All podcasts / Invest Like the Best / Summary

Dan Wang - The US vs China In The 21st Century - [Invest Like the Best, EP.444]

2025-10-16 - 94 min - source - Read full transcript

Patrick O'Shaughnessy (host)Dan Wang

Key insights

It is structurally harder for the US to rebuild manufacturing scale-up capability than for China to close the gap in scientific research.

Wang argues the US still leads at 'zero to one' invention but China has already gotten significantly better at frontier science (rising share of top 1% cited papers, a plausible lunar landing before the US returns), while US apex manufacturers like Boeing, Intel, and Detroit automakers have kept declining and the US builds almost no iPhones or desktop computers domestically.

manufacturing-innovation-asymmetry

China's technology strategy is to let others invent and then scale manufacturing faster than anyone else, using the 'spark and prairie fire' pattern.

Solar photovoltaics were invented at Bell Labs in New Jersey in 1954 but treated as a scientific curiosity in the US; China later industrialized the technology and now controls roughly 90% of the global solar supply chain, from polysilicon processing to final module assembly.

manufacturing-innovation-asymmetry

China's enormous manufacturing workforce compounds tacit, hard-to-encode process knowledge that shifts entire industries over time.

Wang cites roughly 70 million Chinese manufacturing workers versus about 12 million in the US; a worker who assembled iPhones in Shenzhen in 2008 could plausibly move into Huawei handset production, then drones, then EV batteries, explaining why Shenzhen became a hub for hardware, drones, and battery manufacturing in succession.

manufacturing-innovation-asymmetry

Nobel Prize counts are a poor, backward-looking proxy for judging China's current scientific capability.

Prizes are awarded for work often done decades earlier (frequently in the US or Japan in the 1970s), so the small historical count of Chinese laureates undersells present-day Chinese research quality, which Wang says is now visible in a rising share of the world's most-cited scientific papers.

manufacturing-innovation-asymmetry

ByteDance and Alibaba trade at steep valuation discounts because of three compounding, largely uninsurable risks rather than weak fundamentals.

Wang identifies a structurally flat Chinese stock market despite roughly 9% average GDP growth over 20 years, unpredictable Communist Party intervention (ByteDance's 2018 forced self-criticism letter, the 2020 blocking of Ant Group's IPO after Jack Ma's remarks), and geopolitical/Taiwan sanction risk compounded by holding shares through offshore VIE structures rather than direct ownership.

china-investing-risk

China exhibits higher social 'agency,' defined by Francis Fukuyama's notion of trust as the ability to coordinate and act quickly, than the US.

Early in the COVID pandemic, Chinese manufacturers like Foxconn and JD.com rapidly retooled factory lines to produce masks and cotton swabs while, in Wang's account, American manufacturers asked whether making such goods was part of their 'core competence' and largely declined to.

state-capacity-and-governance

China's 'engineering state' can execute rapidly but the same mindset that treats society as another building material has produced some of its worst policy disasters.

Wang traces the pattern from the Great Wall and Grand Canal through the modern Politburo (historically dominated by engineers by training) to the Three Gorges Dam's displacement of roughly a million people, repression in Xinjiang and Tibet, and especially the one-child policy, which he describes as a campaign of rural terror involving roughly 300 million abortions and 100 million sterilizations, alongside the zero-COVID lockdowns.

state-capacity-and-governance

America's 'lawyerly' system protects pluralism and legitimate businesses but also empowers lawsuits and local objections that slow public infrastructure to a crawl.

Wang cites the Amtrak Acela's proposed Connecticut route being made 'squiggly' and delayed after homeowners sued to keep the line out of their backyards, and notes the Metro-North train from New Haven to New York was faster in a 1914 timetable than it is today; he wants the US to become roughly 20% more like an engineering state and China to become roughly 50% more lawyerly.

state-capacity-and-governance

China studies the Soviet Union's collapse and Japan's stagnation specifically to avoid repeating their mistakes, which is part of why Wang doesn't expect a Soviet-style implosion.

He argues Gorbachev's error was attempting economic and political reform simultaneously, which China's leadership has consciously avoided; and unlike Japan's closed, nearly all-domestic-value-add export model, China stayed open to foreign manufacturers (building Apple's iPhones and Tesla's vehicles), which helped it absorb technology and capital Japan's system never did.

state-capacity-and-governance

AI competition over the next five years is a genuine toss-up rather than a settled US lead, largely because of a looming power-generation gap.

The US currently leads on compute and Nvidia chip access, but China is expected to add roughly 500 gigawatts of solar capacity this year versus about 50 gigawatts in the US, has 33 nuclear plants under construction versus zero in the US, and Wang flags the possibility of a reverse brain drain of Chinese-origin AI researchers (several of Meta's publicly disclosed superintelligence-lab hires attended Chinese universities) back toward Chinese labs.

ai-race-and-talent

Vertical integration and manufacturing speed let a much smaller Chinese company beat a vastly larger US rival to market on hardware.

