The Market Is Betting on Two Things That Can't Both Be True

On Thursday, the S&P 500 hit a record high because a weak jobs report made investors think the Fed would hold off on rate hikes ^[1]. The same index carries Wall Street's most aggressive five-year earnings projections since 1995: 25.5% average annual growth, nearly four times the 6.5% historical average ^[2]. The market is betting on booming profits and on a deteriorating economy in the same trade. These two bets cannot both be right. Tobias Carlisle, an investment manager who tracks analyst projections, has described what happens to five-year estimates at extremes ^[2].

Analysts are notoriously bad at five-year earnings projections, and the estimates tend to be most wrong precisely when they're most extreme. — Tobias Carlisle

The 25.5% figure is not a prediction about corporate America. It is a measure of how much optimism is already priced into the index at its record high — and the optimism is priced on both sides of the trade simultaneously. The market is long a boom and long a slowdown in the same position. The earnings bet requires an economy expanding fast enough to generate profit growth at four times the long-run rate. The rate-hike-prevention bet requires an economy deteriorating enough to stay the Fed's hand. A single quarter of GDP cannot satisfy both demands, and the second quarter delivered 1.5% annualized growth ^[3] — too slow to support 25.5% earnings growth, too robust to force a Fed pause. The number fails both bets at once. There is one story that would make both bets right. It is the story Kevin Warsh was nominated to deliver. When Trump tapped Warsh to lead the Fed in March, the explicit theory was that AI-driven productivity is structurally disinflationary — a Greenspan 1990s replay in which technology gains keep inflation in check while growth accelerates, justifying both high equity valuations and accommodative monetary policy ^[4]. Warsh himself made the promise in explicit terms ^[4].

Our nation can see productivity boom like we did in the ’90s when we are not encumbered by a Federal Reserve which throws the brakes on — Scott Bessent

Warsh has since elaborated the analogy, explicitly linking AI to the 1990s IT productivity boom and invoking Greenspan's 1996-97 decision to hold rates steady while tech gains suppressed inflation ^[5]. The intellectual framework is coherent. If AI delivers a structural disinflationary expansion, the contradiction at the heart of the market resolves: earnings surge without triggering rate hikes, and the two bets become one. But Warsh is not holding rates steady. In June, he signaled a rate hike, with the FOMC projecting the funds rate rising to 3.8% by the end of 2026 as inflation hit 4.2% ^[6][7]. In July, he warned that persistent inflation prevents imminent rate cuts, triggering a global sell-off ^[8]. The escape hatch is delivering the opposite of what it was installed to provide. JPMorgan's economists pulled their rate hike forecast forward to December 2026 after Warsh's press conference, which they called the most troubling since the practice began in 2012, citing a lack of forward guidance and shifting inflation goals that damaged Fed credibility ^[9]. The same bank's CEO, Jamie Dimon, argued the same week that massive AI infrastructure spending will pay off as a strategic national investment ^[10].

play out and pay out — Jamie Dimon

The contradiction is no longer between two market bets. It is inside a single institution: JPMorgan's economists are bracing for the rate hikes that its CEO's AI thesis is supposed to prevent. The AI-productivity story is not impossible. It is simply not in the data yet. The U.S. personal saving rate dropped to 2.6% in April, one of the lowest readings since early 2008, while consumption grew only 2% over the past year ^[11]. Job cuts rose 38% year-over-year in April, driven in part by AI adoption itself ^[12]. GDP grew at 1.5%. A productivity boom that has not yet appeared in output, savings, or employment is a thesis, not a fact — and the market has already priced the fact. The bull case is real and it deserves its due. Deutsche Bank raised S&P 500 earnings forecasts to $358 for 2026 and $420 for 2027 after 87% of companies beat second-quarter estimates, with growth broadening across all eleven sectors ^[13]. U.S. corporations have reported nine consecutive quarters of net profit growth ^[14]. JPMorgan Private Bank's Kriti Gupta projects the S&P 500 to 8,200 by mid-2027, arguing that higher rates alone cannot end bull markets without a sharp deterioration in economic growth ^[15].

We're looking at double-digit returns again this year. — Kriti Gupta

The condition she does not currently see is already in the data. The saving rate at 2.6%, the job cuts up 38%, the GDP at 1.5% — these are not forecasts. They are reported numbers. The bull case explicitly depends on deterioration not materializing, and the deterioration signals are already present in the numbers the market is choosing to read as bullish. The market has bought the resolution before the evidence has arrived. It has priced the AI-productivity disinflation that lets Warsh cut rates while earnings surge, and it has done so on the basis of 1.5% GDP growth and a Fed chair signaling hikes. The contradiction does not resolve slowly. It resolves when the data forces a choice between the two bets, and the data is already forcing it. Either the productivity boom materializes in the numbers — soon — or one of the two bets has to give.