When you buy an index, you're not buying diversification — you're increasingly buying concentration into one trade (AI/semiconductor).
| Index | Const | Eff-N | Top-10 | AI-core | Largest |
|---|---|---|---|---|---|
| KOSPI (Korea) | 832 | 8 | 56.0% | 53% | SK Hynix 21% |
| Nifty 50 (India) | 50 | 24 | 52.9% | 0% | HDFC 10.6% |
| NASDAQ-100 | 101 | 34 | 45.3% | 57.5% | Nvidia 8.0% |
| EURO STOXX 50 | 54 | 27 | 43.2% | 18.1% | ASML 12.7% |
| MSCI EM | 1,170 | 27 | 40.4% | 34.6% | TSMC 14.7% |
| S&P 500 | 505 | 49 | 37.1% | 38.2% | Nvidia 7.8% |
| MSCI World | 1,270 | 95 | 26.6% | 29.7% | Nvidia 5.5% |
| TOPIX (Japan) | 1,650 | 111 | 22.3% | 4.8% | MUFG 3.3% |
Eff-N = how many equal-weight stocks the index really behaves like. AI-core = share of the whole index in the AI complex (chips + hyperscalers).
The reading — and its limits
Two exceptions prove the point — Japan (TOPIX) is genuinely diversified (behaves like 111 stocks, AI-core just 4.8%), and India (Nifty) is concentrated but in the old economy (AI-core 0% — its “tech” is IT services, the AI-disrupted category, not AI hardware). Everywhere else, buying the index increasingly means buying the same handful of AI names. KOSPI's eff-N is approximate (the uncapped Korea file is gated) but its top-2 of ~47% is exact.
Method Effective-N and concentration measures computed from each index's own full constituent file (index-owner or full issuer-holdings files). AI-core on a named-basket definition (semiconductors + hyperscalers) that crosses GICS sectors.
Computed 22 June 2026