Where investors may find the next 'big wave' for AI trade

Emerging markets, especially Taiwan and South Korea, are highlighted as the next major AI trade, with the iShares MSCI Taiwan ETF up almost 67% YTD and the iShares MSCI South Korea ETF up 109% as of Thursday's close. Urbanowicz said valuations outside the U.S. have not rerated as much, leaving room for further outsized gains in AI-linked chip names. He also pointed to the Goldman Sachs ActiveBeta Emerging Markets Equity ETF as a way to gain exposure, while still seeing U.S. AI positioning as intact.

Analysis

The cleaner expression of this theme is not a generic EM beta long, but a relative-value bet on the parts of the AI supply chain where pricing power has not yet been arbitraged away by U.S. enthusiasm. Taiwan and Korea sit upstream in memory, advanced packaging, and components, which means they can enjoy a second-order margin expansion if AI capex stays elevated while end-demand broadens beyond hyperscalers. The market is still treating this as a factor rotation story, but the better framing is a multi-year earnings revision cycle for the narrow set of firms whose products are bottlenecks, not just beneficiaries.

The biggest risk is that this trade is crowded through the same small set of large-cap exporters and semiconductor proxies, which makes it fragile to any pause in capex guidance or currency strength. In the near term, the move can keep running on flows for weeks, but the real test is over the next 1-2 quarters when inventory and order-book visibility matters more than narrative. If AI capex decelerates, these markets can de-rate quickly because their multiples have expanded on forward earnings, not current cash generation.

A more contrarian take is that the U.S. AI trade is not being “replaced” so much as broadened, and the best risk-adjusted expression may be to rotate within AI rather than chase EM index exposure. Investors are underestimating how much of the upside in Taiwan and Korea is already embedded in the most obvious ETFs after the sharp year-to-date move. The opportunity now is to selectively own the enablers with the cleanest direct AI linkage while fading parts of the EM basket that are merely along for the ride.

Catalyst-wise, watch for updated hyperscaler capex commentary, any memory pricing inflection, and FX moves versus the dollar; those are the fastest ways the trade either extends or snaps back. If the market starts rewarding earnings quality over beta again, the broad EM AI basket should lag the more focused semiconductor supply-chain names. That makes this a good tactical theme, but one that needs active position sizing and a willingness to trim into strength.

Key Decisions for Investors

• Long SMH vs short EEM for 3-6 months: express AI supply-chain leadership without taking broad EM macro risk; target 10-15% relative outperformance if earnings revisions stay concentrated in semis.
• Overweight TSM and/or ASML on pullbacks over the next 2-4 weeks: best risk/reward on bottleneck exposure with direct benefit from sustained AI capex; cut if hyperscaler guidance rolls over.
• Long KWEB? No — instead avoid broad EM ETF chasing and use call spreads on EWY or EWT only if you want beta with defined risk; 2-3 month tenor to capture flow momentum, not long-duration exposure.
• Pair long Taiwan/Korea semis against short non-AI EM cyclicals within MSCI EM proxies: reduce dilution from commodity and financial names that won’t see the same earnings uplift.
• If memory pricing inflects higher, add via options in MU / SK Hynix proxies where accessible; asymmetric upside over 1-2 quarters as supply discipline can translate into faster EPS revisions than the market expects.