European Datacenters: Why 2026 Is the Year Late Investors Pay the Full Price

Two years ago, an investor watching the European datacenter market closely saw a niche segment reserved for digital infrastructure specialists. Today, they see one of the most contested markets in Europe. In two years, they will look back and understand that an exceptional entry window existed between those two moments — and that this window is now closing.

The mechanics of the imbalance

Demand for computing capacity in Europe grew 40% between 2023 and 2025, driven by three simultaneous forces: the massive rollout of language models by American hyperscalers across Europe, digital sovereignty pushing states and large enterprises to repatriate their data, and the growth of AI inference workloads that require minimal latency close to end users.

Supply has not kept pace. Primary markets — Amsterdam, Frankfurt, Dublin, London — are in a state of advanced saturation. Grid connection constraints in Amsterdam have frozen any new development above 10 MW. In Frankfurt, delivery timelines exceed eighteen months for existing assets. In London, price per kW has risen 35% in eighteen months.

The shift to secondary markets

This shift produces a predictable movement: hyperscalers are moving toward secondary markets. Madrid, Barcelona, Milan, Lyon, Warsaw, Athens — markets where available power, underutilized assets, and regulatory regimes favorable to green investment still exist.

This shift is not hypothetical. It is documented in the research mandates that hyperscaler Real Estate teams have engaged since early 2025. Microsoft, Google and Amazon are simultaneously seeking 20 to 80 MW campuses in at least six European secondary markets. The signing of 15-year triple-net leases in these markets grew 180% between 2024 and 2025.

The transformation premium as a return driver

In this context, the brown-to-green thesis is no longer a niche conviction — it is a direct response to an identifiable and measurable market tension. Underutilized existing assets in secondary markets trade at cap rates of 9% to 10%. A hyperscaler on a 15-year NNN lease in the same market, on a Tier III/IV certified asset with PUE below 1.15, trades at 5% to 5.5%.

This 400 to 500 basis point compression — multiplied by a stabilized NOI — produces investment multiples that liquid markets can no longer offer in a normalized rate environment. It is arithmetic, not speculative.

The window is closing

Two signals indicate that this cap rate compression window is reaching maturity. The first: European Core infrastructure funds — Blackstone, Brookfield, DigitalBridge — have announced significant allocations toward second-tier European datacenters for 2026-2027. When Core capital enters a market, entry premiums disappear.

The second: PPAs — Power Purchase Agreements — for renewable energy in European secondary markets are increasingly contested. Access prices for decarbonized power, which until now constituted a structural advantage for assets in development, are converging toward those of primary markets.

What this means for allocators

For investors who have not yet positioned capital on this thesis, 2026 is likely the last year in which entry points remain consistent with double-digit return objectives. Transformation alpha is still accessible — but it requires a technical execution capability that generalist vehicles do not possess.

The winners of this cycle will not necessarily be the most capitalized players. They will be those who combined, at the right time, technical transformation expertise, access to hyperscaler tenants, and financial structuring that makes assets liquid for Core buyers at exit.

This combination is rare. And in rare markets, the premium always goes to those who arrive first.