Summary
- Colliers identifies 12.5GW of operational EMEA data centre capacity, anchored by the core FLAP-D markets.
- A 7GW planned pipeline and 78GW of early-stage projects show demand is not the main shortage.
- The report points to power availability, planning, and delivery certainty as the real tests for new capacity.
Colliers says power availability has become the defining constraint for data centre development across Europe, the Middle East, and Africa, with grid access now shaping where capital can realistically become operational capacity.
The firm’s H1 2026 EMEA Data Centres Market Report identifies 12.5GW of operational capacity across the region, anchored by the core FLAP-D markets of Frankfurt, London, Amsterdam, Paris, and Dublin. It also points to 7GW of planned pipeline across the largest project markets, including London, Frankfurt, Paris, Milan, Helsinki, and Riyadh.
The larger figure sits further back in the development funnel. Colliers says 78GW of early-stage projects are now in development, including gigawatt-scale campuses from groups such as WBS Power, DataVolt, Freo Group, Mistral AI, DATA4, Sesterce, Merlin Properties, Ecogrid Energy, Tritax, EdgeConneX, and G42.
Demand is no longer the useful question
The report reinforces a shift already visible in planning decisions, utility connection queues, and site marketing across Europe. Hyperscale cloud, sovereign workloads, and AI infrastructure are continuing to pull demand upwards, but the ability to build is becoming increasingly uneven. Announced megawatts only become useful capacity when land, grid connection, substations, water strategy, cooling plant, fibre, permits, and capital all line up.
That distinction is now changing the geography of the sector. The traditional hub markets remain commercially deep, but they are also the places where grid pressure, land scarcity, planning scrutiny, and community opposition are most acute. Newer locations can offer larger sites and a more sympathetic industrial story, but often have weaker ecosystems, longer customer conversion cycles, or utility reinforcement work still ahead of them.
Colliers’ framing also matters because it pulls data centres closer to mainstream infrastructure allocation. The sector is no longer just competing within a digital real estate niche. It is competing with electrification, manufacturing, housing, transport, and energy transition projects for access to the same grid capacity, the same planning bandwidth, and often the same political attention.
Power-rich does not mean shovel-ready
Markets that can show credible power availability are gaining attention, but the power-first sales pitch has limits. Developers still need clarity on connection dates, reinforcement cost, curtailment exposure, permitting risk, and whether local authorities are willing to accept large digital loads. A site described as power-rich can still be years away from a usable connection if transmission works, land assembly, or planning approvals are unresolved.
The report also sharpens the difference between mature colocation markets and large new campus markets. London, Frankfurt, Amsterdam, Paris, and Dublin still provide deep customer ecosystems, dense connectivity, and proven supply chains. Secondary and emerging markets may offer scale, but they have to prove they can support the operational model, not just host a site outline on a map.
AI demand has made this harder. Higher-density deployments increase the importance of power quality, cooling design, and electrical resilience, while also increasing the size of loads being taken to utilities. The strongest projects are therefore likely to be those that can combine land control, credible energisation, water or heat-reuse strategy, and a customer route to market.
For European development, Colliers’ report leaves a blunt conclusion. Demand can justify a pipeline, but only power, planning, and delivery capability can convert it into capacity.
