
GDP GROWTH · AI
65%
PEAK AI CAPEX
2026
MASS LAYOFFS
~0%
SLOWDOWN STARTS
2027
JPMorgan's chief U.S. economist Michael Feroli has declared 2026 the probable high-water mark for AI capital expenditure. In the bank's 2027 outlook, Feroli writes that AI-related spending growth was already expected to decelerate before the U.S. election, and that the slowdown will be modest rather than precipitous. The framing is reassuring — perhaps deliberately so.
Extraction Map
The numbers from Apollo Global's Torsten Sløk put the dependency in stark terms: AI spending has powered 65% of U.S. GDP growth in 2026. That is not a healthy diversification profile. When a single capital-expenditure cycle accounts for nearly two-thirds of aggregate growth, calling the peak is not an academic exercise — it is a macroeconomic stress test. Feroli's cushion thesis rests on two pillars: falling retail gasoline prices (as energy markets route around stalled Middle East peace talks) and a labor market that has remained resilient throughout the AI boom. Neither is guaranteed. Gas prices are a geopolitical variable, not a structural buffer, and the labor market's apparent strength may mask a subtler extraction — Feroli himself notes that the mere perception that AI destroys jobs is suppressing worker bargaining power, allowing firms to capture a larger share of profits. This is the quiet extraction buried in the reassurance. Workers are not being fired en masse, but the threat of AI replacement is functioning as a wage-suppression mechanism. Employers benefit from the narrative regardless of whether AI actually automates the jobs in question. The productivity gains flow upward; the anxiety flows downward. NYU's Aswath Damodaran had flagged a potential AI peak as early as August 2026, predicting further pain for AI equities. The tech sector has since experienced multiple sharp selloffs, most recently amplified by the Nvidia scandal involving CEO Jensen Huang's resignation. The confluence of peaking capex, executive turmoil at the sector's bellwether company, and unresolved questions about when AI spending translates into revenue creates a volatile setup for 2027. Feroli's core message — that the economy will muddle through because other sectors will pick up slack — deserves scrutiny. The 65% GDP contribution figure from Sløk suggests the slack-absorption thesis requires those other sectors to accelerate meaningfully, not merely persist. If AI capex falls even modestly while consumer spending fails to fully compensate, the arithmetic gets uncomfortable quickly. The structural question is whether the U.S. has built a one-engine economy. Feroli acknowledges this directly: "When an economy runs on a single engine, concerns about that engine's health are amplified." That is a risk disclosure dressed as analysis. The honest read is that nobody — not JPMorgan, not Apollo, not the Fed — knows whether the other engines can start fast enough if AI capex stalls harder than expected.
Worth reading?
SKIP
Score
28
/100
Extractive — AI capex concentrated value in a narrow tech-infrastructure corridor while suppressing wages through threat narratives; the deceleration socializes risk across the broader economy.
What Is GRIN?
GRIN stands for Generativity, Resilience, Innovation efficiency, and Novelty openness. These four dimensions capture the essential characteristics of any memetic intelligence — whether a nation, a corporation, a religion, or a movement.
About GRINGenerativity
2/5
Resilience
1/5
Efficiency
2/5
Novelty
1/5
Extractive
4/5
Inverted
Does this create new capability or just redistribute existing value?
AI capex created real infrastructure but 65% GDP dependence on one spending cycle is concentration, not broad-based generativity. The boom built data centers, not diversified productive capacity.
Verified Facts (7)
Michael Feroli is JPMorgan's chief U.S. economist
AI spending growth was expected to slow slightly in 2027 even before the election
Aswath Damodaran is a NYU finance professor who flagged AI peak as early as August 2026
Apollo Global's Torsten Sløk stated AI has powered 65% of GDP growth in 2026
Jensen Huang resigned as Nvidia CEO following a scandal
Feroli noted AI job destruction perception may be suppressing worker bargaining power
Feroli cited falling retail gasoline prices as a consumer offset factor
More latest news
Read all