There’s no U.S. hyperscaler AI data center buildout without passing through Asia.
Given a resilient but soggy economy, structural rather than cyclical themes remain the most appealing for investors. Clearly, the investment and adoption of Artificial Intelligence (AI) remains one of the top themes. For the third year in a row, AI hyperscaler companies have contributed over half of U.S. equity’s double-digit returns. Additionally, the sectors outperforming the broad index are related to the AI infrastructure buildout: technology, utilities and industrials. As the third quarter earnings season showed, AI monetization is happening at some of the largest technology companies and the AI capex race shows no signs of slowing down. That might lead to a feeling of invincibility for U.S. AI-related companies. However, investors should remember that this is a global theme: U.S. tech supply chains are deeply interlinked with Asia and fierce competition and investment is accelerating in the region, especially China. Rather than picking “which country wins the AI race”, the better question to ask is “how to broaden exposure to the AI theme”.
Amidst an excellent year for international equities, EM Asia has been a standout up +31% in U.S. dollars, led by Korea +93%, Taiwan +39% and China +36%. Rather than a reflection of the region’s usual drivers (exports and Chinese economic growth), this has a lot to do with this year’s welcomed broadening of AI beneficiaries. Digging beneath the hood of Chinese markets makes this clear: H-shares technology +33% and Chinese ADRs (weighted heavily to internet names) +48%. In comparison, China’s domestic equity market of A-shares (with smaller technology representation) is underperforming, up “only” 19%, signaling muted enthusiasm for China’s domestic economic momentum.
Investing across Asia offers a variety of ways to capture the AI theme – at cheaper valuations than those in the U.S.:
- Interlinked tech supply chains: While attempts are being made to change this, now only 11% of global semiconductor production occurs in the U.S. Instead, 75% of it occurs in Asia1. Additionally, the global assembly, testing and packaging of semiconductors occurs in Southeast Asia, with Malaysia alone representing 13%. There’s no U.S. hyperscaler AI data center buildout without passing through Asia.
- China’s AI ecosystem: As DeepSeek showed, it’s risky to underestimate China’s AI innovation (despite U.S. export controls on advanced technology). China’s “edge” is on developing an open-sourced, cheaper AI ecosystem and “solutions based” AI applications. This is already driving usage within Chinese internet companies and across a variety of apps – and could become the foundation for several emerging markets. In China’s recently outlined Five-Year Plan (2026-2030), the core focus remains on “technological self-sufficiency and innovation”. This suggests that economic resources will be further channeled to advanced technology (hardware and software)2 and related sectors (advanced machinery, energy and materials).
- Asia’s lead in robotics and automation: A key way that AI can be transformative for the economy is with physical applications – such as robotics and broad automation. Asia’s lead in the space is large, with industrial robot density by far the highest in Korea, followed by China and Japan (and Germany). As populations continue to age, immigration becomes more restrictive, pressure builds to manufacture domestically and AI adoption becomes exponential, Asia’s lead in robotics is key for the profitability of global companies, including U.S. ones.