Gen Z is losing the most in the AI economy—and Goldman warns it’s about to get worse

Goldman Sachs says AI is still eliminating about 11,000 net U.S. jobs per month, with 21,900 AI-attributed layoffs in April alone, even as data center construction has added 212,000 jobs since 2022 and is now creating roughly 9,000 positions a month. The article highlights a growing generational divide, with AI adoption showing a slight positive correlation to unemployment among workers under 30, while broader payroll growth remains resilient. Census-based adoption is only 19.5% of U.S. establishments, but the labor mix suggests ongoing occupational churn rather than a clean unemployment shock.

Analysis

The key market read-through is not that AI displacement is slowing; it is that the labor pain is being geographically and sectorally re-allocated from office labor to capex-intensive physical buildouts. That matters because the construction offset is cyclical and front-loaded, while the displacement in entry-level knowledge work is sticky and compounding. In other words, the headline job count can look benign right up until the temporary infrastructure phase rolls over, at which point the economy loses both the buildout jobs and the onboarding pipeline for young white-collar workers.

The more important second-order effect is margin structure. Companies adopting AI are likely to keep headcount flat-to-down while extracting productivity gains, which supports earnings near term even if hiring softens. But the beneficiaries skew toward incumbents with distribution, data, and workflow lock-in; pure labor-arb businesses and staffing intermediaries face an asymmetric risk because AI can compress billable hours faster than it creates new categories of demand.

For markets, the generational unemployment signal is an early-cycle warning rather than a macro recession trigger. If under-30 unemployment continues to diverge, expect a lagged hit to discretionary spending categories tied to first-job income formation: rent-sharing demand, entry-level autos, fast fashion, quick-service, and consumer credit utilization. Conversely, industrial electrification, data-center power, cooling, and grid equipment names have a second leg if the buildout broadens into manufacturing and utilities.

The consensus may be underpricing how temporary the construction offset is and overpricing how linear AI adoption will be. The real risk is that by the time labor weakness shows up in aggregate payrolls, the market has already normalized the “manageable churn” narrative. If the next 1-2 earnings seasons bring more explicit AI headcount actions without a matching pickup in greenfield hiring, the trade shifts from AI-as-productivity to AI-as-demand-destruction, especially for small-cap labor-intensive services.

Key Decisions for Investors

• Short UPWK or other labor-arbitrage / gig-marketplace exposure on a 3-6 month view; thesis is margin compression from AI replacing low-end digital tasks before new demand offsets it. Use a tight stop if management shows accelerating enterprise spend tied to AI services instead of substitution.
• Long VRT / ETN / PWR as a basket against short KBE or a staffing proxy over 6-12 months; the better risk/reward is in infrastructure picks-and-shovels that monetize the buildout even if office hiring weakens. Expect a cleaner earnings path than labor-exposed cyclicals.
• Buy 3-6 month put spreads on manning/staffing names or HR platforms with high exposure to entry-level hiring cycles; target a 2:1 to 3:1 payoff if AI-driven hiring freezes broaden into tech, ops, and back office.
• Pair long MSFT or AMZN against short a basket of junior software/service enablers over 12 months; the productivity uplift accrues to platform owners, while weaker vendors face slower seat growth and tougher renewals. Prefer call spreads on the longs to reduce multiple-risk.
• Watch for a tactical long in consumer credit-sensitive names only after the under-30 unemployment signal stabilizes; until then, avoid adding to discretionary exposure that depends on first-job wage growth, as this lagged effect can bite 2-3 quarters after the labor headlines.