The 2026 Goldilocks Trade Rests on One Number

The constructive case for 2026 is no longer a contrarian whisper. AI capex, fiscal accommodation and a labour market that is loosening without breaking have produced the rare configuration where growth, disinflation and earnings expansion coexist. The question is not whether the setup is real — Q1 prints already confirm it — but whether the consensus has now priced enough of it that the marginal surprise must come from disappointment. The Desk's view: the bull case survives one oil shock, not two.

Start with the earnings spine, because everything else depends on it. Aggregate S&P 500 earnings rose 19 percent year-on-year in the first quarter, and FactSet's analyst panel projects 24 percent growth over the next twelve months. That is not a late-cycle number. It is a post-recession recovery number, arriving in the eighth year of an expansion. The composition matters: Nvidia is forecast to grow earnings by nearly 80 percent in 2026, Broadcom roughly the same, and the second-tier semiconductor names — Micron, Sandisk — are projected at multiples that only make sense if AI infrastructure spend continues at current pace through the year.

The valuation objection is real but narrower than headline P/E figures suggest. Maverick Equity Research notes the index trades at roughly 22 times forward earnings — outside the historical interquartile range and pricey on every conventional screen. Strip out the Magnificent Seven and the multiple compresses to 19, still elevated but no longer extraordinary. The 2025 rally, importantly, was earnings-driven rather than multiple-driven: margins expanded, revenue grew 8 percent, and net profit margins reached 13 percent. That is a different animal from the multiple expansion that defined 2020-2021 and ended badly. A market paying 22 times for 24 percent earnings growth is not a bubble. A market paying 22 times for earnings that disappoint by even 500 basis points is.

Where the trade breaks

The asymmetry is what makes 2026 interesting rather than comfortable. Mining and metals profits grew nearly 90 percent in Q1, suggesting commodity-linked inflation is already feeding through corporate income statements rather than being absorbed at the margin. If oil holds above current levels into Q3, the disinflation thesis that underwrites the Fed's easing path weakens, the steeper curve flattens, and the small-cap rotation that depends on lower funding costs stalls. The Desk puts the probability of S&P 500 earnings hitting the 24 percent consensus at roughly 55 percent — better than a coin flip, worse than the index price implies. The path to 7,400 exists. It is narrower than the deck suggests.

The wider point, which Surowiecki has pressed and which the data supports, is that pricing power has shifted decisively to incumbents. Inflation and consolidation have given large firms the ability to defend margins through a cost cycle that would have crushed them a decade ago. AI is accelerating the divergence by lowering marginal labour cost in precisely the functions — software, customer operations, content — where the Magnificent Seven concentrate revenue. That is why the gap between investor sentiment and household sentiment keeps widening, and why it can keep widening longer than feels comfortable. The trade is not that everyone is wrong about the economy. It is that the listed equity complex has structurally decoupled from the median consumer, and 2026 is the year that decoupling becomes the base case rather than the anomaly.