Goldman Sachs Reaffirms A-Share Hard Tech Leadership, H-Share Internet Enters Watch Zone
0xBroomberg
Goldman Sachs reiterated on July 13 its tactical preference for A-share hard tech over H-share soft tech, while noting that H-share internet valuations have been compressed enough to begin gradual positioning — whether the trade holds depends on whether profits actually recover over the next few quarters.
How extreme is the A-share vs. H-share divergence this year?
The STAR 50 index has outperformed the Hang Seng Tech index by 68 percentage points year-to-date. ChiNext has beaten the Shanghai Composite and CSI 300 by 19 and 17 points respectively.
This means → capital has made a near-one-way bet on onshore AI hard tech this year, dumping offshore internet platforms hard enough to push the gap to a historical extreme.
In plain terms = investors who bought A-share hard tech won; those holding H-share internet lost — and the spread is historically rare.
H-share internet bounced — why isn't Goldman fully bullish yet?
Hang Seng Tech rallied 11% over the past two weeks but remains down 14% year-to-date. Goldman attributes the bounce to improved newsflow and depressed starting valuations, not profit support.
Earnings pressure is the core hurdle: MSCI China profits fell 8% year-on-year in Q1, dragged down visibly by internet — a sector accounting for roughly 35% of the index's earnings weight.
Since Q2 2025, internet companies have accumulated over RMB 180 billion in subsidy losses while shouldering massive AI capex — estimated at over $100 billion this year and $120 billion next year.
This means → the stock price has bounced, but profits haven't caught up. Goldman's stance is "start nibbling," not "go all-in."
When could internet profits recover — and what signals matter?
Goldman flags three watch-points: ① can subsidy losses narrow; ② can AI-driven revenue streams — cloud, agentic AI, AI tokens — start contributing; ③ can legacy e-commerce hold steady cash flow.
If an operating-profit inflection appears in Q2 or Q3, the market's valuation approach may shift from blended-earnings discounting to sum-of-the-parts.
In plain terms = right now the market values internet firms at a discount because subsidies and AI spending weigh too heavily on profits. Once earnings genuinely stabilize, the valuation method switches — and the upside opens up.
Can investors still chase A-share hard tech? What about bubble risk?
Goldman concludes that China's AI-related stocks do not constitute a bubble overall — since DeepSeek's release, the economic value AI creates through efficiency gains and new profit pools could be 50% to 100% higher than current prices reflect.
But risk is concentrated in pockets: semis and select A-share hard-tech names trade at elevated valuations versus history and global peers, with rising concentration and leverage risk.
Goldman's advice: limit one-directional exposure, diversify across different links in the AI value chain, and elevate earnings-delivery capability as a priority filter.
What is the fundamental difference between hard tech and soft tech?
Hard tech sits closer to the AI infrastructure supply side — it benefits from demand for compute, semiconductors, equipment, and infrastructure. It sells the shovels.
Soft tech and internet platforms are more like AI capex bearers — they still need to prove that AI spending and subsidies won't keep eating into profits. They buy the shovels and dig for gold.
This reflects the deeper logic behind the market split: one side has higher certainty and profits already booked; the other is spending big, with payoff still ahead.
How are foreign investors positioned on China — have they pulled out entirely?
Hedge-fund net exposure to mainland China stocks sits near the bottom of its range, but gross exposure is near a cycle high — funds haven't left, they've shifted toward market-neutral and single-stock alpha strategies.
Mutual-fund allocation to mainland China has risen to a multi-year high, with a mild overweight for the first time since records began in 2011.
On the IPO front, 100 companies have listed on the HKEX year-to-date, raising a combined $35 billion, with foreign cornerstone investor participation near 2021 highs.
This means → foreign capital isn't avoiding China — it's no longer betting on index direction. The shift is toward picking stocks and sectors: alpha, not beta.
Content is for reference only, not financial advice.