The U.S. and China are today’s leaders in AI. These countries could be the next-stage champions.

Bank of America says the next AI winners will be countries embedded in the AI build-out supply chains, expected to benefit over the next 1–3 years. It further argues that over 3–10+ years, economies that successfully adopt AI to raise productivity will gain a durable advantage. The note is positive on the opportunity set beyond the U.S. and China, but does not cite specific company or market-moving financial figures.

Analysis

The market is likely underpricing where the economic rent accrues in the AI stack: not to the broad set of countries that participate, but to the narrow set of chokepoints that cannot be easily substituted. Over the next 12-36 months, that argues for the highest-quality semicap equipment, advanced packaging, foundry, and memory bottlenecks to capture the margin pool, while assembly-heavy or low-value-added nodes see little incremental profitability.

The next leg is more about capital allocation than narrative. If enterprise AI spend stays above budget, the beneficiaries should be upstream industrials and infrastructure enablers before application software shows durable monetization; if spend stalls, the software premium deflates faster than the hardware complex because the latter is already tied to committed capex. BAC is at most a second-order beneficiary via financing and capital-markets activity, but the linkage is too indirect to justify a standalone factor bet.

Contrarian view: consensus is treating “AI winners” as a broad geographic basket, but equity returns will likely be far more concentrated than GDP effects. The market may also be too early on the productivity adoption leg; that is a 3-10 year story, while the next 1-3 years are about capacity scarcity and supply-chain bargaining power. The key falsifier is a capex pause, HBM oversupply, or tighter export controls that reroute volume without expanding demand; that would pressure the equipment complex within 1-2 quarters.

Key Decisions for Investors

• Go long SMH or SOXX on pullbacks over the next 1-3 months; best risk/reward is in the upstream semi complex, with a stop if capex guidance rolls over or export restrictions materially worsen.
• Pair trade: long ASML/AMAT vs short IGV over 6-12 months. Thesis is that scarce AI bottlenecks retain pricing power while application software faces delayed monetization and multiple compression.
• Build a modest long TSM position as the cleanest public-market expression of supply-chain concentration, but size for geopolitical risk; the thesis breaks if Taiwan risk premium widens sharply or foundry capex is cut.
• Do not initiate BAC on this headline; use it only as a watch item for spillover into loan growth and fee activity. The read-through is too indirect to carry standalone expected value.