Goldman Sachs Raises Bets on South Korea and Taiwan Tech Stocks on AI Demand
Goldman Sachs Raises Bets on South Korea and Taiwan Tech Stocks on AI Demand
Goldman Sachs has raised its 12-month forecast for South Korea's KOSPI index to 12,000, up from 9,000. At the same time, the bank upgraded Taiwan to a favorable rating and set a target of 51,000 for Taiwan's Taiex benchmark. Both moves rest on the same bet: that artificial intelligence is driving demand for chips and technology that will keep growing.
The new KOSPI target suggests the index could rise 33% from where it trades now. These moves put both South Korea and Taiwan near the top of Goldman's picks for Asian stock markets.
Why Tech Companies Matter in These Markets
South Korea and Taiwan are home to some of the world's biggest chipmakers and manufacturers. Companies like Samsung and SK Hynix make the memory chips that power AI systems. Taiwan hosts Taiwan Semiconductor Manufacturing Company (TSMC), which makes chips for many global technology companies. Both countries also have packaging specialists and design firms that benefit when AI demand rises.
When global tech companies spend heavily on AI infrastructure, they need more chips. That spending flows directly to these markets' largest listed companies. South Korean and Taiwanese firms control key parts of the supply chain—they make the memory, they design the chips, and they package them for use.
Goldman's strategists say this advantage should last, not just be a temporary spike. They point to orders for memory products that extend well into 2027, and to new technical demands from AI chips that require more advanced manufacturing and specialized packaging. These are signs of structural demand, not a flash in the pan.
What This Upgrade Really Signals
The broader context here matters. Goldman has raised its KOSPI target multiple times during 2026 as it watched AI spending grow. The firm's analysts have become more confident in what they can see: orders and revenue in the pipeline, not just hopes or guesses. Taiwan's upgrade from neutral to overweight—and the price target of 51,000—suggests the bank believes not just that earnings will grow, but that investors will be willing to pay higher prices for those earnings over the next year.
In my view, this resembles earlier technology cycles where genuine new infrastructure was built—like cloud computing investments in the 2010s—more than it does the dot-com bubble. The difference is that cloud had real revenue streams and productivity gains within a few years. AI infrastructure shows those signs already: enterprise customers are deploying it, costs are dropping, and spending is baked into major cloud providers' budgets through 2027.
What Could Go Wrong (and What Goldman Is Betting Won't)
These forecasts rest on several things staying true. They assume companies keep hitting earnings targets. They assume that chip demand doesn't cool or that inventories don't swell unexpectedly. They assume the Korean won and Taiwan dollar stay stable so international investors keep buying these stocks.
Trading volumes and bets placed on these markets have grown sharply during the recent rally. That's usually a sign that the move is broadly backed, not concentrated among a few big players. But it also means these markets are now sensitive to AI sector news in ways they weren't before. If AI spending slows, or if earnings disappoint, these stocks could fall just as fast.
Goldman's emphasis on 12-month targets reflects confidence that the company earnings they're counting on will actually arrive. That timeline aligns with when cloud companies and enterprise customers say they'll spend. Whether they do is the open question.
