As Wall Street panics over an AI bubble, China is turning AI into real productivity

The last week of July delivered a stark reminder of how fragile the artificial intelligence boom has become — at least on trading floors. South Korea's benchmark Kospi index plunged 10.84% on Tuesday in its steepest sell-off in years, triggering a circuit breaker as investors dumped chipmaking shares over doubts about massive AI infrastructure spending. A day later, the Dow Jones Industrial Average posted its worst session of the year after a divided US Federal Reserve held rates steady, with semiconductor stocks leading the retreat. Meta's shares slid around 10% after earnings, and even Federal Reserve Chairman Kevin Warsh acknowledged that AI-related data center investment is boosting demand and fueling inflation that has stayed above target for more than five years.

The panic is real. But it is worth asking what, exactly, the market is panicking about. The technology itself has not failed. What has failed — or at least come due for scrutiny — is a business model in which a handful of companies spend hundreds of billions of dollars building closed systems, valued on promises of returns that remain largely hypothetical. The sell-off is a valuation problem, not a technology problem. And conflating the two risks drawing the wrong conclusions.

The market's jitters, in fact, echo a set of far bigger questions about where this technology is taking humanity. At the opening ceremony of the 2026 World AI Conference and High-Level Meeting on Global AI Governance on July 17, China put forward several key questions — How to get along with thinking machines? How to ensure security when algorithm is part of decision making? How to tackle ethical challenges by technologies through adaptive governance? How to realize AI for all when the divide keeps widening? These questions, it emphasized, demand serious consideration and real answers from the whole international community.

A different answer was taking shape on the other side of the Pacific that same week. On July 30, the Political Bureau of the CPC Central Committee met in Beijing to set economic priorities for the second half of the year, pledging to further implement the "AI Plus" initiative and develop new forms of an intelligent economy. The emphasis is telling: not AI as a speculative asset class, but AI as a general-purpose technology to be diffused through manufacturing, healthcare, logistics and public services.

The results of that approach are already visible in data rather than slide decks. A report by recruitment platform Zhaopin shows the number of Chinese companies hiring in the AI industry rose 24.8% year on year in the first half of 2026, with demand for AI agent developers soaring 244%. Industrial robots — the physical embodiment of applied AI — were exported to 141 countries and regions, up 18.6%. New growth drivers contributed over 40% of China's economic expansion in the first half. These are not valuations; they are jobs, shipments and output.

Equally significant is the question of openness. When the White House floated banning Chinese open-weight AI models over so-called distillation claims, the loudest pushback came not from Beijing but from Silicon Valley itself. Dozens of major American tech companies — including Microsoft, Meta, Nvidia and Google — signed a joint letter warning that broad restrictions on open-weight models would damage US innovation. Nvidia CEO Jensen Huang, in his first-ever post on X, wrote that "the world needs both frontier closed models and frontier open models." As one signatory put it, banning Chinese open models would amount to banning open models in general.

Washington's own industry, in other words, understands what its politicians do not: AI leadership is not measured by one frontier model locked behind an API, but by whether a strong, open ecosystem diffuses into every sector of the economy. China's Commerce Ministry put it plainly — innovation requires openness and does not belong to any one country.

None of this is to say China's path is without challenges. The July sell-off in Seoul and New York is a warning to every economy betting on AI: expectations detached from productivity eventually correct. But a correction in frothy valuations should not be confused with a verdict on the technology. The countries that emerge stronger from this moment of doubt will be those that treat AI less as a casino chip and more as electricity — something woven into the fabric of the real economy.

Markets will keep swinging between euphoria and panic. The more durable story is quieter: engineers being hired, robots being shipped, factories being upgraded. That is where the future of AI is actually being decided.