Global 2nm Chips Are Here. Where Does China’s Semiconductor Industry Stand?
The global semiconductor industry has reached a new milestone. TSMC, Samsung, and Intel have all started mass production of 2nm-class chips in 2026. These chips represent a major technological shift from previous generations, using a new transistor architecture called gate-all-around (GAA) that improves both performance and power efficiency.
China, by contrast, remains at the 7nm level for its most advanced commercially produced chips. The country’s leading foundry, SMIC, first achieved this with the Huawei Mate 60 series in 2023. Now a second Chinese manufacturer, Hua Hong Group, is reportedly preparing small-scale 7nm production at its Shanghai facility, with initial capacity of a few thousand wafers per month expected by the end of 2026. Industry sources indicate Huawei has been involved in supporting this new production line.
A separate development involves chip design rather than manufacturing. Xiaomi announced in late August 2026 that it would use a self-designed 3nm chip called Xuanjie O3 in its upcoming Xiaomi 18 Fold phone, scheduled for a September launch. The chip, manufactured by TSMC, marks the first time a Chinese smartphone brand has placed its own flagship processor in a top-tier foldable device. With reported 24 billion transistors and performance gains of 60% for the CPU and 85% for the GPU, it represents a significant design achievement. However, it does not reflect China’s domestic manufacturing capability, as the chip is produced abroad.
The gap between China’s manufacturing capability and global leaders is significant. TSMC’s 2nm process achieves a transistor density roughly 2.5 times higher than SMIC’s current 7nm-class production. The performance difference in real-world applications is estimated at about four to five years behind global leaders.
What holds China back is equipment. Without access to ASML’s extreme ultraviolet (EUV) lithography machines due to US export controls, Chinese manufacturers must use older deep ultraviolet (DUV) tools with a technique called multi-patterning. This means exposing a wafer multiple times to create finer circuits. It works, but it increases production costs and reduces the number of usable chips per wafer. A Goldman Sachs analysis noted that China’s domestic lithography technology is roughly 20 years behind ASML’s capabilities, though this assessment has been debated within the industry.
Despite these constraints, China is making progress. Goldman Sachs projects that the supply-demand gap for 7nm and below chips in China will narrow from 92% in 2025 to 34% by 2035, driven largely by SMIC’s capacity expansion. Chinese semiconductor capital spending is expected to reach $82 billion by 2030, up 79% from earlier forecasts. The country’s chip self-sufficiency rate by volume will reach roughly 70% by mid-2026, though value-based self-sufficiency remains much lower.
For Chinese tech companies like Huawei, having a second domestic 7nm supplier provides some relief. SMIC’s fabs have been running at over 90% capacity, and the company has prioritized AI server chips over smartphone processors. Hua Hong’s new capacity could help balance that supply.
The technology gap remains substantial, and the cost disadvantage of multi-patterning production makes Chinese 7nm chips more expensive to produce than comparable chips from TSMC. For now, the two semiconductor ecosystems are operating at different tiers, with the global leaders pushing toward 1.4nm while China consolidates its position at 7nm and works on domestic equipment development. The Xiaomi announcement shows that Chinese companies can compete in chip design, but manufacturing constraints remain the fundamental bottleneck. The next five years will show whether alternative architectures and advanced packaging can help China narrow the gap, or whether the technology divide continues to widen.
