This week the KOSPI fell more than 14%, but UK stocks actually rose 3%. Can spreading money across several countries avoid a shock like this?

This week's winners were decided not by country but by AI weighting

The ranking lines up almost inversely to AI hardware weighting. In the KOSPI, Samsung Electronics and SK hynix together made up 43.1% of market cap at the end of April (Korea Exchange data, cited by Financial News). In the Taiwan TAIEX, TSMC alone accounts for over 40% (Focus Taiwan). Conversely, per Goldman Sachs data, about 60% of new data center capacity in 2026 is in the US and Canada and about 10% in Europe (cited by Yahoo Finance). European indexes had few AI capex beneficiaries to begin with.

The UK rose for the same reasons as the sectors that rose in the US

The FTSE 100 had its best week in six months at 10,881.90. Defense, pharmaceuticals such as AstraZeneca and GSK, banks and mining stocks led the gain (BBN Times, citing Reuters, 2026-10-16). In the US this week too, energy, health care, financials (XLF +1.44%) and defense (DFEN +2.98%) rose.

In other words, buying a UK ETF was not diversifying by country but switching sectors. The same effect was available inside the US through sector ETFs.

Europe did not escape the AI shock either. The European tech sector index was -4.8% for the week through 10-15, and semiconductor stocks such as ASML and Infineon fell 5-8% in a day on some days (Sharecast and others, 2026-10-15). The index held up because the tech weighting was small. Bloomberg reported that European stocks had their biggest weekly outperformance versus the S&P 500 since March (Bloomberg, 2026-10-16).

Korea and Taiwan are both semiconductors, so why was the drop three times as large

In both markets one or two semiconductor stocks account for over 40%. Yet the KOSPI fell -14.46% and Taiwan -4.47%. Concentration alone does not explain the gap. Three factors remain.

• Memory versus foundry. Of this week's negatives, China's CXMT announcement targeted memory. TSMC is a foundry and not a direct competitor.
• How much each had risen. One tally has year-to-date gains at the end of May at about 87% for the KOSPI and about 52% for Taiwan (Yardeni Research, cited by StockWireX). Profit-taking was larger in the market that had risen more.
• Foreign investor flows. USD/KRW fell 1.00% to KRW 1,329.60, a stronger won, yet selling continued. That means the reason was position reduction, not currency losses.

We could not calculate how much each of the three factors contributed. Treat it as directional only.

The trap of country diversification

Buying a country ETF means buying the sectors that index holds.

The UK's advantage this week makes next week's risks just as clear. At the October 28 UK budget, additional taxes on banks are being discussed, weighing on NatWest and Lloyds, and the UK 10-year yield is near 5.44% (BBN Times, 2026-10). If oil drops or ceasefire news emerges, the FTSE's energy and defense names would be the first to reverse. The US FOMC decision also comes the same day.

What investors can do

• Recount your holdings by sector. If you hold both a KOSPI ETF and a Nasdaq ETF, there are two countries but the AI exposure overlaps into one. For a domestic index ETF, first check the weight of the two semiconductor stocks.
• Do not chase this week's No. 1 market. The FTSE had its best week in six months and is about 1% from its 52-week high. Buying now is a bet on oil and defense.
• If diversification is the goal, pick a market with a different sector mix. Build small positions in developed-market indexes with low AI weighting, and add only after the October 28 FOMC and UK budget have passed.
• Country diversification worked against a sector shock like this week's. But it does not protect against shocks that hit all markets at once, like oil and interest rates. The short-term bond share in answer 1 covers that side.