China’s AI sector posted a blistering 65% gain across the first half of 2026, yet the mood among professional investors has shifted from euphoria to caution. What looked like a broad, sustained rerating of an entire technology ecosystem is now colliding with warnings from major banks and headline-grabbing bearish positions that could presage a more volatile second half.
Market performance and drivers
The surge in AI-related equities reflects several reinforcing dynamics: a wave of high-profile IPOs and secondary listings, rapid deployment of AI services across consumer and industrial applications, and concentrated flows into indices and ETFs that overweight AI and semiconductor names. Companies from cloud providers to chipmakers and model houses have benefited as investors reprice long-term growth potential into current valuations.
According to a July report by the South China Morning Post, China’s AI stocks rose 65% in the first half of 2026, a performance that outpaced most global tech indices and drew comparisons to past sectoral manias. Domestic structural advantages, abundant state-led investment in compute capacity, fast-growing token-based model usage, and an active startup-to-unicorn pipeline, have sustained optimism. EastFrontier reporting has documented how AI is powering the formation of new unicorns and reshaping capital markets across China.
Early signs of froth and professional skepticism
Still, several market signals now point to froth. Institutional voices that earlier celebrated China’s technology catch-up have begun issuing cautionary notes. Morgan Stanley, for example, has provided more guarded guidance to clients, highlighting valuation risk and the narrowing gap between promise and near-term revenue conversion. High-profile hedge fund manager Michael Burry’s publicly reported bearish position on parts of the sector has amplified attention, serving as a reminder that concentrated rallies breed concentrated downside.
On the ground, passive index reweights and retail enthusiasm have amplified flows into AI names, pushing prices ahead of earnings and prompting analysts to question whether multiples reflect sustainable business-model improvements or merely momentum. EastFrontier reported on structural flows into AI-favored indices and the passive capital rerouting that turbocharged share prices in recent months. That dynamic can accelerate upside and downside alike.
The valuation paradox: strong narrative, mixed fundamentals
Beneath the headline gains, fundamentals look mixed. Many AI firms are still investing heavily in R&D and compute and have yet to show consistent profit improvement; revenue recognition from tokenized APIs, model-as-a-service and agent deployments remains nascent. Economists and some sell-side analysts have noted that China’s AI spending boom has not yet translated into proportionate macro growth, a divergence that raises questions about the durability of current valuations. In short, market prices may reflect a plausible future that has not yet been realized.
Compounding the risk picture: policy and geopolitics
Regulatory and geopolitical factors add another layer of uncertainty. Beijing’s active industrial policy and procurement can underwrite demand for domestic chips and cloud services, but it can also produce cliff-edge risks if enforcement priorities shift or if political considerations limit certain foreign listings and capital flows. Meanwhile, export controls and international scrutiny over dual-use AI technologies complicate cross-border partnerships and could constrain revenue opportunities for some firms.
Why the caution matters now
Cautionary notes from Morgan Stanley and Michael Burry’s widely reported bearish stance matter because they shift the narrative among institutional investors. Where momentum-driven rallies were previously self-reinforcing, the arrival of respected skeptics can catalyze re-pricing, especially when combined with high levels of passive and retail exposure. The balance between long-term structural growth and short-term valuation discipline is therefore fragile.
What a correction would look like
A significant pullback could take several forms: selective weakness in the most richly valued model and platform stocks; a broader sector-wide retracement if passive flows reverse; or episodic volatility tied to regulatory headlines. Because many firms are capital-intensive and still scaling, even a modest correction could meaningfully slow hiring, investment in compute, and smaller IPOs, with knock-on effects for suppliers and regional economies.
Implications
• For investors: The market has entered a more binary regime where careful differentiation matters. Long-term exposure backed by revenue traction, diversified customer bases, and strong cash generation will fare better than momentum-driven positions. Risk management, such as position sizing, hedges, and attention to liquidity, is crucial.
• For companies: The pressure to convert token usage and deployments into durable, paid revenue streams will intensify. Firms that articulate clear paths to profitability and demonstrate unit economics for AI services are more likely to maintain investor confidence.
• For policymakers: Balancing stimulative support for domestic AI infrastructure with measures to prevent leveraged speculation will be important. Beijing’s role as both market promoter and regulator creates an opportunity to channel capital into productive investments while containing systemic risk.
• For global competition: A correction in prices would not erase the structural gains China has made in AI adoption, chips and open-model deployment. But volatility could slow international dealmaking and the pace of overseas expansion, providing competitors time to regroup.
The stretch run of 2026’s rally has shifted from being a unilateral celebration of China’s AI ascendancy into a more complex test of durability. The 65% H1 gain underscores rapid progress, but the newfound chorus of caution — from Morgan Stanley’s notes to Michael Burry’s short bets — signals that investors, companies and policymakers now face the harder work of converting technological momentum into stable, long-term value. According to the SCMP, bubble fears are already beginning to weigh on sentiment. How that tug-of-war resolves will determine whether the sector’s rally is a sustainable revaluation or a cycle in which rapid gains are followed by painful, but cleansing, reset.