The recent surge in U.S. real GDP growth, beating consensus by 1.5 percentage points in Q3 2025, the largest positive surprise in a decade, has sparked debate about its sustainability. Some have said things such as, "But if people aren't working then they won't spend, if they don't spend then we will have recession," worry that AI-driven efficiency could lead to widespread unemployment, reduced consumer spending, and economic downturn.
While this concern is valid on the surface, it overlooks the nuances of our increasingly bifurcated economy, where productivity gains from AI are creating a resilient structure that favors asset owners and high-earners, preventing a broad recession.
The Productivity Boom Is Real—and It's Not Labor-Driven
The Q3 GDP acceleration was almost entirely fueled by productivity, not increased labor hours or employment. Output rose sharply while unit labor costs remained contained, allowing corporate margins to expand faster than inflationary pressures. This is the hallmark of an economy becoming structurally more efficient: companies are doing more with fewer workers, thanks to agentic AI tools and automation. As Ed Yardeni notes, hyperscaler AI spending levels are "appropriate and sustainable," stimulating broader business activity through demand for GPUs, servers, networking, data-center power, and cooling infrastructure. This creates a virtuous cycle that boosts revenues without proportional cost hikes, widening operating leverage and raising earnings ceilings.In this environment, market leadership has rotated from long-duration growth stocks to cyclicals, value, and small-caps—sectors that historically thrive during genuine economic acceleration without overheating.
Addressing the Job Loss Concern: A Bifurcated Economy Absorbs the Shock
The fear—that AI-induced unemployment will slash spending and trigger recession—assumes a uniform impact across society. But our economy is deeply bifurcated: the upper class, including much of the former middle class elevated by property and asset holdings, increasingly determines economic direction. These groups own the majority of stocks, real estate, and other appreciating assets, which continue to benefit from AI productivity.
- Asset-driven wealth cushions the blow: Homeowners and investors (often the same people) have seen massive gains from rising property values and stock portfolios. This "wealth effect" sustains spending even if wages stagnate for lower earners. The middle class has effectively become upper-middle through these holdings, providing a buffer against job displacement.
- Consumption is concentrated: In the U.S., the top 20% of earners account for ~40% of consumer spending. As AI boosts corporate profits and asset values, this cohort's purchasing power grows, offsetting reduced spending from displaced workers. The labor market's shift (from 2-to-1 jobs-to-unemployed ratio in 2022 to one-to-one now, with hiring plans at historic lows) hasn't yet caused a spending collapse, precisely because high-earners and asset owners keep the economy humming.
- No broad recession signal: Initial jobless claims spiked to 231k recently, but overall data shows contained risks. AI efficiency means output grows without labor inflation, keeping the Fed patient and markets supported.
Caveats: Adoption Must Accelerate, Bottlenecks Avoided
The bullish case isn't without risks. While AI will eventually create more jobs than it displaces, job displacement that is too rapid, the rising inequality of ability, and potential demand weakness if productivity gains remain concentrated at the top are real risks. The "no broad recession" conclusion is plausible in the short-to-medium term but not guaranteed if AI adoption outpaces job creation/retraining or if bottlenecks slow the productivity boom.
AI's productivity gains depend on rapid enterprise and government adoption. Slowdowns from integration challenges, skill shortages, regulatory hurdles, or ROI doubts could moderate the boom. Key bottlenecks: power availability, grid capacity, high-end GPU supply, advanced chip production, cooling, data-center timelines, and skilled labor—must be resolved. If they persist, the efficiency surge could decelerate, making the recession fears more plausible.
Yet, as long as AI spending remains stimulative and these hurdles are overcome, the bifurcated economy—where asset-rich upper classes drive growth and stock buying—should prevent a downturn. The Nasdaq's recent choppiness reflects near-term overspend worries, but the structural tailwinds point to new highs ahead.
In short, job losses from AI won't trigger mass recession because spending and markets are increasingly decoupled from broad employment. The economy's winners, bolstered by assets and productivity, will carry the load.