Goldman's Warning: 75% of the S&P 500's Value Now Hangs on AI's Long-Term Payoff

Goldman's core point: AI demand is real, but market reliance on distant profits is fragile

About 75% of the S&P 500's equity value rests on profits more than a decade away, near a 25-year high. Goldman's warning is that this makes the market more sensitive than usual to shifts in long-term growth assumptions. When so much value sits in terminal assumptions rather than nearer-term cash flows, even a modest change in sentiment can move prices sharply.

That backdrop matters because the AI trade remains intact, at least based on recent megacap tech earnings. Strong results from a handful of leaders can reinforce confidence across the theme, but confidence is not the same as broad proof. If investors start to question how widely AI profits will spread, the market has less near-term earnings durability to fall back on.

The practical divide is no longer AI versus non-AI. It is between durable monetization and capex-heavy speculation. Goldman's filter is straightforward: be more cautious with names where operating earnings growth is under pressure and capex is debt-funded. Strong AI demand at the top of the stack does not make every spender deserving of the same valuation premium.

Why concentrated AI earnings are being mistaken for broad AI value creation

The market is making an expensive categorization error: treating concentrated AI earnings power as if it were broad AI value creation. AI is expected to drive roughly 40% of S&P 500 earnings growth this year, and the largest cloud companies are planning about $670 billion in 2026 spending. Recent megacap results also suggest the AI trade remains intact. Even so, that points to real demand among a small group of infrastructure leaders, not proof that the economics are spreading evenly across the market.

Investors are already being selective

This is where confirmation bias can distort pricing. Investors see strong results from a few platform leaders and then assume the entire AI chain is equally valuable. In practice, Goldman says investors have continued to favor AI platform stocks and productivity beneficiaries while rotating away from firms where operating earnings growth is under pressure and spending is debt-funded. That suggests the market is already distinguishing between companies capturing real AI economics and companies that are simply labeled AI-exposed.

Software remains the most exposed layer

The valuation risk is not evenly distributed. The software sector is down 17% so far this year, and Goldman estimates that every one-percentage-point decline in assumed long-term growth would cut combined enterprise value of S&P 500 companies by about 15%. High-growth stocks would be hit even harder. That helps explain why software is so vulnerable when long-dated growth assumptions come under pressure.

A useful takeaway is simple: AI demand can be real while AI economics remain narrowly distributed. The investment question is who captures the value, not whether AI exists.

How to position around AI: favor proof over narrative

The portfolio implication is not to abandon the theme, but to finance narratives more carefully. After a stretch in which investors rotated toward AI platform stocks and productivity beneficiaries and away from weaker earners, the market is already signaling which AI stories deserve a premium. The cleaner approach is to favor companies that can show current revenue linkage, better earnings conversion, and enough balance-sheet flexibility to keep investing.

What to watch next

The clearest signal is continued divergence. Over the next earnings cycle, watch three things:

• whether capex expectations stay broad or become concentrated in a few dominant platforms,
• whether more companies can connect spending to revenue in guidance, and
• whether financing plans remain disciplined rather than shifting toward weaker earners.

What could change the view

The setup improves if small-cap expansion starts to include real execution stories, not just thematic hope, and if semiconductor cap equipment remains a key scaling bottleneck. It weakens if the market decides even the current winners are still priced for too much future payoff after concerns about profits expected beyond the next decade. In that case, the safer move is to lean back toward companies with the clearest present-tense proof.