Goldman Sachs Says Europe Is the AI Trade's Next Winner

By Olivia Hart

The Stoxx 600 returned 2% on average across the eight biggest short-term corrections in the AI trade since 2023, while the S&P 500 fell 4% and a broader AI basket dropped 19%. Those numbers come from a Goldman Sachs global strategy paper published Monday, and they anchor the bank's new argument: Europe, the region that missed most of the AI rally, is now positioned to outperform.

The note, lead-authored by Guillaume Jaisson, makes a contrarian case built on exposure arithmetic. The S&P 500 is three times as sensitive to the "AI factor" as the European benchmark. AI infrastructure stocks make up only 9% of the Stoxx 600 and contributed just two percentage points of the index's forecast 9% earnings-per-share growth for 2027. In plain words, Europe has less to lose when the AI trade wobbles — and October's volatility, extending into the first week of November, has already rewarded that insulation.

Why does underperformance suddenly look attractive?

Goldman's core claim is that the factors behind Europe's lag are turning into advantages. The region can still give investors exposure to the AI theme, the bank says, through the "picks-and-shovels" hardware layer and through eventual productivity gains.

The report positions Europe for what Jaisson and his team call the second phase of the AI revolution. The region holds computing and semiconductor manufacturers, electrical and cooling-equipment suppliers, data-center and physical-infrastructure companies, and electricity providers. Earnings in Europe's AI infrastructure sector are set to improve by 20% in 2027, yet the sector's price-to-earnings multiple has fallen from 23 to 18. That combination — rising earnings, compressed valuation — is the cheap entry point the argument rests on.

The growth data backs some of this up. These companies grew around 8% per year on average before ChatGPT arrived; since then the pace has accelerated to 12%. "AI capex growth may be peaking," Goldman said, "but the spending is not." The caveat is worth weighing: peaking capex growth has historically been the moment when late-cycle arguments sound most persuasive.

What about the third phase — and the biggest prize?

The third-phase beneficiaries are companies whose revenues are enabled by AI, such as software and IT services. Here the news is worse for existing holders. Concerns about whether business moats survive the new AI era have hit the sector on both sides of the Atlantic, and the price-to-earnings multiple has compressed from almost 25 times to 17. Goldman is effectively telling investors that the market has already punished this cohort, without saying precisely when the derating ends.

The least-proven opportunity is also, in the bank's framing, potentially the largest: productivity. Europe carries long-standing burdens — labor shortages, rigid labor markets, overregulation. Because it lags the U.S. in AI adoption, automating tasks could give Europe more room to improve productivity economywide. Goldman calculates that a third of the Stoxx 600's aggregate wage bill could be automated, which could lift earnings by as much as 24% if implemented successfully. That "if" is doing heavy lifting; the figure is a ceiling, not a forecast.

Which stocks does Goldman name?

The report sorts its picks by phase. By market capitalization, Europe's largest AI infrastructure plays include:

• ASML
• Siemens
• Iberdrola
• Vinci

In the revenue-generation cohort, Goldman points to Prosus, RELX and SAP. For long-term productivity benefits, the bank suggests banks Barclays and Deutsche, chemicals group Croda, technology names Capgemini and Sage, and insurers Zurich and Admiral.

The spread matters as much as the names. A list spanning semiconductor tools, Spanish utilities, Dutch internet holdings and British insurers is a bet on the AI theme diffusing through the whole economy rather than concentrating in a handful of U.S. megacaps.

Goldman's case ultimately rests on two verifiable facts and one assumption. The facts: Europe's AI exposure is low, and its related valuations have compressed while earnings growth accelerates. The assumption: productivity gains arrive at something close to the scale the model projects. Investors who buy the thesis are buying the first two and hoping for the third.