Apollo chief economist says there’s ‘zero evidence’ AI is killing jobs—in fact, he says it’s creating them

The article argues that there is "zero evidence" AI is causing broad job losses, citing ADP data and Torsten Sløk’s view that AI spending is instead stoking employment and inflation. It contrasts that with corporate layoffs and workforce reductions at Block, Klarna, and Amazon, where executives have tied cuts to AI-driven efficiency gains. Overall, the piece centers on the debate over whether AI is truly displacing labor or companies are using AI as cover for broader restructuring.

Analysis

The market is likely misclassifying this as a labor-demand bear story when the nearer-term effect is a capex-and-services bull story. If management teams believe AI can preserve output with fewer layers, the first beneficiaries are not generic software vendors but the picks-and-shovels stack: data-center REITs, power equipment, networking, semis, and implementation consultancies. That supports the modestly positive read for ADP and Apollo-linked capital allocation narratives, while creating a near-term negative bias for labor-sensitive enterprise software and staffing proxies.

The important second-order effect is inflation persistence, not immediate unemployment. AI adoption at scale is disinflationary only after a multi-quarter buildout phase; before that, it pulls forward demand for chips, energy, and specialized labor, which can keep services and capex inflation sticky. That matters for JPM because a hotter-for-longer rate path would offset any margin gains from internal automation and can keep credit costs from falling as quickly as the market expects.

The consensus is probably overestimating how quickly headcount savings show up in P&L and underestimating how quickly firms overstate AI-driven layoffs to justify restructuring. In practice, the next 6-12 months should feature more hiring of implementation, compliance, and workflow redesign talent than net corporate headcount destruction. KLAR and APOS look like beneficiaries of that transition, but the bigger long-only expression is through the infrastructure layer rather than the application layer.

The trade setup favors owning the physical buildout and shorting labor-displacement narratives. The risk is a regulatory or macro shock that slows AI capex before it translates into earnings, which would compress multiples in the infrastructure cohort first. For the next quarter, the catalyst to watch is payroll and inflation prints; if both stay firm, the market will be forced to reprice AI as an input-cost boom rather than a labor shock.

Key Decisions for Investors

• Long APOS vs short a labor-sensitive software basket for 1-3 months: the market is likely underpricing the duration of AI capex and consulting demand while overpricing immediate headcount destruction.
• Add ADP on weakness over the next 2-4 weeks: payroll data strength implies resilient employment demand and better pass-through to service pricing, with limited downside unless labor prints roll over sharply.
• Long AMZN only tactically, not structurally, into the next earnings cycle: AI should improve fulfillment/corporate productivity, but the stock can lag if investors focus on margin expansion delaying rather than accelerating growth.
• Avoid extrapolating JPM margin upside from internal AI automation; consider a hedged position if rates remain sticky: long JPM against a regional bank basket only if credit data improves, otherwise AI-driven efficiency can be offset by slower loan growth.
• If you want the cleanest expression, buy the AI buildout winners on pullbacks over 1-2 weeks: semis, networking, and data-center infrastructure outperform when the narrative shifts from job loss to energy and capex demand.