Two Major Chinese Hedge Funds Warn AI 'Super Bubble' Nearing Collapse
0xBroomberg
Wealspring Asset and Shanghai Banxia Investment, managing a combined $1.7 billion-plus, have told investors that global AI stocks have formed a "super bubble" — with some of China's hottest names at risk of falling over 80%. It is one of the clearest collective bearish calls on AI from China's top private funds.
What exactly are these two funds afraid of?
Wealspring founder Yang Dong — known for correctly calling the 2007 market top — wrote that many Chinese AI infrastructure companies lack durable moats, run "fairly ordinary" business models, and need constant capital spending just to keep growing.
He compared the current market to the "buy anything" phase of China's 2015 A-share bull run, warning that some of the hottest A-share names are "very likely" to fall over 80% — without naming specific stocks.
This means → his core concern is not that AI is useless, but that prices have raced far ahead of business fundamentals — valuations and actual earning power have disconnected.
What different signal did Banxia Investment spot?
Banxia looked offshore: founder Li Bei predicted Anthropic's annualized revenue run rate will fall short of market expectations.
Two reasons: big tech companies are cutting AI spending as token costs — the computing fee paid each time an AI model is called — rise; and rivals are eating into Anthropic's share among developers.
In plain terms = if even the AI sector's most-hyped company cannot sustain its expected revenue growth, stock prices built on those expectations are in even greater danger.
How costly has the bearish bet been so far?
Per CITIC Securities' May fund-view summary, at least four Chinese hedge funds are bearish on AI, four are bullish, and seven have not taken a stance — the majority are sitting on the fence.
Being bearish has not been free: Wealspring's Zhiyuan fund and Banxia's low-volatility macro fund both posted small losses in the early months of this year, though both remain profitable over a longer horizon.
This reflects a real cost: the CSI AI Index is up over 35% year-to-date, and South Korea's KOSPI — lifted by SK Hynix and Samsung — has rallied nearly 100%. Bears are paying a tangible opportunity cost.
What should investors watch next?
Despite the short-term pain, neither fund has reversed its bearish stance.
This means → they are not making a short-term timing call; they are betting that AI monetization data will eventually reveal a material gap — revenue growth failing to keep pace with capital expenditure.
In plain terms = the referee in this bull-bear standoff is not market sentiment but the actual earnings data AI companies report over the coming quarters.
Content is for reference only, not financial advice.