The winds have changed for Chinese entrepreneurs. Xi Jinping is driving a crackdown on traditional software sectors like fintech, ecommerce, social media, and gaming. Policy is shifting incentives toward materials science, AI, biotech, and the like. The effects are everywhere: juggernauts like Alibaba and Baidu’s stock have fallen 23% in the past weeks, and venture capital appetite for Chinese consumer SAAS has pulled back severely. Jack Ma disappeared after a speech critical of the government, and numerous CEOs resigned, most notably Zhang Yiming of $400B Bytedance. Meanwhile, investment in hard tech is surging.

This breaks a 30-year trend of China aggressively growing its software industry. The Great Firewall gave Chinese startups a safe space to innovate and compete against one another without being disrupted by further-along Western players. From that Petri dish, extraordinary companies and products have emerged: today, US innovators look to WeChat and Pinduoduo for inspiration. Especially with the rise of TikTok, it seemed like we were on the cusp of Chinese consumer software taking over, and Silicon Valley moving to second place.

Xi has put the brakes on the global ascent of Chinese consumer SAAS. It’s a costly move for China, and raises an important question: is he right? Is it true that (1) consumer SAAS is less valuable than manufacturing and (2) the two are mutually exclusive to some extent?1

The consensus in the West is that (1) we are moving from a manufacturing economy to a service economy and (2) software consumes manufacturing. The second point is subtle: companies that rely on any kind of iterative development become software companies. Amazon is not a retailer, but a software company. SpaceX is a software company insofar as their manufacturing is enabled by breakthroughs in computationally modeling rocket launches. Tesla is a software company, whose manufacturing is similarly enabled by computational modeling, and much of the customer experience is defined by the touchscreen. In general, software eats manufacturing since (1) any manufacturing/R&D process gets digitized and (2) any manufactured good will have a software component.

Importantly, software cross-pollinates: for example, Facebook’s public research in computer vision gets adopted by hardware companies all the time. Great work has come out of ecommerce companies that is now used in life sciences, construction, and so on. Not to mention that founders in biotech wouldn’t necessarily start companies in fintech, or vice-versa. In short, I believe the two are more mutually complementary than exclusive.

Whether consumer SAAS is less valuable than manufacturing is hard to answer, because it’s about hard-to-measure externalities. By the numbers, consumer SAAS is far more valuable. Google prints money, steel mills don’t. However, highly financialized software carries a whiff of bullshit. If I give you $20 and you give me $20, GDP is up by $40 but we’re both no better off. The entire advertising industry seems like a Red Queen problem, where despite enormous competition and spend, there’s no benefit to either buyer or seller. You might argue that advertising and fintech companies create liquidity in markets and lubricate the economy to facilitate more transactions, but it’s not obvious to what extent increasing the velocity of money is a social good. We do know that things like advertising and social media come at significant social and cognitive costs, but those aren’t priced in at all. Facebook makes money, but how great are the intangible costs that it imposes on our society?

It’s a challenging question. If you’re Xi Jinping, you may pull out Occam’s razor and say whatever, if I can’t figure out whether this is good or bad, I’ll just focus on the thing that I know to be good for sure. Hence, the push into life sciences, aviation, etc. that produce tangible, non-bullshit things with uncontroversial benefits and more readily understood externalities.

China’s crackdown shows a belief about the value of certain kinds of infrastructure. Digital financial infrastructure is falling out of favor, and physical infrastructure is the big priority. That certain types of infrastructure could be undesirable (or deemed threatening) is a surprise found only in authoritarian countries. In the US, we are generally in favor of all kinds of infrastructure. We’re falling behind on physical, but we’re all-in on digital: new generations of payments, advertising, business creation, and so on. We are doing everything we can to drive the digital economy to hyperefficiency and put the velocity of money on hockeystick growth. In the West, we have a deep belief in the power of the markets to solve allocation and coordination problems, and our tech industry reflects that fact – collectively, it is working to make the gears of the markets turn as quickly and efficiently as possible.

The fundamental question remains whether Xi made the right choice – and conversely, whether we in the West are making the right choice. It’s hard to know, because the externalities of different types of technology are hard to figure out in real time. As for me, I believe the history of innovation shows that the next big thing always comes in guise of a toy or something first perceived as useless. Technology evolves in unexpected ways. As such, I think it is dangerous and possibly counterproductive to rein in toying around. But theory won’t suffice for settling this debate. Thankfully, we now have a practical experiment, with different approaches in different nations: time will tell.



  1. This is an important question because – if it were true – it would mean that “less important” companies were being run at the expense/opportunity cost of “more important” companies. In a world in which an undesirable social media company competes over the same talent as a desirable biotech company, one might enact policy to curb social media companies.