Goldman Sachs: No Broad Bubble in China's AI Sector; Massive Foreign Allocation Gap Remains
Claire Weston
Goldman Sachs chief China equity strategist Kinger Lau argues that Chinese AI stocks' total market cap has not yet priced in AI's full economic potential, meaning no systemic bubble exists; meanwhile, foreign funds allocate just 1% of their AI holdings to China — against an 11% share of global AI market cap.
Why does Goldman say China's AI sector isn't in a bubble?
Goldman estimates Chinese AI companies account for roughly 11% of global AI market cap, but the ratio of that cap to AI's potential economic contribution suggests overall pricing is not inflated.
At the June peak, some A-share AI hardware names plus ChiNext and STAR Market valuations hit five-year highs — clear signs of localized overheating.
After a month-plus pullback, valuations have returned to reasonable levels. This means → Goldman classifies this correction as a healthy reset, not a bubble bursting.
Why hasn't foreign capital moved in at scale?
Foreign funds allocate only about 1% of their total AI positions to China, against an 11% market-cap share — a massive gap. In plain terms = global funds loaded up on AI stocks but barely touched Chinese ones.
Among actively managed global mutual funds Goldman tracks — roughly $1.4 trillion in assets — emerging-market-focused investors have moved to a slight overweight in Chinese equities for the first time in about fifteen years.
Hedge fund gross exposure is near its cycle peak, yet net long positioning remains low. This reflects a stance of actively hunting opportunities while still prioritizing hedges — no broad directional bet yet.
Will Hong Kong IPOs drain market liquidity?
Year to date, more than 100 new listings have priced in Hong Kong, raising over $30 billion; foreign investor participation has returned to its highest since 2021.
The market worries large IPOs will siphon existing liquidity. Goldman's counter: dividends and buybacks from A-share and Hong Kong-listed companies may exceed RMB 4 trillion this year — more than enough to cover IPO-related funding demand.
In plain terms = the cash companies return to shareholders outweighs what new listings take from the market, keeping liquidity pressure manageable.
What's the key anchor for the second half? Earnings delivery
Goldman maintains its forecast of roughly 10% earnings growth for MSCI China full-year, and is more bullish on A-shares at 20% year-on-year growth.
Q1 MSCI China earnings fell 8% year-on-year, dragged mainly by heavy subsidy spending at internet platforms; Q2 and Q3 should see a meaningful recovery.
An earnings inflection for Hong Kong-listed internet companies is expected within one to two quarters, driven by three tailwinds: subsidy competition cooling lifts margins, AI commercialization accelerating creates new revenue, and legacy businesses (e-commerce, gaming) hold stable fundamentals.
Which sectors benefit most from earnings catalysts?
Materials — improving global supply-demand dynamics plus sustained AI-chain capex provide a dual driver.
Full-chain AI hardware — the industry upcycle has not been derailed by the short-term pullback; earnings visibility remains intact.
Energy — geopolitical tensions support the oil-price floor, and current valuations sit at lows. This means → Goldman sees these three sectors as having the highest probability of earnings delivery, underpinning potential alpha.
Inside the AI chain, which three tracks does Goldman favor most?
Power infrastructure: Chinese power-equipment makers are steadily gaining global share, sector valuations are low, and large-scale compute buildouts will keep pulling demand. In plain terms = AI needs compute, compute needs electricity, and power equipment is the most foundational beneficiary.
Hardware infrastructure (PCBs, optical modules, data centers): earnings visibility is very high, with steady delivery expected over the next two to three years.
Physical AI (industrial intelligence, humanoid robots): leveraging China's deep manufacturing base, these companies have broad overseas expansion potential. Lau also noted that Chinese large-model token costs are far below overseas rivals — AI tokens could become China's next major export growth driver after physical goods.
Content is for reference only, not financial advice.