China has ambitions to become a powerhouse in semiconductors. All too often, their efforts are misunderstood in the US. This leads to companies underestimating brewing competition, while at the same time it leads the US government to overestimate the policy implications and any perceived national security threats.
Several years ago, Xi Jinping, the General Secretary of the Communist Party of China gave a speech where he pointed out that China spends more on its semiconductor imports than it does on oil imports. The obvious implication being that semiconductors had a become a Strategic Asset for nations. Since then the Chinese government has built a massive apparatus for acquiring a domestic semis industry.
This has led to today’s situation where semiconductors are a major part of the US-China Trade War and a perception that this could be a flashpoint between the two countries. We have written on the Trade War and are working on on an expanded piece, but for this post, we are going to look beyond the current headlines to more closely examine how China is building its semiconductor complex.
There are two sides to the semis industry – manufacture and design. Manufacturing is (incredibly) capital intensive. China has been trying for a long time to develop its chip manufacturing capability including both foundries and the tools used in chip manufacturing. Neither effort has fared well. This has led to Taiwan’s TSMC becoming a major focus point for both sides. Given the technical requirements and the amounts of money involved, it seems unlikely that China will close this gap any time soon.
However, the outlook on the design side is very different. Today, the US is unarguably the world leader in chip design – companies like Intel, Qualcomm, Broadcomm and a few hundred more. Until recently, China has lagged on this front. For many years, China’s electronics sector had been reluctant to invest, preferring the technically less demanding, and much more tangible, investment in electronics manufacture and assembly. Designing a chip is a more challenging engineering task, in that time to return on investment is much longer and less predictable.
That has now changed. The ee Times reports that there are over 1,000 chip design companies in China today. Many of them are small, custom shops doing small jobs, but many far larger. The ee Times says China minted 38 chip billionaires in 2019. Today, none of them could displace an Intel for CPUs or, with one exception, Qualcomm for wireless modems But this is an active, highly competitive market, and smart companies are building niches which they can then grow into major businesses.
One of the major impetus for this has been major government support for venture investing in chip companies. However, it is important to understand the form of this support is largely done on commercial terms. We know that many people in the US assume government support comes in the form of inefficient state-owned companies building products by committee, but this is completely off the mark. Aside from a few companies developing chips very narrowly targeted at military uses, almost all of those 1,000+ chip companies are run professionally and allowed to compete freely.
The irony is that just as the US is looking at ways to provide direct government subsidies for specific chip companies, China is using its own turbocharged capitalism to build a next generation of chip companies.
In the past we have noted that there is effectively an infinite amount of money available for chip companies in China. Tapping into this capital does not require filling out endless forms and submitting an application to the Committee on Chip Investing. Instead, it is just a matter of raising venture capital – pitch decks and coffee meetings – that would feel familiar to anyone working in Silicon Valley. It is true that these investors are backed by State sources, but the various State backers largely function as would an LP in a US venture fund. There are exceptions, but to a surprising degree these investors make decisions largely independent of the government.
There are about 40 or 50 investment firms doing major chip investments in China. For reasons we will not get into, they mostly call themselves Private Equity firms, not Venture firms, but the difference is largely semantic.
They tend to be staffed with principals who have impressive resumes working for US technology firms. These are people who understand chips and have experience in running large chip operations. Again, defying the stereotype, personal relationships and 关系 carry no more weight here than they do on Sand Hill Road. Which is to say they help, but they cannot carry an investment. These investors ask tough chip questions and know how to judge a product, a market and a team. We have met with many of these firms over the years, and we recognize kindred spirits, with matching scars.
They expect venture-type returns. Waving hands and claiming National Interest and Patriotism carries little weight. The end result is a growing field of Darwinaly-honed competitors.
China’s chip industry still has a long way to go. As would be expected of any country in that stage of development, many companies still focus on low-cost alternatives. But if anything, China has demonstrated that its companies and industries can learn how to move up the value chain.
Take the use of PCs as an example. China’s government is acutely aware of its own reliance on PCs running on Intel CPUs. There are probably no companies in China today that could entirely replace those x86 CPUs. However there is also no reason to assume that every government employee or border inspector needs a PC. Everything can run on the cloud, so why not replace that PC with a lower-powered Chromebook-like device or a tablet running the required cloud interface. These could make do with some homegrown ARM-based processor just as well. (Let’s save a discussion for ARM in China for another day). This will not happen tomorrow, but the point is to not assume today’s framework will be valid in a few years.
We know that people will view this as some form of threat to the US, but this overstates the case. China’s companies are operating under their own conditions and will optimize their strategies to suit their individual economics. This is not some nefarious, underhanded plot so much as a rational development outcome.
Where the US should be concerned is that while China has many firms investing in chips, the US has maybe three. Our rough math is that about 50% of Chinese venture (including PE) firms will invest in chips. In the US that figure is closer to 0.3%. That should be of far greater concern to the US government and the US Technology Industry.
Pingback: The State of RF Semis in 2020 | DIGITS to DOLLARS·
Pingback: Intel Follow-Up – Data Centers | DIGITS to DOLLARS·
Pingback: Tears in the Rain – Arm and China JVs | DIGITS to DOLLARS·
Pingback: The Five Year Plan for Semiconductors | DIGITS to DOLLARS·
Pingback: So what is the right model? | DIGITS to DOLLARS·
Pingback: Arm for sale? | DIGITS to DOLLARS·
Pingback: China RF Semis | DIGITS to DOLLARS·
Pingback: What will happen to Huawei? | DIGITS to DOLLARS·
Pingback: Who benefits from Huawei’s predicament? | DIGITS to DOLLARS·
Pingback: Why settle for 1,000 chip companies when you can have 10,000? | DIGITS to DOLLARS·
Pingback: Should Intel Split in Two? | DIGITS to DOLLARS·
Pingback: Cognitive Dissonance and Semis Funding | DIGITS to DOLLARS·
Pingback: Tsinghua Unigroup Defaults Widen | DIGITS to DOLLARS·
Pingback: Corporate Shells are 9/10s of the Law | Digits to Dollars·
Pingback: High Level China Semis Overview | Digits to Dollars·