Xiaomi, worth a small fraction of Apple's roughly $3-4 trillion valuation, announced a $10 billion car program in 2021 and shipped an electric SUV that set a lap record at Germany's Nurburgring, while Apple spent about a decade on Project Titan before cancelling its car effort entirely.

manufacturing-innovation-asymmetry

The most likely flashpoints for a US-China conflict are not a full invasion scenario but unpredictable regional or maritime incidents.

Wang sees Taiwan as the obvious long-term risk but currently unlikely because the status quo works for all three parties; he views a repeat of the 2020 fatal India-China Himalayan border clash, or an escalation from Philippines-China ship-ramming incidents in the South China Sea, as more plausible near-term triggers than either a Chinese invasion of Taiwan or a US invasion of the mainland.

china-investing-risk

Books referenced

Breakneck - Dan Wang - Wang's new book and the subject of the episode; frames China as an 'engineering state' and the US as a 'lawyerly state.'
The Hundred-Year Marathon - Michael Pillsbury - Patrick raises it to ask whether China's long-term rise has been deliberately kept below the radar so compounding growth goes unnoticed.
Seeing Like a State - James C. Scott - Cited (via a recommendation from Wang's friend Eugene Wei) on whether top-down systems are more fragile long-term than bottom-up ones, despite being faster short-term.
The Art of Not Being Governed - James C. Scott - Wang recommends it directly; about highland Southeast Asia (including his own family's home region, Yunnan) where people historically fled state conscription and taxation.
The House of Huawei - Eva Dou - Wang's friend and former Wall Street Journal (now Washington Post) reporter's book on Huawei, published earlier in 2025; Wang endorses it (transcribed in the source as 'the cult house of Huawei').
Middlemarch - George Eliot - Wang mentions it as the novel he plans to finally read now that the book project is done.

Media referenced

New York Times op-ed by Stephen Greenblatt on Chinese university rankings - article - Cited by Wang as a recent example of American commentators invoking a 'Sputnik moment' for China's rising science output, part of a pattern he thinks is more rhetoric than action.
Andy Grove's Bloomberg Businessweek essay on US job creation - article - Published around 2010; Wang cites Grove's argument that the US is great at setting technological 'ladders' but has stopped being the one to climb them at scale.

Companies

ByteDance - Central valuation puzzle of the episode: throws off enormous free cash flow yet trades at a steep discount because of Communist Party unpredictability; humiliated in 2018 when its founder wrote a self-criticism letter after an app was accused of violating 'core socialist values.'
Alibaba - Cited alongside ByteDance as a cash-generative company trading cheap; founder Jack Ma's 2020 public criticism of regulators triggered a broader crackdown on Chinese tech.
Ant Group (Ant Financial) - Its IPO was blocked by the Communist Party in 2020 after Jack Ma's remarks and remains unlisted.
Meituan - Wang's example of a sprawling 'octopus' conglomerate with roughly 50 core business lines; survived a 'battle royale' against 5,000 Groupon-clone competitors around 2015.
Huawei - Discussed as the archetype of Chinese vertical integration, from switchboards to 5G equipment to handsets; alleged close ties to the People's Liberation Army.
Xiaomi - Handset and household-appliance maker worth a small fraction of Apple's valuation that nonetheless shipped an electric SUV, which set a lap record at Germany's Nurburgring, faster than Apple's decade-long, ultimately cancelled Project Titan car effort.
Foxconn - Taiwanese iPhone assembler; retooled its Chinese factory lines to produce masks early in the COVID pandemic when American manufacturers hesitated.
JD.com - Chinese e-commerce retailer that also retooled operations to produce masks and cotton swabs during early COVID shortages.
DeepSeek - Cited as evidence Chinese AI talent can produce world-class reasoning models; the state reportedly restricts its researchers from speaking with foreign journalists.
Meta - Its publicly disclosed superintelligence-lab hires include several engineers of Chinese origin, raising Wang's concern about a possible reverse brain drain back to China.
Apple - Most iPhones and desktops are still built in China; its decade-long Project Titan car program was cancelled, contrasted with Xiaomi's fast EV launch.
Nvidia - Referenced as a roughly $4 trillion company that needs substantial legal protection to operate, illustrating the productive side of America's 'lawyerly' system; also central to the US-China chip and export-control dynamic.
TSMC - Its Arizona fab is cited as one of the few genuine recent US manufacturing successes.
BYD - Named as a company Wang would like to study in more depth alongside Huawei, Xiaomi, and Alibaba.

