AI’s boost to the U.S. economy is nearing its peak, Goldman says, as Apollo finds ‘zero financial payoff’ outside Big Tech

Goldman Sachs estimates the AI-related wealth effect added 0.4 percentage points to U.S. consumer spending growth over the past year, but expects that boost to fade; it also sees the AI growth impulse peaking in late 2026 or early 2027. Apollo’s Torsten Slok says AI has delivered no financial payoff to corporate America outside tech, while Goldman’s Jan Hatzius expects an eventual productivity and profit payoff but warns near-term labor-market disruption could weigh on growth. Hatzius estimates developed-market long-term rates have risen about 100 basis points since early 2026, potentially subtracting a cumulative 0.6% from GDP if the increase persists.

Analysis

The key risk is not simply weaker AI spending; it is a two-sided earnings squeeze. If infrastructure outlays keep rising while non-tech firms fail to monetize adoption, capital intensity and token costs pressure margins before productivity gains arrive. If spending then slows, the first-order hit shifts to chip, networking, data-center and power supply-chain demand, while the expected broad-market earnings diffusion remains absent. This makes the AI trade increasingly dependent on a narrow set of capex buyers and beneficiaries rather than economywide profit growth.

Over the next 1–3 months, higher fuel costs and long rates can compound that risk: energy-sensitive consumers lose discretionary capacity, and duration-sensitive equities face valuation pressure. A sustained oil shock or further yield increases would also reduce the room for the Fed to cushion weaker growth. Over 6–18 months, labor substitution may lift margins at adopters but weaken household income and demand; productivity upside is therefore not automatically bullish for aggregate earnings.

The contrarian point: weak measured payoff today does not establish that AI investment has failed. Adoption and workflow redesign lag model access, and cheaper inference could widen usage. But the market may be over-crediting eventual monetization while underpricing the transition’s costs. Avoid treating AI capex as a broad-market growth engine until non-tech revenue-per-worker, margins or productivity data improve.

Key Decisions for Investors

• Consider a modest XLE-versus-XLY relative-value position over the next 1–3 months as an energy-price hedge against consumer margin pressure. Keep sizing limited: unwind if oil and refined-product prices retreat materially on de-escalation, or if consumer discretionary earnings revisions stabilize while energy estimates roll over.
• Avoid adding broad exposure to AI infrastructure solely on capex growth. Track hyperscaler spending plans alongside supplier order trends and non-tech adoption economics; reduce exposure if capex guidance slows before evidence of monetization broadens.
• Favor balance-sheet quality over long-duration, cash-flow-distant growth if developed-market yields continue rising. Falsify this stance if yields stabilize and non-tech earnings revisions improve; the article supplies no valuation or positioning data to justify an outright short.
• Set a 1–3 month macro alert: persistent fuel-price strength plus further yield increases would raise downside risk to consumer demand and capital spending. Reassess if energy prices normalize or broad non-tech productivity and margin measures turn sustainably higher.