#What are the implications of US restrictions on AI chips for Chinese entities?
The United States is taking significant steps to limit the access of Chinese-owned companies to advanced AI chips produced by major industry players like NVIDIA and AMD. This move extends the existing restrictions that have been in place, which began tightening in 2022 with the Biden administration halting sales of leading AI GPUs, including the A100 and H100 chips.
#How have restrictions evolved over time?
The process began in April 2025 when the Trump administration barred NVIDIA’s H20 and AMD’s MI308 chips from being sold to China without specific licenses. NVIDIA had developed the H20 as a less powerful alternative to stay compliant with prior regulations. By mid-2025, these sales licenses were reinstated, but new rules required both companies to pay 15% of their revenues from any sales linked to China directly to the US government.
By November 2025, there were reports indicating plans to restrict sales of another NVIDIA chip, the B30A. The situation evolved further in January 2026, when new guidelines permitted exports of advanced chips like the H200 and AMD’s MI325X, provided that shipments to China did not exceed half of US volumes. Additionally, the US broadened its Entity List to add more Chinese entities while implementing measures to counteract potential circumvention through third-party countries.
#Why is this relevant for semiconductor companies like NVIDIA and AMD?
These ongoing restrictions are proving financially challenging for companies like NVIDIA and AMD. The requirement to remit a percentage of revenue from sales to the US government adds pressure to profit margins, discouraging engagement with the Chinese market as it becomes less lucrative on a per-unit sales basis.
In contrast, companies such as Huawei are ramping up their chip development efforts to bridge the gap left by restricted access to US suppliers. This shift is particularly critical as it fosters a competitive environment in semiconductor technology, impacting global supply chains.
#How does this affect investors focused on crypto and AI?
The limitations on acquiring NVIDIA chips could significantly alter the landscape of AI computing power distribution. This shift has notable implications for decentralized GPU networks like Render and Akash, which pool hardware resources from various providers globally. Should Chinese-owned entities find themselves unable to access the latest NVIDIA and AMD technologies, there could be a shift in computing resource allocation, favoring Western-aligned regions.
Overall, understanding the evolution of these restrictions provides vital insights for investors. Each new round of controls has been increasingly comprehensive, transitioning from specific chip restrictions to broader performance thresholds and geographic restrictions, thereby reinforcing the complexities of the semiconductor industry.