In 2021, US venture investors funded 99 deals for semiconductor companies, in China the number was 364. That ratio was the same in 2020. If you dig into the numbers, it gets more stark. If we just look at early stage deals, the US saw 33 deals, China 186. We culled these numbers from Pitchbook, which is a great platform but has real gaps in its coverage for China, so the disparity is probably even greater. By our (very rough) estimate, Chinese VCs have been funding close to one semi company a day in 2022. As much as the press and the policy people talk about renewing investments in US semis, the reality on the ground is totally different.
We suspect that if we extend that analysis further afield to include other advanced electronics that the disparity holds as well. Over the years, we have covered this topic on occasion, but given the changing state of the world, we think this under-investment in “Deep Tech” or “Hard Tech” has become an urgent problem. And we do not need to cast this in purely patriotic terms. Michael Pettis, who has some of the most prescient China economic analysis out there, pointed to this paper which shows US policies detract not only from US innovation but global technology improvements.
Some may argue that US investments reflect its ability to drive growth through higher-value areas, and the US certainly has some of the most valuable technology companies in the world. But they all rest on top of semiconductors and Deep Tech. Apple is an obvious success, a showcase of some of America’s most advanced capabilities – totally dependent on chips. What powers Google, Amazon’s and Facebook’s immense expertise – semiconductors. Look at any area of the economy – agriculture, manufacturing, logistics, retail – all things the US does incredibly well – and the story there is about the growing use of electronics to drive productivity gains. And don’t get us started on chips for automotive (and here and here).
Others will argue that the economics of venture investing in semis just do not work out. Why invest in a chip company which will need $200 million to get to breakeven when they can invest in two developers in a garage who will only need $5 million to get to breakeven, and maybe $20 million to get to runaway growth? This may have held true 10 years ago, but conditions have changed dramatically. VC valuations for hot consumer tech companies are through the roof. At the same time, the cost of building a chip have fallen substantially. We know many chip companies in the US who designed chips for less than $5 million in angel funding, and only needed to take institutional venture money to take the chips to production. Unfortunately, we know many more aspiring chip companies who could not make that transition to institutional fund raising. There are roughly 1,000 venture firms in the US, but maybe 20 of them will even look at semis deals.
With semis so important to all of technology, and to the entire economy, but funding barely enough to qualify as anemic, there is a huge opportunity for technology-savvy investors to step in. Is there another deca-unicorn waiting to be born in social media? There are definitely dozens of incredibly high-potential semis companies out there poised for success, and poised to deliver incredible returns to true Deep Tech investors willing to take the plunge.
Hi – what do you mean that the costs of designing a chip have fallen? Mask costs at TSMC at 5nm are very high, maybe $50 million. I thought bringing a leading edge chip to market today might cost $200 million or more. Thanks
Building, manufacturing a chip – yes, those costs have gone up a lot. But the cost of designing a chip has fallen considerably. There is a lot more, highly capable talent out there who can do the work. The EDA tools have made their pricing more tolerable for start-ups (not cheap, but tolerable), and the cost of some of the IP has also gone down (looking at you RISC V) – those kinds of factors make it much less expensive to design the chip.