The Global Regime Change | Jason Hsu on AI, Factor Investing and What Investors Miss About China

Episode Summary

Executive Summary: The conversation argues that China is best understood as a fiercely competitive capitalist economy, not a monolithic state-planned system, and that this has powered manufacturing, EVs, and likely AI innovation. It also contends AI will displace many professional jobs, increase labor non-participation, and favor those who can blend human judgment with machine scale. Finally, it says factor investing is evolving toward broader, AI-assisted, more esoteric approaches, with China and emerging markets offering fresh opportunity.

Main Topics: China as competitive capitalism, not central planning (Priority: 5/5): China is framed as a market-driven economy with intense internal competition, hard bargaining, and profit motive, even if the state plays a strong strategic role. The speaker stresses that Beijing often acts more like a giant VC/LP than a command planner. Manufacturing dominance and trade power (Priority: 5/5): China's manufacturing scale is described as a near-monopoly at the global level, which changes tariff dynamics and makes it difficult for the U.S. to win broad trade fights against Chinese production. China's real estate reset and capital reallocation (Priority: 4/5): The property downturn is portrayed as an intentional and necessary bubble-pricking move to shift capital from unproductive real estate into more productive sectors like AI and EVs. AI as the next labor-disruption wave (Priority: 5/5): AI is compared to globalization's impact on manufacturing: it will cheapen labor, disrupt professional services, and force major changes in employment, policy, and human work roles. China, U.S., and the move toward a G2 world (Priority: 4/5): The speaker argues geopolitics is shifting from U.S.-dominated exceptionalism to a two-bloc world where the U.S. and China each develop separate innovation ecosystems that still cross-pollinate. Factor investing, China alpha, and AI-driven quant evolution (Priority: 5/5): Traditional factors have weakened in the U.S. but remain powerful in China due to retail dominance and less institutional arbitrage. AI and machine learning are viewed as tools to mine broader factor sets and esoteric signals. Portfolio implications: diversification and China tech access (Priority: 4/5): The discussion supports more diversified exposure away from heavy U.S. concentration, including separating China from broad EM and using vehicles like CNQQ to access Chinese tech directly.

Key Arguments: China is not simply state socialism; it combines state direction with fierce internal competition, which better explains the quality and cost advantages seen in manufacturing and EVs. The Chinese government behaves more like a massive VC/LP allocating capital across many bets, then letting entrepreneurs compete, rather than centrally selecting winners. Broad tariffs on China are ineffective because China functions like the world's manufacturing monopoly, while tariffs on individual firms can be absorbed through substitution inside China. China's real estate bubble needed to be pricked to prevent wealth from being locked into unproductive assets and to redirect capital toward higher-return sectors. AI will replicate the outsourcing effect of globalization, but this time against professional services, not just factory labor. Labor non-participation is likely to rise as AI makes more workers economically redundant, pushing governments toward larger fiscal support and public-sector-style employment. In the U.S., growth stocks have not been irrationally overpriced for years; market efficiency and real innovation explain why factor/value strategies have struggled. China remains a strong alpha reservoir because retail-heavy trading, limited institutional arbitrage, and behavioral biases create persistent inefficiencies. Machine learning can improve factor investing by distinguishing between truly decaying factors and cyclical ones, and by exploiting a larger universe of esoteric signals. Investors should separate China from the broader EM bucket because China is too large and distinct; it deserves dedicated analysis and allocation. AI will not eliminate the need for humans entirely; protected classes, liability-bearing roles, and human-connection jobs are likely to persist. The best edge in AI investing will come from human judgment plus machine scale, not from the machine alone.

Data Points: Relative AI training cost: 1/20th of the money - DeepSeek is cited as having achieved roughly 80% of the result at about one-twentieth of the training cost, illustrating underfunded innovation. Training result: 80% of the way - Used alongside the 1/20th cost claim to describe DeepSeek's efficiency in AI development. U.S. labor non-participation: Under 20% to close to 30% - The speaker says labor non-participation rose over the last 30 years as jobs were outsourced to Asia/China. Projected labor non-participation: 30% to 40% - Forecast for how AI could further reduce workforce participation. ETF comparison: CNQQ vs. QQQ - CNQQ is presented as a China version of the Nasdaq QQQ for tech exposure. EM outperformance period: About 9% per annum for almost a decade - The speaker cites emerging markets outperforming the U.S. from 2000 to 2010. Global market structure: G1 plus six - A phrase used to describe the old G7 era as effectively U.S. dominance plus six minor players. Trade bloc shift: G2 - The speaker says the world is moving from U.S. exceptionalism toward a U.S.-China two-power framework.

Pivotal Quotes: "China has become the world's factory. In fact, it may be the world's only factory." — Jason: Used to argue that China's manufacturing scale creates global pricing and trade leverage. "The Chinese government probably is the largest LPGP in sort of private equity and VC funds." — Jason: Explains the claim that Beijing acts more like a strategic capital allocator than a command planner. "AI is the same thing, right? We are just finding a cheaper source of labor." — Jason: Summarizes the central analogy between globalization's labor arbitrage and AI-driven labor substitution.

Implications: Investors may need to rethink U.S.-centric portfolios, separate China from EM, and prepare for AI-driven labor disruption. The biggest winners may be those who combine human judgment with machine tools and identify new sources of alpha in China and beyond.