AI-Driven US Equity Rotation

The thesis of AI-driven rotation from mega-cap to mid-cap equities is partially validated but requires significant nuance. While hyperscaler capex spending totaled roughly $400 billion in 2025, nearly 70% growth relative to 2024, and Goldman Sachs estimates hyperscaler capex will swell to $539 billion in 2026 and $629 billion in 2027, the investment reality is more complex than a simple rotation away from mega-caps. Hyperscaler capex for the "big five" (Amazon, Alphabet/Google, Microsoft, Meta/Facebook, Oracle) is forecast to exceed $600 bn in 2026, a 36% increase over 2025, with roughly 75%, or $450 bn, directly tied to AI infrastructure. This massive spending continues to concentrate wealth in mega-cap infrastructure builders. However, observable market rotation is occurring: small-cap companies' gains have reached 5.57% in the year to date while large caps have gained a mere 0.56%, with tech currently the worst-performing sector, losing 0.40% this year. The earnings gap between Magnificent Seven and the S&P 493 has contracted significantly—consensus expectations predict the gap in earnings growth between the Magnificent 7 and the S&P 493 will narrow from an estimated 30 percentage points in 2024, to 6 percentage points in 2025, and to 4 percentage points in 2026. Over a 5-year horizon, the thesis faces a critical bifurcation: mega-cap tech will continue benefiting from AI infrastructure capex and monetization, while mid-caps may outperform on valuations and earnings growth acceleration, but true "rotation" depends on whether mega-cap AI investments deliver sufficient returns to justify elevated valuations and whether broader economy benefits materialize. Valuation dynamics sharply contradict pure rotation thesis. The forward 12-month P/E ratio for the S&P 500 is 21.5, which is above the 5-year average (20.0) and above the 10-year average (18.8), indicating broad market premium rather than mega-cap excess. More critically, S&P 500 ex the Great 8's P/E on 2025 consensus EPS is 21.1x, with the Great 8 at 35.1x, and overall S&P PE at 24.7x. Small caps are trading at a forward P/E of 18x versus the S&P P/E of 24x, with this 22% discount near a historic low. For much of 2024, the S&P 500 traded at a forward P/E of 22x–26x, while mid-caps lingered at 16.5x, creating a valuation arbitrage that became too significant for institutional investors to ignore once the Fed's rate-cutting cycle provided the necessary catalyst. This valuation gap provides upside for mid-caps, but while earnings continue to weigh in favor of the Magnificent 7, macro factors such as growth and trade policy lean towards the S&P 493, with economists' expectation of a steady and above-trend pace of US growth in 2025 favoring the performance of the S&P 493, which is more sensitive to changes in growth compared with the Magnificent 7. The 5-year outlook hinges on whether mid-cap earnings growth acceleration is sustainable or temporary, and whether mega-cap ROI from AI capex validates current valuations. Institutional consensus reflects cautious bullishness on both mega-cap and mid-cap exposure, not forced rotation. Three main factors support ongoing strength within tech: earnings momentum, current valuations, and a supportive macro environment, with valuations high but justifiable among mega-cap names given their structurally higher profitability. Conversely, institutional observers expect 2026 to mark a shift toward broader market leadership as valuations normalize, creating opportunities beyond mega-cap growth. Goldman Sachs analysts expect investor interest in AI to transition from AI infrastructure to a broader AI "Phase 3" of application rollout and monetization, referring to companies that are likely to see AI-enabled revenues beyond those that build the infrastructure underlying AI, including software and services firms which offer longer-lasting, secular growth. This suggests mid-cap beneficiaries are concentrated in software/services with AI applications, not broad-based mid-cap rotation. Over 5 years, the thesis succeeds if: (1) AI infrastructure capex sustains cumulative multi-trillion deployment, (2) monetization unlocks via enterprise software and services adoption (where mid-caps dominate), and (3) valuations remain anchored despite uncertainty. The thesis fails if: (1) AI ROI disappoints, causing mega-cap capex cuts and multiple compression, or (2) returns on AI investment cannot support mega-cap valuations and broader economic acceleration proves transitory.