US equities are deeply split right now: published research is openly at war with itself on whether AI earnings momentum justifies current prices or whether the market is forming a cyclical top.
The bull case, as one camp of analysts frames it, leans on Micron's blowout quarter as evidence the AI build-out is intact. Researchers in this group argue roughly $3 trillion in AI capex through 2028 will rotate through memory, chip equipment, and data center construction, sustaining earnings. Some recently-published views also point to broader earnings resilience beyond mega-cap tech.
A contrarian thread of recent research pushes back hard. These analysts flag valuations near dot-com peaks, a record-high market cap to GDP ratio, and compressed risk premia, arguing semis and mega-cap tech are over-earning. Drawing on past memory cycles, they contend Micron's margins are unsustainable on a 1–3 year view.
The published research consensus is that the real disagreement is about timing, not direction: both sides agree memory is cyclical — they differ on when it rolls over. On the S&P 500, some see a constructive 2026; others warn of a lost decade.