Artificial intelligence has become corporate America’s favorite buzzword — but the payoff hasn’t arrived yet, according to Goldman Sachs.
In the second quarter, a record 58% of S&P 500 companies mentioned AI on their earnings calls, Goldman analysts wrote in a Thursday note.
Companies are quick to tout new AI tools for customer support, software coding, and marketing, but “the share of companies quantifying the impact of AI on earnings today remains limited,” they wrote.
Few companies tied AI directly to profits, echoing a McKinsey survey showing more than 80% of firms said generative AI hasn’t meaningfully affected their bottom line.
Still, that hasn’t slowed the stock market’s love affair with AI.
Shares of AI-exposed companies are up 17% this year following a 32% surge last year, according to Goldman’s analysis.
The S&P 500 is now trading at one of its most expensive levels ever — though it’s still below the peaks of the dot-com bubble and the 2021 tech boom, the Goldman analysts said.
The 4 phases of AI trade
To make sense of where the market is at — and what could come next — the bank maps the AI trade into four stages.
Phase 1 was driven by Nvidia, the chipmaker powering AI models.
Phase 2, where markets are now, is powered by the hyperscalers: Amazon, Microsoft, Google, Meta, and Oracle. Together, they’re expected to spend $368 billion on capital projects in 2025, up from $239 billion in 2024 and $154 billion in 2023.
Related stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
That wave of investment has lifted semiconductor makers, power suppliers, and other infrastructure stocks.
However, the next two phases — AI-enabled revenues for software firms in Phase 3 and broad-based productivity gains across industries in Phase 4 — remain uncertain.
Phase 3, in which companies incorporate AI into products to boost sales, is trickier. Some investors worry AI could hurt software-as-a-service firms by reducing prices and lowering barriers for new rivals. Investors would likely wait for clear evidence of earnings gains before embracing these stocks.
“For AI-native companies to take share from SaaS companies, the AI product has to be meaningfully better and meaningfully cheaper than the incumbent, and SaaS companies continue to progress with their own AI-enabled products,” they wrote.
Phase 4 is the long-promised productivity boom.
Right now, Goldman says the US economy is still in the “early innings” of AI adoption. Usage is higher at large firms and in industries like information and finance.
The risk, Goldman cautions, is that the hype gets too far ahead of reality.
If AI spending reverted back to 2022 levels, Goldman estimates it would shave $1 trillion off 2026 sales forecasts — and wipe 15% to 20% off the S&P 500’s value.
