Chinese equities have once again slipped into a bear market, but the headline obscures a far more interesting reality: there are effectively two distinct markets operating inside China right now, moving in opposite directions. Understanding this split is the key to making sense of both the weakness and the opportunity.
The Tale of Two Chinas
The most striking feature of the Chinese market this year is its internal divergence. On one side sits the onshore, mainland market — particularly the technology and semiconductor sector — which has been extraordinarily strong. On the other side sits the offshore, overseas-listed market, centered in Hong Kong, which has been persistently weak. This pattern has become almost a daily ritual, a kind of deja vu: semiconductor and hardware stocks on the mainland rise, while internet stocks and the broader market offshore fall.
This divergence can be tracked precisely through two contrasting baskets. One tracks the STAR Market in mainland China — specifically the top fifty science and technology names listed there, which are heavily weighted toward hardware technology. The other tracks China's internet companies, which are predominantly listed offshore. The gap between them has been dramatic: the mainland hardware-tech index is up roughly 50% or more year-to-date and even hit a record high recently, while the offshore internet basket has been under heavy, sustained pressure.
What Is Driving the Weakness?
The question naturally arises: what explains the renewed weakness — growth concerns, geopolitics, or investor fatigue? The answer is a combination of factors, but two stand out.
First, there is intense AI exuberance concentrated in the mainland semiconductor stocks. Money is flooding into the hardware names that are perceived as direct beneficiaries of the AI buildout.
Second, there is a genuine pullback and a sluggish consumer within China. This soft domestic demand weighs disproportionately on the internet platforms — the software names like Alibaba, Meituan, and Tencent — which happen to be mostly listed offshore. Crucially, investors have not yet captured these internet companies as part of the AI story, even though they are in fact central to AI development and deployment in China. The market is treating them as consumer-exposed laggards rather than AI players.
The Capex Recipients Versus the Capex Spenders
The deeper logic behind this split mirrors a dynamic playing out in the United States and around the world. Investors are rewarding the recipients of capital expenditure while punishing the spenders of it.
The recipients — the chipmakers and hardware suppliers — are seeing this AI capex land on their books as revenue, and investors applaud. This group includes Nvidia, the Chinese semiconductor names captured in the mainland tech basket, SK Hynix in Korea, and TSMC in Taiwan. These have all been favored. The same pattern appears among U.S. megacaps: the heavy spenders, such as Apple and Google, have underperformed the recipients, with Apple down year-to-date.
The spenders, by contrast, are being penalized. The market looks at companies like Alibaba — which is pouring billions into artificial intelligence — and reads that spending as money going out the door rather than as investment in the future. But that interpretation misses the point. This capex is not waste; it is investment whose payoff will come later.
The irony is that in China the internet platforms are precisely the entities that will deliver AI productivity to the end user — to individuals, to businesses, to the broader economy. Alibaba and Tencent are the primary channels through which people and companies will actually use AI. Yet these are the very stocks that have been sold off most aggressively, even as their suppliers have benefited enormously from the same trend.
A vivid illustration of this imbalance comes from MiniMax, a large language model provider and a new Hong Kong IPO this year that has performed very well. MiniMax outsources most of its cloud compute to Alibaba. So MiniMax's stock is up — while Alibaba, its main cloud revenue customer and the entity actually supplying the compute, is down. This kind of disconnect simply does not make logical sense, which is why the market currently looks so imbalanced.
The Case for Mean Reversion
Given this distortion, an eventual rebalancing — a mean reversion between the two sides — seems likely. The honest caveat is that no one can say when it will happen; the timing is genuinely unknowable. But the direction of the imbalance, with the AI-serving platforms sold off while the hardware suppliers have soared, suggests it cannot persist indefinitely.
Importantly, investor sentiment toward the internet names has not fundamentally collapsed. Analysts at Daiwa, for instance, recently trimmed their price target on Alibaba — but only modestly, while maintaining a buy rating on the shares. The bullish thesis is still intact among professionals; the share prices simply have not reflected it. By technical measures, the offshore internet basket is deeply oversold, sitting at a 14-day RSI of around 23 — about as clear an oversold signal as one can get.
The Consumer and the Question of Stimulus
A meaningful catalyst for rebalancing would be more aggressive consumer stimulus from the government in Beijing. This would help the sluggish consumer get off the ground, directly benefiting the consumer-facing internet platforms.
Here the candid assessment is sobering. Markets have been wondering and waiting for more consumer stimulus for years, and the market currently appears to assume it simply will not happen — the prospect has fallen off the radar. The recent 618 (June 18th) shopping festival produced decent sales, but nothing eye-watering or earth-shattering. The consumer still needs to find firmer footing, and that remains the open question hanging over the offshore names.
Valuations: Cheap, But Is Cheap Enough?
The valuation gap underscores how far the offshore names have fallen. The China internet basket trades at roughly 12 times price-to-earnings. By comparison, the broader MSCI Emerging Markets Index trades at about 18, and U.S. internet companies average a price-to-earnings ratio with a 30 handle. The current multiple on the China internet portfolio is the lowest recorded in four years — genuinely bargain-basement territory.
But cheap does not automatically mean good. A central risk with Chinese stocks is the threat of sudden, out-of-the-blue regulatory crackdowns — the regulatory and jurisdictional risk inherent to investing in China. A discount to account for this risk is entirely justified. The key argument, however, is one of magnitude: even applying a substantial discount — in the range of 30% to 50% — to account for that regulatory and jurisdictional risk, these companies still come out as undervalued relative to their U.S. peers. The risk is real and should be priced in, but it does not erase the underlying value.
Reframing the Regulatory Overhang
The nature of recent regulation is itself widely misunderstood, and a closer look reverses the usual gloomy interpretation. Most current regulation has been aimed at curbing excessive competition, particularly in the instant-commerce industry. There, companies have been operating essentially at a loss, pouring out subsidies to capture the same wallet share. Because the consumption pie is not growing fast, these firms have been chasing the same urban consumers, handing out so many subsidies that none of them make any money.
The regulatory response has been designed to assert the rights of and protect mom-and-pop shops, and to stop these large companies from eating into their own bottom lines. Over the long term, this is actually positive for the companies' balance sheets, because it should allow margins to improve as the destructive subsidy war is reined in.
This reframing points to a broader shift in posture: the Chinese government now appears more supportive of the tech sector and of private industry than in previous years — a meaningfully positive development. The paradox is that even regulation intended to improve margins over time has, in the short run, weighed on the sector and added to the offshore weakness.
The Bottom Line
The picture, then, is coherent once the two Chinas are separated. The mainland hardware-tech complex is riding a wave of AI capex flowing in as revenue. The offshore internet platforms — the very firms that will ultimately deliver AI's benefits to end users — are being treated as capex-burning, consumer-exposed laggards and have been sold down to four-year-low valuations that remain attractive even after generously discounting for regulatory risk. The imbalance is stark and, in the long run, hard to justify. A rebalancing looks probable; consumer stimulus would accelerate it. The only thing that cannot be pinned down is when.
