The next major electricity-demand story in South East Europe may not come from households, steel mills or electric vehicles. It may come from data centres. Greece has already shown how quickly the curve can change. Applications from data-centre developers seeking access to the Greek high-voltage network have risen above 4.5 GW, compared with about 1.2 GW at the end of 2025. Most of the new interest is concentrated outside Athens, in regions such as Western Macedonia, Central Greece, Epirus and Evros, as available capacity around the capital becomes increasingly limited.
The scale is striking. ADMIE has issued final connection terms for eight data-centre projects with combined capacity of around 520 MW, while active applications in the wider Attica region are estimated at less than 500 MW as developers shift toward areas with more transmission headroom. The bottleneck is not just demand. It is the order and structure of connection processing. Many applications are reviewed according to submission sequence, meaning projects in congested areas can block queue movement even when other regions have available capacity.
For the SEE region, Greece is a warning and an opportunity. Data centres are not ordinary consumers. They require high reliability, continuous load, redundant grid supply, strong cooling conditions, fibre connectivity, bankable green electricity and credible proof of origin. They can improve baseload demand for generators, but they can also intensify local network stress. A few hundred megawatts of 24-hour demand can materially change the economics of a regional node.
June’s power-market data makes this more important. HU+SEE consumption rose by 3,360 MW in the second half of June as temperatures increased, pushing the market into heat-driven scarcity. Data centres would add a different type of pressure: less seasonal, more constant and more creditworthy, but also less forgiving in terms of interruption risk. They would deepen the value of transmission capacity, storage, flexible generation and long-term green PPAs.
This is where Serbia and Montenegro enter the broader story. Serbia has scale, industrial land, a growing renewable pipeline, cross-border connections and the possibility of structuring data-centre supply around solar, wind, storage and future pumped-storage capacity. Montenegro has hydropower, growing solar interest, strategic Adriatic connectivity and regional optic-network potential, but it also has tighter domestic generation margins and clear exposure to import risk when large assets are unavailable. Both markets can attract data-centre interest only if green electricity is treated as engineered infrastructure, not marketing language.
The financing logic is equally important. A hyperscale or AI-focused data-centre project needs more than a power contract. It needs a bankable electricity supply stack, grid-impact study, reserve strategy, metering architecture, Guarantees of Origin or equivalent documentation, and in many cases hourly matching between consumption and contracted low-carbon supply. For international tenants, the power package becomes part of the site-selection decision.
That creates an opening for hybrid renewable projects. Solar alone is not enough for a data centre because the load is continuous. Wind improves the profile, hydro or gas adds dispatchability, and battery storage supports ramping and short-duration resilience. The strongest regional offers will combine grid-secured land, renewable PPAs, backup capacity, battery storage and transparent measurement, reporting and verification.
Data centres are therefore becoming a power-market theme before they become a real-estate theme. The limiting factors will be transmission queues, connection timelines, cooling demand, land permitting and the ability to prove clean electricity. Greece has already reached the point where grid connection requests have become a strategic market signal. The rest of SEE should treat that as the first draft of its own coming demand story.
