JPMorgan initiated coverage of China's AI value chain for the first time, estimating $544 billion in cumulative operating profit from 2026 to 2030 and rating Alibaba, Tencent and Z.AI Overweight with targets of HK$205, HK$690 and HK$2,000. The bank expects hardware's share of industry profit to drop from 93% to 30% by 2030 as models and applications take 54%. Six other names received Overweight ratings, while Meitu, MiniMax, Kingdee, JST Group and Manycore started at Neutral.
JPMorgan Chase put China's artificial intelligence supply chain under research coverage for the first time, telling clients that $544 billion of operating profit will be generated across the industry between 2026 and 2030 and that Alibaba, Tencent and Z.AI are its top picks.
The bank rates all three Overweight, with price targets of HK$205 for Alibaba (09988.HK), HK$690 for Tencent (00700.HK) and HK$2,000 for Z.AI (02513.HK), according to the report dated Sept. 24. It assigns a constructive view to the sector.
JPMorgan's thesis rests on a redistribution of profit rather than a simple expansion of it. Hardware accounts for 93% of industry operating profit in 2026, a share the bank expects to fall to 30% by 2030. Cloud's share rises from 7% to 15% over the same period, while the model and application layer climbs from zero to 54%, moving from losses to the largest single slice of the profit pool.
The revenue math behind that shift is lopsided. Between 2026 and 2030, JPMorgan expects hardware revenue to grow 1.8 times, cloud 12.5 times and models plus applications 26.9 times. Total industry operating profit rises from $27 billion in 2026 to $243 billion in 2030. The bank attributes the divergence to token demand, which it expects to expand roughly 60-fold over the period, with enterprise customers accounting for 69% of consumption by 2030 and consumers the remaining 31%.
The selection logic differs by name. Alibaba is cited for vertical integration across chips, cloud infrastructure and proprietary models, a closed loop from computing power to applications. Tencent is credited with distribution channels, proprietary user scenarios and transaction execution that JPMorgan says are not yet reflected in its valuation. Z.AI is the only pure-play model developer among the three picks, and the bank expects differentiated intelligence capabilities to drive paying-user growth.
The initiation extends to six further Overweight ratings: SenseTime (00020.HK) at HK$2, Phancy (06682.HK) at HK$50 and Xunce (03317.HK) at HK$160. Meitu (01357.HK) starts at Neutral with a HK$4.5 target, and MiniMax-W (00100.HK) at Neutral with HK$300. Kingdee International (00268.HK), JST Group (06687.HK) and Manycore Tech (00068.HK) also received Neutral ratings, with target prices not disclosed in the report.
The calls land against heavy short positioning in the same names. Alibaba's shorted share volume rose from 9.61 million shares in late August to 14.02 million by Sept. 23, peaking at 19.50 million in a single session, the highest since Aug. 26, with short-selling turnover climbing from HK$1.091 billion to HK$1.554 billion. Tencent's shorted volume went from 2.91 million shares to 3.43 million, peaking at 7.40 million, the most since July 22, on peak turnover of HK$3.367 billion. Z.AI's shorted volume jumped 71.4% to 1.43 million shares, with a single-day record of 4.22 million and peak turnover of HK$3.098 billion.
As of 12:25 p.m. on Sept. 24, short-selling ratios stood at 21.855% for Alibaba, 13.703% for Tencent and 9.217% for Z.AI, according to exchange data.
For holders of Hong Kong-listed AI names, the report reframes the question from who builds the infrastructure to who converts it into revenue, which favors platform operators with existing distribution over hardware suppliers. The concentrated short base in Alibaba, Tencent and Z.AI leaves room for a squeeze if liquidity improves or a positive catalyst lands, though it also caps near-term rebounds. The next test is whether the model and application layer's profit share tracks the forecast in coming quarterly disclosures from Alibaba and Tencent.
This article is for informational purposes only and does not constitute investment advice.
