D2D China: The Economy Rears Its Head
Back alley GPU deals, China WFE and Nuclear Tidal Waves
We are publishing this for all readers as a sample of our work. But Digits to Dollars is supported by subscribers. Please sign up below to receive the full content of this note and many other benefits. Thank you for reading.
Noteworthy Items
Semis and Deep Tech
Spacemit is designing a RISC V CPU and they just raised several hundred million RMB in their Series A fundraising round. There are a at least half dozen companies in China trying to build this kind of product, all raising eye-popping amounts. On the one hand it is great to see semis companies with big ambitions, raising big amounts and coming up with next-gen names. On the other hand, it is hard to see how the market can support even a fraction of them. Also, it would be great if US start-ups could pull off these kinds of funding rounds.
Memsonic, a company which seemingly no one had heard of before, is reportedly ready to ship BAW (fBar) filters to Huawei. We have been waiting for this kind of surprise for many years. Very little is known about the company and their products, let alone their yields and quality, but if this is true it is significant for the mobile RF industry.
Electric vehicle start-up Nio plans to launch their internally designed driving assistance chip in 1-2 years. On the other hand, handset maker Meizu, owned by car marker Geely, is scaling back their internal chip ambitions. One of the benefits of China’s frenetic pace of activity (and funding) is that it allows companies to move very quickly. We would not be surprised if more companies dip in and out and then back into the self-design market in coming years.
Digitimes makes the case that as much as we are in a boom for AI chips and a boom specifically for AI chips in China, the market is too fragmented to promote healthy, sustainable growth.
Apparently, a third of RISC V cores were manufactured in China. RISC V has a lot of momentum right now, and China is a big part of that.
Auto IC maker Will Semi (韦尔股份) reported some dismal earnings and saw their stock down sharply. There are many who think Will is just the first causality of a coming blood bath in China’s Analog Semis stocks. As one market observed noted Profitability is the only hope.
Wuxi, a city in East China, announced a new set of policies to attract semis companies to town, seemingly with a focus on EDA companies. Worth reading to get a sense of the long list of incentives that local governments offer companies. Chinese government support of semis comes in many flavors beyond direct subsidies.
With the downturn in smartphones globally this year, TSMC was left with a lot of capacity at 7nm. This is being taken up by Chinese AI companies.
Buying Nvidia GPUs on the street corner with strangers met on WeChat. The GPU Gold Rush is in full effect.
Automotive, Industrial and Macro-Economics
European think tank Merics published a deep study of China’s approach to industrial policy. Their central thesis is that the government is opening up state support to private companies in designated sectors, recognizing China’s need for the private sector to drive technical innovation. The ways in which this differs from past policies are more nuanced than we can get into, but they are significant, if they are allowed to fully play out. Worth noting that the favored sectors are medical and pharmaceutical, robotics, materials (i.e. semis manufacture) and energy (i.e. EVs). These matter as China has a lot of success in some of these areas already (which is how the program is structured), and many of these companies could emerge down the path as global players.
China’s government is making headlines by pulling ever more data from the public domain. Many are now familiar with the disappearance of youth unemployment data this month, but this is happening across China government sources. A big concern is the large number of court records, especially around Intellectual Property (IP) cases which do not appear anywhere in official records. As much as China wants to portray itself as a rising power, these moves belie a deep lack of confidence.
This blogger claims that China policymakers’ goal is for the country to produce 80 million EVs a year by the 2030’s. This is both outlandish (the whole world made 85 million vehicles in 2022), and alarming because maybe they can pull it off.
40% of all funds raised in Chinese IPOs this year went to semiconductor companies. But don’t call it a Bubble….
A good overview of China’s Wafer Fabrication Equipment (WFE) companies. This sector has disappointed in its inability to catch up to Western peers, but they continue to chip away at the market.
Software and the Cloud
The Information got a hold of Bytedance’s financials. They are big, but growth in China is slowing, and they are becoming more dependent on foreign markets where they face growing regulatory and competitive attention.
If you like this content you should listen to our podcast. In the latest episode we look at the reality behind “de-globalization” and the ongoing connectedness of the semis supply chain.
