正是这种集中度放大了新兴市场一贯的风险。此类指数覆盖国家范围非常广，容易引发“是否应将巴西与波兰放在同一篮子”这类质疑，也伴随对中国权重偏高的担忧。以美元计算，近15年中新兴市场仅有5次跑赢 MSCI 发达市场指数，其中仅2次幅度足以让美国投资者调整配置。上年新兴市场指数上涨31%，比发达市场高12个百分点。
当前的主要催化是疲软美元与AI、电动车硬体需求。由亚洲主导的AI基础设施供应链，包括台湾 Quanta Computer、晶片企业及资料中心设备，若美元维持窄幅整理甚至下行，受益机会较高。电池环节也受追捧，关键企业有中国 CATL、BYD，以及韩国 LG Energy Solution；Delta Electronics 今年上涨137%。但估值分化明显：根据 S&P Capital IQ，只有 Delta 与 BYD 目前估值倍数高于长期均值；三星电子与 SK Hynix 则仍低于历史水平，分别约8倍对比13倍、6倍对比22倍。文章核心观点是，AI“pick and shovel”供应商可能比应用与品牌端更耐久，因为他们能在每一轮竞争中持续赚取需求。
MSCI’s emerging markets index hit a fresh high this week, and its 17% rise since January has been unusually concentrated: Samsung Electronics (+122%), SK Hynix (+146%) and TSMC (+48%) delivered roughly half the gain. Samsung Electronics has crossed $1tn in market capitalisation, and TSMC is close to twice that level. In other words, the broad “emerging markets” advance is being driven largely by a narrow set of Asian chip and hardware champions.
That concentration makes traditional emerging-market risks more visible. The index covers a very wide range of countries, raising worries about whether Brazil and Poland should be grouped together and about concentration in China. In dollar terms, emerging markets have beaten MSCI’s developed-world index only five times in the last 15 years, and only two of those occasions were large enough for many U.S. investors to reconsider allocations. In the last year, EM gained 31%, 12 percentage points ahead of developed markets.
The immediate tailwind is a weak U.S. dollar and demand for AI and EV hardware. Asian-led AI infrastructure chains, including Taiwan’s Quanta Computer, chip firms, and data-center hardware, stand to benefit if the greenback remains range-bound or weakens. EV battery names such as CATL, BYD, and Korea’s LG Energy Solution also draw attention, while Delta Electronics is up 137% year to date. Valuation support is uneven: according to S&P Capital IQ, only Delta and BYD trade above their long-term average multiples, while Samsung and SK Hynix remain below history, around 8x versus 13x and 6x versus 22x. The article’s key claim is that AI “pick-and-shovel” suppliers may be more durable than app and brand players, since they can earn from every winner in the hardware race.