The discussion about Chinese semiconductors often falls into a comfortable trap for the West: measuring everything by the most advanced node. If China isn’t producing at an industrial scale in 2 nanometers, the conclusion is that they still have a long way to go and that true leadership remains intact. That interpretation is partly correct, but it’s increasingly less useful for understanding what’s really happening.
Because the competitive threat doesn’t only come from the top. It mainly comes from the bottom and volume. While the U.S. and parts of Europe focus the debate on preventing Beijing from accessing frontier chips and state-of-the-art manufacturing equipment, China has taken that time to strengthen where it already could compete: mature chips, power semiconductors, those that drive cars, appliances, industrial automation, and much of the real electronic infrastructure supporting the economy. That’s where it’s gaining strength, and where it can do damage.
The U.S. export controls from October 2022 continue to define the geopolitical background. Washington tightened restrictions on advanced chips and manufacturing equipment to slow China’s leap in AI, supercomputing, and dual-use capabilities. But those measures haven’t halted the mainland industry. Rather, they’ve prompted a reorganization: less dependency on hitting the cutting edge immediately and more focus on mastering capacity, costs, and internal ecosystems where there’s room to maneuver.
What the numbers really say
2025 figures help ground the debate in real terms. SMIC, China’s leading foundry, closed the year with revenues of $9.327 billion—a 16.2% year-over-year increase, an average utilization of 93.5%, and more than 1 million 8-inch equivalent wafers of monthly capacity. Hua Hong, the second largest Chinese foundry, earned $2.402 billion and operated at an average utilization of 106.1%, meaning above its nominal capacity. This isn’t industrial propaganda; it’s real demand pushing factories to their limits.
The following table summarizes the most relevant data:
| Indicator | Number | Interpretation |
|---|---|---|
| SMIC revenues in 2025 | $9.327 billion | Historic high for China’s leading foundry |
| SMIC YoY growth | 16.2% | Expansion continued despite sanctions |
| SMIC average utilization | 93.5% | Capacity remains highly strained |
| SMIC monthly capacity | Over 1 million 8” wafer equivalents | Established industrial scale |
| SMIC R&D expenditure | $774 million | Tech efforts continue to grow |
| Hua Hong revenues in 2025 | $2.402 billion | Second Chinese foundry also growing |
| Hua Hong average utilization | 106.1% | Operating above nominal capacity |
| China’s share in legacy chips | ~30% | China is now a dominant player in mature chips |
| Projected global legacy capacity share in 2027 | 39% | Competitive pressure will grow, not diminish |
Sources: SMIC 2025 Annual Report, Hua Hong Q4 2025 Results, EUISS, Rhodium Group.
That approximate 30% in legacy chips is not a trivial detail. The EU Institute for Security Studies notes that these semiconductors—though less technologically sophisticated—will still account for around three-quarters of global demand and are critical for automotive, medical devices, robotics, defense, and industry. Additionally, the institute highlights that the European Union lags behind with about 13% market share in this segment. Put differently: Europe risks obsessing over advanced chips while becoming dependent on basic ones.
And here’s the more uncomfortable part. When China increases its supply of mature nodes with government support, scale, and favorable internal conditions, it doesn’t need to surpass NVIDIA or TSMC to pressure the market. It’s enough to flood categories where price matters a lot, margins are tighter, and many Western manufacturers lack the same backing. Rhodium Group estimates that China will account for 39% of global legacy chip capacity in 2027 and 55% of the projected capacity expansion in that segment. That’s no longer just catching up; it’s about shaping the rules of the game.
The frontier of cutting-edge technology remains
However, it’s also important not to buy into the Chinese narrative of full self-sufficiency without nuance. The technological gap remains significant. TSMC confirms that its N2 2nm technology entered volume production in Q4 2025. Meanwhile, China has demonstrated the capability to approach 7nm class nodes in certain contexts but remains far from the leading production standards that define leadership in performance, efficiency, and cost per transistor.
That gap is clearer in this second table:
| Area | China | Current Leaders | Interpretation |
|---|---|---|---|
| Strongest segment | Mature, power, industrial chips | Distributed across Asia and the West | China wins by scale, not by cutting-edge |
| Most cited advanced node | Class 7nm | TSMC N2 in volume production | The frontier remains distant |
| China’s share of legacy chips | ~30% | EU ~13% | European dependence could grow |
| Projected legacy capacity in 2027 | 39% global | The rest of the world is below in aggregate | China consolidates volume power |
| AI strategy | Cost-effective, open, practical deployment | More focus on closed frontier models | Two increasingly different competitive models |
Sources: TSMC, EUISS, Rhodium Group, TrendForce.
This calls for a more sober interpretation. China has not yet broken through the barrier of advanced semiconductors. But it doesn’t need to do so tomorrow to influence the global industry. If it gains ground in mature chips and, additionally, pushes AI platforms that are sufficiently good and much cheaper, the impact on competitors could be very serious. TrendForce reported in January that generative models developed by Chinese companies already accounted for about 15% of the global market share in November 2025—up from around 1% the previous year. It’s not market dominance, but it’s rapid growth too significant to ignore.
The real challenge for the West
The biggest weakness in the West isn’t that China will start manufacturing the world’s best chip tomorrow. It’s another: while the U.S. and Europe focus their capital and discourse on the top tier of the market, Beijing could end up dominating much of the industrial base and the “good enough” segment. And in technology, what’s “good enough” and cheap often turns out to be politically very influential.
This pattern has been seen in other sectors. First, the product is dismissed as not being the most cutting-edge. Then, it proves useful, sells well, gains market share, and puts pressure on more expensive competitors’ margins. Mature chips may seem less sexy than an AI GPU, but if they become concentrated in China, the problem for Europe and much of the Western industry will be equally strategic. Perhaps even more so, because it will affect many more sectors beyond frontier model training.
The uncomfortable conclusion: China hasn’t yet surpassed TSMC at the forefront of semiconductors, but it has advanced enough to force rivals to fight two wars simultaneously—the cutting edge and volume. And the latter, despite fewer headlines, might be the one that begins to hurt most first.
Frequently Asked Questions
Can China already produce 2nm chips comparable to TSMC?
No. TSMC confirms that its N2 2nm technology entered volume production in Q4 2025. China remains far from that frontier, although it has advanced in much more mature nodes and certain class 7nm implementations.
Why is there so much concern about legacy or mature chips?
Because these semiconductors support much of the real economy: automotive, industry, medical devices, defense, consumer electronics, and automation. The EUISS notes they will continue to represent around three-quarters of global demand.
Which Chinese foundries are growing fastest?
SMIC and Hua Hong. SMIC made $9.327 billion in 2025, and Hua Hong $2.402 billion, with very high utilization rates in both cases.
Is China also gaining ground in AI, not just chips?
Yes. TrendForce reported that Chinese generative models accounted for about 15% of the global market share in November 2025, driven by players like DeepSeek and Qwen.