Wall Street is at record highs. Beijing is deflating asset bubbles. The divergence, analyzed in a new South China Morning Post piece, is creating a structural tension at the heart of China’s AI economy: the country is generating some of the most spectacular AI-driven valuations in the world, CXMT surged 471% on its debut, briefly becoming China’s most valuable listed company, while simultaneously pursuing financial policies designed to prevent the kind of speculative excess that produces those valuations.
The divergence: two financial systems, two approaches to AI euphoria
The contrast between Wall Street and China’s A-share market has rarely been sharper. US equity markets are pricing in an AI-driven productivity revolution, with technology valuations at historic highs and investor appetite for AI-adjacent stocks seemingly insatiable. China’s markets, by contrast, are operating under a deliberate policy of bubble prevention, regulators have been tightening margin lending, scrutinizing IPO pricing, and discouraging the kind of retail speculation that drove previous A-share boom-bust cycles.
The CXMT debut illustrates the tension perfectly. The memory chipmaker’s 471% first-day surge was extraordinary by any measure, it briefly topped ICBC as mainland China’s most valuable listed company and eclipsed Intel’s market cap. But the surge happened despite, not because of, Beijing’s financial policy preferences. The 212x retail oversubscription that preceded the listing was exactly the kind of speculative frenzy that regulators have been trying to dampen. The fact that it happened anyway suggests that investor demand for AI-adjacent exposure is overwhelming the policy guardrails.
The capital flight problem
The SCMP analysis argues that Beijing’s bubble-deflation instincts are creating a structural capital outflow problem. When domestic markets are deliberately suppressed while US markets are surging, rational investors move capital to where returns are higher. That dynamic is already visible in the crypto derivatives market, where Chinese investors have been using platforms like Hyperliquid to gain synthetic exposure to AI stocks that Beijing has restricted or made difficult to access domestically.
The analysts quoted in the SCMP piece argue that China needs to build “good bubbles”, speculative enthusiasm directed at strategically important industries like AI and semiconductors, rather than suppressing all speculative activity. The argument is that the US AI boom is partly a product of the willingness to tolerate and even encourage speculative excess in strategically important sectors, and that China’s risk-averse financial culture is a competitive disadvantage in this environment.
The AI investment paradox
The deeper paradox is that China’s AI industry needs both things simultaneously: the disciplined, long-term capital allocation that state-backed investment provides, and the risk-tolerant, return-seeking capital that drives the kind of speculative valuations that attract talent, signal ambition, and create the financial incentives for frontier model development.
As EastFrontier analyzed in its coverage of China’s household savings as a tech financing engine and investment with Chinese characteristics, Beijing is trying to thread this needle by channeling household savings into capital markets and state-guided funds. The CXMT IPO, oversubscribed 212 times by retail investors, suggests that the demand for AI exposure is there. The question is whether Beijing’s financial policy framework can harness that demand productively, or whether the bubble-deflation instinct will continue to divert capital to Wall Street.
What the divergence means for the US-China AI competition
The Wall Street-Beijing divergence has a direct bearing on the AI competition. The US AI boom is being financed, in significant part, by equity markets that are willing to assign enormous valuations to AI companies on the basis of future potential rather than current earnings. That willingness to fund speculative AI bets has allowed US frontier labs to raise the capital needed for the next generation of training runs, data center buildouts, and talent acquisition.
China’s more cautious financial policy means that its AI companies must rely more heavily on state-backed funds, corporate investors like Tencent and Alibaba, and the kind of patient capital that prioritizes strategic alignment over financial returns. That is not necessarily a disadvantage, patient capital can sustain long-term R&D programs that would be cut by return-seeking investors. But it does mean that China’s AI companies face a different set of constraints and incentives than their US counterparts, and that the financing gap between the two ecosystems is as important as the technical gap in understanding how the competition will unfold.
The SCMP analysis does not resolve this tension, it identifies it. The “good bubbles” argument is a call for Beijing to rethink its financial policy in the context of the AI race, not a prediction that it will. Whether Chinese policymakers are willing to tolerate the kind of speculative excess that has fueled the US AI boom, with all the systemic risks that entails, is one of the most consequential policy questions in the global AI competition.