The China Integrated Circuit Design Industry Exhibition (ICCAD) is China’s annual semiconductor design focused show. Exhibitors are mostly semiconductor IP and electronic design automation (EDA) tool suppliers, fabless design firms, foundries, and packaging companies. It’s the perfect place to get a feel for how China’s industry has developed over the year and where it is going. I went there last week to get a sense of things on the ground, the mood, and see how the industry is progressing.
China’s struggles
In the last year, the number of Chinese semiconductor design companies has increased by 175 companies (5%) from 3,451 to 3,626, its lowest growth in the past five years. Industry revenue grew at 11.9% though to around $90.9bn, which averages out at around $25m per company, an increase of about $23m from the previous year. This is a positive sign, but we have to remember that only 731 companies actually have revenues above RMB 100m ($13.7m), meaning while looking at averages may be useful to provide a general trend, in reality, most companies have revenues below $13.7m.What’s more, only 32 companies have more than 1,000 employees, two fewer than last year; in fact, 87.9% (or 3,187 companies) have fewer than 100 employees.
Most of these companies are designing telecommunications or consumer electronics chips, mostly in the mid to low end, which in my opinion, looking around ICCAD, is still where most of China’s industry is. There were many similar IP companies, for example, low-end RISC-V cores, or interface IP like SerDes, with no clear difference between any of the firms.Speaking to salespeople it seems these companies basically compete on price and relationships, rather than any specific USP.
Last year, China’s industry growth was only 8%, 0.2% less than the global industry, and this trend has continued this year, with its 11.9% being lower than the globally predicted 19% for 2024. If this happens again next year, then it will become a clear trend.
What’s more, despite the aforementioned economic issues, employees can still demand world-leading salaries. Many companies struggle to afford these wages while employees continue to move from company to company, taking advantage of the situation as firms with new funding rounds use the influx of capital to bring in the best talent. This is a vicious circle that still continues, and it is hard for companies to escape from if they want to have the best talent.
On top of this, customers continue to demand lower prices. For example, recently BYD was in the news demanding suppliers reduce prices by 10%. This kind of mentality makes it hard for smaller chip design companies to survive and traps companies in this low-to-mid end market, as doing anything more will push costs onto their customers. This is forcing many of the companies I spoke with to look abroad for customers as they believe they can demand higher prices and thus make acceptable profits.
Some positives
There are some positives to take out of all this for China. I did notice that Chinese design companies seem to be working closer with their foundry partners, helping each other improve, rather than being transactional, which was definitely the case previously.
It is also clear that TSMC, Samsung, and GlobalFoundries are still doing their best to supply Chinese customers with what they are allowed to. All companies had large booths and even pitched their latest process nodes. Perhaps some Chinese companies can still gain access. What’s more, they were not the only foreign companies present; most large IP and EDA companies had a significant presence, so it seems the Chinese market is still profitable and attractive in this sense.
All the pressures, domestic and foreign, for Chinese companies are also opportunities. They used to rely on US EDA tool suppliers’ innovations or foundry innovations to improve their chip PPA (power, performance, and area), but now they are forced to find innovative architectures or subsystem designs to achieve results – something they might not do otherwise.
Conclusions
From a macroeconomic level, the Chinese industry is flat, and even falling behind the global industry. China needs to accept it is cut off from certain technologies, focus on mature node development, and innovate its way out of the situation at the bleeding edge, be that with innovative architecture designs or new materials. Prof. Wei Shaojun said it best at the show, and I paraphrase here: “Chinese companies should stop complaining they don’t have access to TSMC. Even if they did have access, do they really think they are innovative enough to create a world-leading GPU?”. Stop complaining and get on with it – very practical words. But it remains to be seen how easy getting on with it will be. It will be interesting to see whether the Chinese industry can outgrow the global industry in 2025. If it can’t, then a new trend is set.
