China’s AI-linked technology sector has suffered a sharp financial correction, with the STAR50 index plunging about 40 percent and wiping roughly $480 billion from market value, Reuters reported. The sell-off has exposed a gap between investor expectations and the earnings of companies tied to AI, semiconductors and robotics.
Market exuberance had driven some listings to extraordinary highs, only for prices to reverse quickly once trading opened. Reporting showed Unitree Robotics shares fell around half at their market debut. Chipmakers and optical component suppliers also slid, with CXMT, MetaX and Moore Threads among firms losing large portions of their earlier gains. Goldman Sachs estimated AI-related stocks made up about 42 percent of the relevant segment of Chinese and Hong Kong market capitalization and accounted for roughly 57 percent of trading turnover.
Part of the strain comes from heavy investment in data center capacity that now appears underused. An MIT Technology Review report found many facilities sitting idle or poorly matched to current AI workloads. The emergence of DeepSeek suggested capable models require less computing power than previously thought, leaving firms that invested heavily in infrastructure exposed. Bloomberg reported DeepSeek raised $15 billion in a major funding round.
Beijing faces a policy choice between broad bailouts and selective support. October Party-State guidelines urged more AI capacity while cracking down on wasteful investment and holding officials accountable for losses. Reuters reporting suggested the government will likely let private and local actors absorb many losses while using state-backed loans and state-company action to preserve projects tied to domestic chips and public services, mirroring its approach in the property sector.
Implications For Firms Investors And Global Competition
The correction will test frontier research labs and listed hardware firms most severely. At the same time, large Chinese tech groups still show revenue momentum. Alibaba reported AI cloud revenue of $7.1 billion in its latest quarter, up 45 percent, and Baidu’s AI cloud infrastructure sales rose 50 percent, indicating real demand even amid market turmoil.
Global rivals face similar capital pressures. Reuters noted the four largest US cloud providers are on track to spend about $700 billion on capital expenditure in 2026, a level projected to consume 92 percent of operating cash flow versus 41 percent in 2023. Microsoft reported an AI business generating more than $37 billion a year, and Google Cloud’s backlog surpassed $460 billion, underscoring high stakes for returns on vast outlays.
Startups are most at risk, as fresh funding will be harder to secure. Lotus Asset Management’s Hong Hao said many companies may delay IPOs. The downturn could also force firms to build cheaper, leaner AI systems and compete on price, which may expand access in emerging markets. Reporting suggested Beijing’s two-speed approach will favor projects linked to self-reliance while pushing weaker commercial ventures to consolidate or fail.