Tidal Wave of Capacity
In the 1996 not-quite-cult-class film Escape from LA Kurt Russell (spoiler alert) triggers a nuclear explosion off the coast of Los Angeles and then surfs the ensuing tidal wave to accomplish the title of the movie. This is the defining scene of a highly dubious film. We have found ourselves thinking about this scene a lot lately as we read the latest macro-economic news from China.
The headlines paint a dark picture. Youth unemployment climbing to 20% before the PRC government just stopped reporting the data entirely. Steady declines in housing prices in all markets. Looming defaults of major property developers triggering defaults in local government trusts threatening to trigger a banking crisis. The list goes on.
We are not macro-economists, but we have been watching China for long enough to know a few things about the current situation.
First, the outlook is nowhere near as bleak as the Western press makes out. US press coverage of China tends to vacillate wildly between the extremes from “China is Taking Over the World” to “China is Doomed”. The truth is always more nuanced, maybe some place in between the extremes but maybe on some completely separate spectrum. China still has several trillion dollars in foreign exchange reserves and a largely closed capital system, so we can rule out using the word “collapse”.
That being said, China clearly faces some significant challenges. We tend to view the situation through the framework established by an actual macro-economist, Michael Pettis of the Carnegie Endowment. By this framework, China appears to be reaching the end of the economic model it has employed since the 2008 Financial Crisis - over reliance on fixed asset investment and exports, built on the back of financial transfers from the household sector.
And this leads to our second observation. Absent a major shift in China’s political economy, companies in China are likely to resort to the prior playbook. We do not know how that will work in the property sector (aka 25% of China’s GDP), but we are pretty clear what it will mean for the Tech sector, especially semiconductors and electronics. And we can sum that up in one word - Capacity.
China’s economic policy managers’ reflex seems to hold that when times are bad build housing developments and factories. In semiconductors, this means more fabs. China already has ~200 semis fabs with allegedly another 100 under construction. These are all being built to produce trailing edge processes. At the same time, existing plants are already heavily under-utilized. This year we have seen reports of some (old) fabs running at 30% utilization, with foundries offering 30%-50% price cuts. Meanwhile, the global market for these products has finally reached equilibrium after years of constraints, meaning it is poised for a downturn in its own cycle.
This extends beyond fabs. For much of this decade, the pace of new semis company formation was frenetic, with a new fundraising announced almost every day. This ground to halt late last year when the arrests of the executives at China’s Big IC fund signaled a change in government policy towards the sector. But in recent weeks, the pace of fundraising has picked up again sharply. Semis are in favor again, or at least they are seen to be that way by investors. As the post we linked to above mentions, 40% of IPO funds raised in China’s stock markets last quarter went to semiconductor companies. Local and provincial governments have unleashed new rounds of direct and indirect subsidies for companies in favored sectors notably Wafer Fabrication Equipment (WFE), EDA tools and anything touching AI.
And of course there are electric vehicles (EV) OEMs who seem to be expanding as fast as they can. EVs provide a good case study in why all of this has to come to end at some point. China has almost overnight gone from being a net importer of cars to one of the world’s largest exporters. With trade tensions already heightened, it seems impossible that China will be allowed to become a dominant player of the industrial world’s central pillar.
Nonetheless, for the foreseeable future it seems inevitable that PRC companies will be exporting large amounts of analog semis, EVs and a host of other electronics and industrial products.
And to be clear, this is not just a problem for the rest of the world. Take any segment of semiconductors and there far too many competitors ruining profitability for everyone. Many in China’s semis sector are calling for consolidation, but everything seems to be going in the other direction. If history is any guide, we will see consolidation across China semis some day, but that day does not seem to be coming soon.
It is unclear how this will ultimately play out. We have heard scenarios that run the range from apocalyptic - war over Taiwan - to the hopelessly optimistic - China will alter course and shift its economy to something more consumption driven.
So far signs are not good. The Central Bank last week lowered interest rates, a page taken from the old playbook. And the Chinese press is not optimistic about any of this. For example, this analyst bemoans the many problems facing problems facing China’s economy, but is very quick to point to problems in foreign economies, largely eliding the root cause of the problems.
Maybe China’s policy makers will alter course and devise a clever way to shift the economy, but it is hard to see how they could accomplish that in anything less than a decade. In the meantime, for anyone competing with anything made or designed in China, keep an eye to the horizon for Kurt Russell on a giant wave of excess capacity.