The case for global diversification has long rested on a simple premise: different markets move to different drummers. That premise is being stress-tested. In a June 2026 paper published by Principal Asset Management, Market Strategist Magdalena Ocampo argues that the convergence of artificial intelligence infrastructure spending and global energy dynamics is compressing the very regional differentiation that diversification depends on — and advisors need to reckon with it.
The Concentration Problem Isn't Going Away
The starting point is the S&P 500. Technology now accounts for nearly 40% of the index, a level of concentration that leaves investors acutely exposed to any softening in hyperscaler capital spending. That risk isn't just domestic. As Ocampo notes, "given the global nature of tech supply chains, any slowdown in U.S. AI investment would likely have meaningful spillovers into regions whose earnings are closely tied to the AI cycle." The implication is pointed: owning international equities as an offset to U.S. tech exposure may provide less protection than the geographic labels suggest.
Last year, diversification worked in a specific and legible way. U.S. policy disruption triggered a rotation toward international markets with supportive domestic policies and improving growth outlooks. This year, the dynamic reversed. An energy shock reinforced the relative resilience of the U.S. economy and its structural advantages in AI, pulling capital back home. Two consecutive years, two different regime-level forces — and in both cases, diversification required active judgment, not passive allocation.
Korea and Taiwan: Deeply Embedded in the AI Supply Chain
South Korea and Taiwan offer a striking case study in the limits of geographic diversification. Both economies are among the most exposed to Middle East conflict given their reliance on crude oil and natural gas transiting the Strait of Hormuz. Yet the macro impact was less severe than feared — markets recovered sharply from the March selloffs, with MSCI South Korea and Taiwan equity indices returning roughly 117% and 63% year-to-date in USD terms as of June 1, 2026.
The explanation is structural. Decades of U.S. semiconductor offshoring have positioned Korea and Taiwan at the centre of the global AI buildout. Multi-year agreements with U.S. hyperscalers, combined with the outsized share of key chip producers' revenues tied to U.S. demand, have created what Ocampo describes as "a deep symbiotic relationship between U.S. tech leaders and their Asian suppliers." With U.S. hyperscaler capex forecast to reach $700 billion in 2026 and exceed $1 trillion in 2027, the semiconductor demand underpinning these economies is structurally supported for the near term.
But the diversification trade-off is explicit. Korea and Taiwan now account for nearly half of the MSCI Emerging Markets Index — roughly double their 2020 weight. The EM index, historically a source of distinct return drivers, has become increasingly correlated with the U.S. AI cycle. Owning broad EM exposure today is, to a meaningful degree, owning the same AI thesis through a different ticker.
Europe: The Energy Sensitivity Drag
Europe enters this analysis as the historical counter to U.S. tech dominance — lower technology exposure, greater sensitivity to macro-fiscal dynamics. That story worked well in 2025, when a pivot from fiscal austerity to expansion drove equity outperformance and genuine diversification value.
The energy shock has changed the calculus. As a net energy importer, Europe is structurally vulnerable to supply disruptions flowing from U.S.-Iran tensions, and that vulnerability has been on display. Since late February, European equities have underperformed U.S. markets and remain below pre-conflict levels. In Ocampo's framing, "Europe has not been perceived as an attractive diversifier from the U.S. AI cycle in recent months."
The path to restoring Europe's diversification appeal runs through geopolitical resolution. Recent reports suggesting a potential U.S.-Iran deal could stabilize energy markets and reactivate interest in the region, particularly given longer-term tailwinds in infrastructure and defense spending. For now, however, Europe's near-term role as a diversifier is constrained.
China: The Outlier Case
The most differentiated diversification signal in Ocampo's analysis comes from an unexpected but logical source: China. Unlike Korea, Taiwan, and Europe, China's equity market has managed to largely sidestep both the AI-cycle dependency and the energy shock sensitivity that are compressing diversification elsewhere.
On the energy side, China's long-term effort to diversify its energy mix — through renewables deployment alongside expanded coal-fired capacity — has allowed it to build crude oil reserves and reduce its exposure to global supply disruptions. During the March selloff, Chinese equities posted only a modest decline broadly in line with the S&P 500's pullback, a sign of relative resilience.
On the technology side, China's strategic push to build out domestic semiconductor supply chains is fostering what Ocampo characterizes as "a more independent technology ecosystem." DeepSeek's recent training of its latest model on domestically produced semiconductors is the most visible illustration of this trajectory — a deliberate pivot away from reliance on U.S. advanced chips.
The result is a portfolio characteristic that is increasingly rare. "At a time when global markets are increasingly driven by common macro factors," Ocampo writes, "China's relative independence from both energy shocks and the U.S. AI cycle makes it one of the few major markets offering more distinct diversification benefits." The added consideration: Chinese valuations remain relatively attractive, amplifying the case.
The Structural Shift in Diversification
Ocampo's bottom line is clear-eyed: "as energy and technology supply chains become more globally interconnected, regional exposures are increasingly shaped by the same underlying forces, making portfolio diversification harder to achieve in practice." The familiar tools of geographic allocation are providing less insulation than their labels imply.
This isn't an argument against diversification — it's an argument for rethinking how it's measured and sourced. Traditional regional weightings may obscure rather than reveal underlying factor exposures. A portfolio split between U.S. equities, EM, and European stocks might still be substantially a single bet on U.S. AI capex and global energy stability.
Key Takeaways for Advisors and Investors
1. Look through the geographic label. Country allocation alone no longer describes risk. Korea, Taiwan, and U.S. tech are increasingly expressions of the same AI infrastructure thesis. EM exposure requires scrutiny of what is actually driving returns.
2. U.S. tech concentration is a portfolio-wide risk. At nearly 40% of the S&P 500, technology concentration is not just a domestic equity concern — it bleeds into international portfolios through supply chain linkages.
3. Europe's diversification case is geopolitically contingent. A resolution to U.S.-Iran tensions could restore Europe's appeal meaningfully, particularly given defense and infrastructure tailwinds. Monitor energy market signals as a leading indicator.
4. China warrants a fresh look for genuine diversification. Its structural independence from both the U.S. AI cycle and global energy shocks — combined with attractive valuations — makes it one of the few major markets offering a differentiated return driver. The geopolitical risk premium is real, but so is the diversification benefit.
5. Reassess diversification at the factor level, not just the regional level. Shared exposure to AI and energy is the organizing risk of the current environment. Advisors should map portfolios against these underlying forces rather than relying on standard geographic splits to deliver true diversification.
Footnote:
1 ”https://www.principalam.com/us/insights/equities/rethinking-diversification-ai-driven-world." 14 June 2026.
Footnote:
1 "Rethinking diversification in an AI-driven world | Principal Asset Management." 14 June 2026,
