US equity research is unusually split right now — and the fault line is AI.
The bull camp, well represented in published research this week, argues the S&P 500 can keep grinding higher. One widely-cited year-end target sits at 8,100, with a separate fair-value model pegging it near 8,300. The case rests on the AI capex supercycle, resilient earnings in a non-recessionary economy, and abundant global liquidity running into late 2026. Several strategists prefer stocks over bonds and stay constructive on Nvidia, the US hyperscalers, and AI compute hardware, citing supply constraints that extend the cycle.
The bear camp, drawing on the same data, calls valuations "absurd" — above dot-com and 2008 peaks — and flags that AI and semis now make up roughly 19% of the index, a concentration historically tied to weak ten-year forward returns. Other cautious voices warn that mega-cap free cash flow is turning negative, that high long-dated yields pressure leveraged tech, and that a "compute glut" echo of the fibre overbuild is possible. A recurring contrarian tilt: large-cap value over mega-cap growth, plus industrials and commodity-linked names as diversifiers.