While American equities spent two years repricing upward on the promise of artificial intelligence, Chinese technology companies developed comparable capability and watched their market capitalization move the other way, a divergence that is now too large to explain away as a lag.
The gap is not about engineering. Chinese labs have shipped frontier-competitive models, Chinese manufacturers dominate the hardware supply chain, and Chinese applications have reached enormous user bases. What has not followed is the equity premium that attached to the same narrative in the United States, where a single disappointing print from a hyperscaler can move indices.
Several explanations compete. Domestic demand for paid software remains weaker than the advertising and enterprise models that fund American platforms. Capital controls limit how easily foreign money can express a view on Chinese technology. And the possibility of export restrictions or listing delistings attaches a permanent discount that no amount of technical achievement offsets.
The consequence for global allocation is that AI capability and AI equity exposure have become separable for the first time. An investor who believes the technology will matter can hold that view without holding the companies best positioned to build it, and vice versa. That separation is uncomfortable for a market that spent two years treating the two as identical.
The unresolved question is which market is pricing correctly. If Chinese technology eventually earns a comparable multiple, the current gap represents one of the larger dislocations in the sector. If it does not, the market is telling us that the ability to build a model was never the scarce asset, and that the scarcity lies in the monetization built around it.
Image source: i.ibb.co