Techniques and frameworks

Engineering state vs. lawyerly state - Wang's central framework: China's leadership (mostly engineers) treats the economy, physical environment, and even society as engineering problems to be solved by building; the US is run by lawyers who protect pluralism and debate but also enable lawsuits and NIMBYism that slow infrastructure.
Spark and prairie fire - Wang's metaphor for China's innovation strategy: it lets American, German, and Japanese labs invent the 'spark' (e.g., solar photovoltaics at Bell Labs in 1954) and then 'sets the prairie fire' by scaling manufacturing (China now controls roughly 90% of global solar production).
996 work culture - Shorthand (9 a.m. to 9 p.m., six days a week) for the intensity of work hours common among Chinese engineers and entrepreneurs, contrasted with lighter US tech-office norms.
The anaconda and the chandelier - Sinologist Perry Link's metaphor, cited by Wang, for Chinese censorship: a sleeping anaconda hangs above every conversation, rarely striking but always inducing self-censorship.
Made in China 2025 - China's major industrial plan, mentioned as something Wang first learned about from his mentor Arthur Kroeber shortly before moving to Hong Kong for GavekalDragonomics.

Summary

Patrick O'Shaughnessy talks with Dan Wang, a technology analyst and author who spent six years in China (as an analyst at GavekalDragonomics and later a research fellow at Yale Law School's Paul Tsai China Center and at the Hoover Institution), about his new book Breakneck. Wang's central framework casts China as an "engineering state," led by officials trained as engineers who treat the physical environment, the economy, and society itself as projects to be built and optimized, and the US as a "lawyerly state," where pluralism and legal protection are guaranteed but the same legal machinery that protects the rich also empowers lawsuits and local objections that grind infrastructure projects to a halt. He traces China's engineering mindset from imperial-era megaprojects like the Great Wall and Grand Canal through the modern Politburo, and argues the mindset explains both China's genuine successes (COVID-era factory retooling, EV and solar buildouts) and its worst disasters (the one-child policy, zero-COVID, repression in Xinjiang and Tibet).

A major thread is the innovation-versus-manufacturing asymmetry. Wang argues it's structurally harder for the US to rebuild manufacturing scale-up capability than for China to close the gap in frontier science, since China has already improved rapidly on measures like top-cited scientific papers while flagship US manufacturers (Boeing, Intel, Detroit automakers) keep declining. His "spark and prairie fire" metaphor captures China's actual strategy: let American, German, and Japanese labs invent the breakthrough (as Bell Labs did with solar photovoltaics in 1954), then out-scale everyone else in manufacturing it, a pattern he also sees replicated in the Shenzhen ecosystem, where iPhone assembly workers moved on to build drones, then EV batteries, compounding tacit process knowledge that can't easily be written down or transferred elsewhere.

On investing, Wang unpacks why cash-generative giants like ByteDance and Alibaba trade at steep discounts to what similar US businesses would command: a Chinese stock market that has stayed roughly flat despite averaging 9% GDP growth over two decades, unpredictable Communist Party interventions (ByteDance's forced 2018 self-criticism letter, Ant Group's blocked 2020 IPO after Jack Ma's public criticism of regulators), and geopolitical risk tied to Taiwan and the offshore VIE structures most foreign investors actually hold. He walks through several vertically integrated Chinese companies as case studies, Huawei, Meituan (with roughly 50 core business lines), and especially Xiaomi, whose electric SUV beat Apple's decade-long, ultimately cancelled Project Titan car program to market and set a Nurburgring lap record.

The conversation turns to governance comparisons and history. Wang argues China studies the Soviet Union's collapse (Gorbachev's simultaneous economic and political reform) and Japan's stagnation (its closed, nearly all-domestic-value-add export model) specifically to avoid repeating either mistake, staying open to foreign manufacturers in a way Japan never was. He is skeptical of near-term political liberalization after Xi Jinping and argues state capacity has a ceiling: too little state capacity leaves the US unable to execute what it says it wants to do (California high-speed rail, Amtrak upgrades), but too much lets China's engineers run society off the rails before anyone can correct course, prompting his half-joking prescription that the US become roughly 20% more engineering-minded and China roughly 50% more lawyerly.

On AI specifically, Wang frames the next five years as a genuine toss-up rather than a settled US lead: while the US currently leads on compute and Nvidia chip access, China is adding roughly ten times as much solar generation capacity this year and has dozens of nuclear plants under construction versus none in the US, and he flags the possibility of a reverse brain drain of Chinese-origin AI researchers back to Chinese labs. He closes on the human cost behind China's technological rise, especially the one-child policy's toll on women's bodies, which became personal for him while writing the book, and on the risk of a US-China conflict, which he sees as more likely to be triggered by an unpredictable regional incident (a border skirmish, a South China Sea confrontation) than a deliberate invasion of Taiwan.

Notable Quotes

"China can just wait for a lot of these labs in the US and Germany, Japan, to figure out these new sparks, and then they set off the prairie fire." - Dan Wang

"Computer chips, potato chips, what's the difference?" - Dan Wang, quoting a former chairman of the Council of Economic Advisers

"I want the US to be 20% more engineering, and I want China to be 50% more lawyerly." - Dan Wang

"Maybe there can be too much state capacity. Maybe you don't want a state to be too efficient." - Dan Wang

"I decided over the course of writing this book that I wanted to write the truest story that I know." - Dan Wang