US equities: the published research consensus leans constructive, with bullish conviction on the index and AI-exposed names — but real pockets of caution underneath.
Several recent research notes anchor the bull case on earnings, projecting roughly 20% EPS growth for US companies in 2026, driven by AI infrastructure spend, margin discipline, and broadening tech adoption. One camp frames further upside as contingent on earnings actually showing up — not multiple expansion — and flags rising bond yields and any AI capex slowdown as the key risks. A separate strand of analysis argues today's AI rally differs from the dot-com era because the revenues and profits are real. Another thread sees consumer discretionary names benefiting from resilient spending, equity-wealth effects, and cheaper gasoline into year-end.
Sub-splits matter. Researchers express the strongest conviction on semiconductors and AI infrastructure, citing persistent compute scarcity, with continued outperformance expected so long as high-yield credit spreads stay contained. The same voices turn bearish on legacy enterprise software incumbents, expecting AI-native challengers to compress their multiples. Housing-linked stocks face their own headwinds in published work — collapsing builder confidence, rising inventory, and visible price cuts from major builders.
On single names, a contrarian thread argues SpaceX's valuation is propped up by a tiny ~4% public float, and that expiring insider lockups through 2027 could force meaningful price discovery lower.
The main divergence: a conditional case for US value stocks if oil keeps falling — coexisting with, not contradicting, the AI bull thesis.