Goldman Sachs Advises Clients to Rotate into China's AI Value Chain
N.R. Finch
Goldman Sachs has issued its first explicit call to rotate positions from Korea's AI trade into China's AI value chain, driven by Korea's concentration-risk exposure colliding with catalysts around China's homegrown chip efforts — a signal that Wall Street is shifting from watching China AI to actively allocating.
Why is Goldman suddenly pivoting to China?
Goldman's thematic strategy team published *Trade Idea: Long China AI Value Chain*. Author Louis Miller wrote: "China AI has officially entered our radar."
This means → after tracking Korea's AI rally for nearly a year, Goldman is switching its primary thesis from Korea to China — not adding an option, but changing the main trade.
In plain terms = Goldman used to think Korea was the AI play. Now it believes China's opportunity is too large to ignore.
What news triggered the rotation?
Three stories broke simultaneously: Reuters reported DeepSeek is developing in-house chips to power its AI systems; The Information reported Zhipu is considering designing its own AI chips amid rising demand and U.S. export controls; the Financial Times reported Apple has begun testing memory chips made by CXMT in China-market products.
Bloomberg Intelligence analyst Steven Tseng said: "This is a clear catalyst — because of geopolitics, DeepSeek has to work with domestic Chinese chipmakers."
This means → Chinese AI companies are no longer just customers for U.S. and Korean chips — they are building their own supply chain. Once that chain takes shape, capital must reprice.
Why did Korea sell off so hard?
KOSPI fell 5.4% in a single day and has now dropped roughly 20% from last month's all-time high, entering a technical bear market — defined as a 20% decline from the peak. SK Hynix fell 5.7%; Samsung Electronics fell 6.3%.
Samsung had just reported quarterly profit up 19-fold year-on-year earlier the same week, yet shares kept falling. This reflects a market worried not about current earnings but about whether AI capital spending can last.
In plain terms = no matter how much money a company makes today, if the market doubts the money will be there next year, the stock still drops.
What does "concentration risk" actually mean here?
Fidelity International portfolio manager Ian Samson explained: global AI-driven semiconductor demand is effectively controlled by roughly $1 trillion in capex from a handful of big tech companies — if that chip spending proves unsustainable, downside risk is very significant.
This means → Korea's semiconductor sector lives or dies by the procurement budgets of a few U.S. tech giants. If those companies cut spending, Korea takes the first hit.
Foreign investors have net sold over $100 billion of Korean equities this year, and leveraged ETFs have amplified volatility in both directions.
How is the China side performing?
The Hang Seng China Enterprises Index rose as much as 4.5% intraday — its biggest single-day gain since February 2025. Alibaba surged over 13% in Hong Kong; Tencent gained more than 4%.
Reed Capital CEO Gerald Gan said: "The divergence between China and the rest of the world is particularly striking, creating attractive value opportunities in Chinese equities." His firm is building positions in major Chinese tech names.
Yet the Hang Seng Tech Index and the HSCEI are still down roughly 15% and 10% year-to-date. This reflects a rally starting from a deep base — plenty of room, but the foundation is not yet firm.
Can the rally last?
KOSPI's year-to-date gain has pulled back from a peak of about 116% to roughly 72% — still the largest among 92 major benchmarks tracked by Bloomberg. But this month the Hang Seng has become Asia's best-performing market, while KOSPI sits at the bottom.
In plain terms = Korea is still this year's big winner, but momentum is rotating toward China.
Whether the rally continues hinges on two things: whether Alibaba and other internet giants can validate bullish expectations with actual earnings in the coming reporting season, and whether Goldman's call draws enough institutional capital to follow through.
Content is for reference only, not financial advice.