Executive Summary
- Through grid limitations and AI demand, the FLAP-D (Frankfurt, London, Amsterdam, Paris, Dublin) markets are becoming ‘fundamentally supply-constrained’ with areas in northern and southern Europe becoming more attractive
- Yet proactive electricity grid planning can be a pull factor for data centres, driving economic activity – particularly with the inexorable move towards AI
- Alongside smarter grid strategies, flexibility in data centre design and output is key, busting the myth that data centres are always firm loads. For AI data centres, this is even more pronounced
“The EMEA data centre market continues to expand at pace, but the past year has marked a clear shift from a growth-led cycle to one increasingly defined by delivery constraints, power availability and locational selectivity.”
So begins a report published this week by Colliers. The EMEA Data Centres Market Report H1 2026 confirms what many in the industry likely suspect: grid access is moving ahead of demand as a key factor in where new EMEA data centres are being built.
In Frankfurt, Europe’s second-largest data centre hub behind London, the report notes, the market is ‘increasingly constrained by power availability.’ Amsterdam is seeing expansion ‘severely limited by stringent planning policies, land scarcity, and acute grid congestion.’ It’s a similar story in Dublin, where in 2024 Google was refused planning permission for its Grange Castle expansion, citing grid capacity limitations and insufficient on-site renewable energy provision.
Paris, meanwhile, despite France’s low-carbon power mix, is also suffering from grid connection and distribution constraints, meaning operators are looking to sites with secured power access, including brownfield locations and areas near high-capacity substations.
Colliers argues, therefore, that while the core FLAP-D EMEA hubs are now ‘fundamentally supply-constrained’, emerging city markets are now gaining traction. These include Madrid, Milan, Lisbon and Helsinki in Europe, as well as Abu Dhabi. These offer ‘improved access to power, renewable energy integration and relatively more flexible planning regimes.’ Ultimately, the report concludes, the industry’s biggest challenge has shifted. It is not the case that attracting customers is the biggest concern, but securing enough electricity to build and operate new facilities.
So, what can be done to remedy it? A report published last year from Ember, a global energy think tank, explored how proactive electricity grid planning can be a pull factor for data centres, driving economic activity – particularly with the inexorable move towards AI.
The report noted, much like Colliers, the surge in power demand Europe is facing, and that demand growth is set to shift from the top five markets to northern and southern Europe. 62% of Europe’s data centre capacity is located in FLAP-D markets, expected to drop to 51% by 2035. “Major industry players will not wait for the necessary infrastructure to be built – they will just move to where it is readily available,” the Ember report noted.
This is where smarter grid strategies need to come in, Ember argues. Long connection queues are becoming increasingly common, and building transmission infrastructure is a long-term solution.
Countries could attract 20% more data centre growth by halving grid connection time, the report notes. Non-firm grid connections, where users agree with system operators that their consumption may be interrupted under certain conditions, is another option for new data centres. This ‘allows new customers to connect to a constrained grid much faster and may bring additional benefits such as lower network tariffs.’
Similarly, phased grid connections can enable collaborative infrastructure build-up. “Data centres could apply for an immediate connection capacity corresponding to around 50% of their needs, with the remaining 50% being delivered over an extended period of time,” adds the report. “This provides both the developer and the grid operator with a clear path to scale.”
The other aspect to consider is with regard to flexibility across data centres. Part of the issue is busting the myth that data centres are always firm loads; even for colocation facilities, electricity consumption can fluctuate up to 30% between days, according to Ember.
For AI data centres, this effect is even more pronounced, with electricity demand depending on model training and usage. “The load profiles can be optimised through the batching of AI queries and different execution schedules that align with electricity prices, grid congestion, and on-site generation,” Ember notes. “This presents an opportunity for more flexible power consumption, with benefits both for the grid and data centre operators.”
