The grip of AI-related technology stocks on the S&P 500 is loosening. The top 10 stocks, which accounted for 42 percent of index market capitalization in Q3 2025, fell to 38 percent by mid-March 2026—a 400-basis-point decline that redistributes roughly $175 billion in relative weight across 490 remaining constituents.
This concentration reversal reflects a structural reassessment, not a tactical rotation. Hyperscalers are deploying rare capital into AI infrastructure—data centers, power grids, compute capacity—while generating no near-term cash return. The balance-sheet impact is dramatic: firms that historically operated as asset-light software businesses now carry significant physical plant, dampening free cash flow and stretching capital efficiency.
The vulnerability is real. Torsten Slok, chief economist at Apollo Global Management, has flagged leverage embedded in AI valuations as a tail risk. If momentum reverses, the same leverage that amplified gains could trigger rapid deleveraging. Neil Dutta, head of economic research at Renaissance Macro Research, noted on May 7 that 15 non-tech companies with combined market capitalization of $2 trillion are tracking AI capex flows as a leading indicator—meaning traditional sectors are tethered to technology spending even as they appear to "outperform."
Jim Paulsen, an economist, identified the critical stress point: market stability has historically required old-economy sectors (banking, manufacturing) to move in tandem with tech. A persistent divergence between the two would signal underlying fragility.
European Central Bank economists have added a longer-term warning: historical patterns suggest valuations are vulnerable even if AI delivers on its technical promise. Current prices reflect AI's significant power; they do not yet price in the execution risk of turning multi-trillion-dollar capex into positive returns.
The 400-basis-point shift in concentration is material but does not prove durability. Previous tech-to-value rotations in 2022 and early 2023 reversed within quarters. The sustainability of this rebalancing depends on whether non-AI sectors can generate independent earnings growth or whether they remain derivative of hyperscaler spending.
