrees with the widely held assertion that despite stringent efforts, domestic Chinese firms such as Huawei have been unable to catch up with NVIDIA in technology. One of the biggest constraints that Chinese firms face when developing AI chips is their inability to rely on the latest chip manufacturing technologies due to US restrictions on the sale of advanced lithography equipment to China. Huawei is also restricted from buying advanced chips from Taiwan's TSMC, which means that the firm is restricted to manufacturing all of its chips using older technology.
As a result, the "resumption of H20 shipments is expected to unlock significant pent-up demand for AI infrastructure deployment in China, especially among major CSPs," believes TrendForce. Another factor that will drive the demand for the H20 is high-bandwidth memory (HBM). According to the firm, while H20 chips shipped in 2024 used HBM3 8hi memory, this is expected to be upgraded to HBM 3e 8hi in 2025 and spur greater interest due to Chinese domestic firms facing export restrictions.
As part of his discussions with the Trump administration to allow NVIDIA to sell the H20 GPUs to China, Huang asserted that the sales resumption would enable American products to form the backbone of the global AI hardware base. He likened the GPUs to the US dollar and outlined that the GPUs could become the global reserve in AI computing.
NVIDIA's H20 GPUs already face strict design rules that degrade their performance to meet US export guidelines. They belong to the older Hopper architecture, while American firms and non-sanctioned countries can buy products designed and manufactured through the Blackwell architecture.
Huang has also countered US national security concerns by pointing out that not only will the Chinese military not rely on foreign-source chips for its computing needs, but added that local chip designers will only expedite their efforts to develop competing products in the case of tight restrictions. However, as per TrendForce, while the H20 sales resumption to China will lead local firms "to stockpile aggressively," the country's "AI market remains vulnerable to geopolitical uncertainty, and risks to NVIDIA persist."
As a result, the research firm concludes that as Chinese OEMs and CSPs continue to stockpile NVIDIA products, "domestic AI chipmakers and ecosystem players—backed by government support—will continue to accelerate their development in parallel."
About the author: Ramish is a seasoned technology writer and editor with more than a decade of experience. He specializes in semiconductor fabrication and market analysis. With a background in finance and supply chain management - via his bachelors in Finance and a micromasters in supply chain management from MIT - Ramish combines financial rigor with deep industry insight to deliver accurate and authoritative coverage.
Follow Wccftech on Google to get more of our news coverage in your feeds.
