I reported on China’s chip design progress last year and figured it made sense to make this an annual occurrence. I was present at the China Integrated Circuit Design Industry Exhibition (ICCAD) once again this year on behalf of our clients. I’ve been attending the show for the best part of a decade now.
China’s struggles
Last year, the number of Chinese semiconductor design companies rose 5% from 3,451 to 3,626; and this year it rose again, to 3901 companies, a 7.6% increase. Overall revenue from these companies also increased, now at approximately $118bn, a 29.8% increase on last year. This means on average a Chinese semiconductor design company is earning $30.3m in revenue, up from $25m last year. This is positive, but the same problem prevails. There are just too many small companies.
Of the 3901 companies, 831 have revenues above RMB 100m ($14.1m). That means that despite the high average, it is still the case that 3,070 companies are struggling to survive, with revenue less than $14.1m. In fact, when we break it down, 2061 companies earn less than $1.4m, and 783 companies earn less than $7m. Only 39 companies have more than 1000 employees, and 3375 have less than 100 employees, or 86.5%, no change from last year. In short, most companies are small, weak, and fragmented.
There has been no significant change in the types of chips being designed, either. Telecommunications and consumer electronics chips in the mid-to-low end make up two-thirds of all revenue. Processors make up approximately 7.7%, much less than the global average of 25%.
Due to market interest in AI and automotive semiconductors, there has been little change in employee wage demands; many employees can still demand world-leading salaries, yet many companies struggle to afford these wages, while employees can easily move from company to company, as companies compete to attract the best talent. This is a vicious circle that has continued for the past five years now.
Involution is still a big problem, too. Chinese customers continue to demand low prices, and foreign customers – where higher margins can be gained – are increasingly looking for non-Chinese suppliers.
Positives
We can’t be overly negative, though; in fact, the positives for Chinese companies this year are starting to outweigh the negatives. The industry saw its largest growth ever at 29.8%, despite geopolitical pressures. Talent is sufficient for mid-low experience workers for now, and China predicts its industry to reach revenue of RMB 1tn ($141bn) by 2030. Although I do doubt this personally, as this would mean China’s IC design industry would account for approximately 14% of the entire global semiconductor industry. We shall see.
For now, at least, it seems many Chinese design companies still have access to US EDA tools and Taiwanese fabs. TSMC, Samsung, Synopsys, Cadence, Siemens, Arm, etc., all exhibited, and most companies I spoke with designing AI chips or GPUs admitted to taping out at TSMC or using US EDA tools.
One specific positive area is China’s development of RISC-V server chips. No other country seems to be developing such products, at least not on this scale. Many companies, such as Spacemit or Sophgo, have moved from designing RISC-V IP to RISC-V server chips and now sell accelerator cards and full server products based on their own chip designs. This progress in part relates to restrictions on ARM IP and geopolitics, but is a good example of restrictions leading to innovation.
Conclusions
Last year, I concluded from a macroeconomic perspective that the Chinese industry is flat and even falling behind the global industry as it struggled for growth. This year has been much different, experiencing its largest growth ever. Rather than accepting using SMIC or more mature process nodes, many Chinese IC design companies are taking the risk and still going with TSMC or Samsung to have high-performance products.
China’s old problems persist, though. Once again, the number of companies grew, further fragmenting the industry. This needs to stop; consolidation is much needed if China is ever going to have companies that can compete globally. It is one thing to have a high-performing chip, but it is meaningless if you cannot produce enough for your customers. Most of China’s IC design revenue still comes from mid-to-low-end consumer electronics and telecoms chips. There has been no change here; despite the growth in AI and automotive chip companies, they have had little effect on the industry from a macro perspective. Chinese companies need to stop competing on price and focus on product quality. This is easier said than done, though, when its domestic customers are struggling, and foreign customers are often looking for non-Chinese alternatives.
Overall, the year has been positive, though. The Chinese industry grew faster than elsewhere, unlike last year, and a new trend of decline has not occurred. Let’s see what 2026 will bring!
