Goldman Sachs sees AI investment adding to European growth in 2026

Global AI investment exceeded $1 trillion this year, and Goldman Sachs estimates the capex cycle will add 0.05–0.15 percentage points to European GDP growth in 2026, with a mean estimate of 0.10 percentage points. The bank says 2026 earnings growth expectations for European AI-exposed companies have risen by around €12 billion since the AI cycle began, equivalent to 0.05% of European GDP. Goldman expects the positive growth impulse to continue into 2027.

Analysis

European AI exposure looks more like a concentrated industrial and infrastructure trade than a broad earnings or GDP rerating. The key second-order beneficiaries are electrical equipment, cabling, cooling, and data-center construction suppliers; Schneider Electric, Siemens, ABB, and Prysmian are names to test for order and backlog exposure, not assume as direct winners. Physical buildouts may also lift local contractors and grid investment in the Nordics and Iberia. Conversely, data centers compete for scarce grid capacity and power: this can raise connection costs or electricity-price pressure for other users, potentially offsetting benefits for energy-intensive industry.

The macro impulse is small relative to the energy-price headwind and does not establish that European firms capture most of the global spend. Domestic construction value added, exports, and earnings expectations are different channels; avoid treating the estimated growth contribution as equivalent to broad-based profit growth. The contrarian risk is that investors capitalize a multi-year theme before European order conversion, margins, and power availability are demonstrated.

Near term, the signal is likely to matter through company commentary on orders, data-center exposure, and delivery schedules over the next 1–3 months. Over 6–18 months, grid connection constraints, permitting, and equipment capacity determine whether announced investment becomes realized revenue. A sharp slowdown in global AI capex, weaker supplier order growth, or renewed European energy stress would reverse the thesis.

Key Decisions for Investors

• Prefer a measured relative-value watchlist in European electrification and grid-equipment suppliers versus broad European cyclicals; initiate only if upcoming results confirm data-center-related order growth and backlog conversion. Do not pay a theme premium without evidence of margin capture.
• Track Schneider Electric, Siemens, ABB, and Prysmian disclosures for data-center orders, lead times, backlog quality, and capacity additions. Treat company-level exposure as unverified until quantified; supplier revenue can lag announced project spending.
• Monitor European power prices, grid-connection queues, and permitting in the Nordics and Iberia. Deteriorating access or higher power costs would weaken the regional growth benefit and could create relative pressure on energy-intensive manufacturers.
• Falsify the constructive view if global AI infrastructure spending guidance rolls over, exposed suppliers report decelerating orders or backlog conversion, or energy costs rise enough to offset the estimated activity benefit. No broad-market trade is warranted from this modest macro estimate alone.