Summary
- Deutsche Telekom is preparing an application for the EU's AI Gigafactory programme and has signalled billion-euro investment.
- Its existing Industrial AI Cloud in Munich is described as largely booked, providing a demand case for additional compute capacity.
- The EU plans to mobilise €20bn for several Gigafactories, with the first construction currently targeted for 2027.
Deutsche Telekom is preparing to bid for one of the European Union’s planned AI Gigafactories and has indicated that it is prepared to commit investment at billion-euro scale if the programme’s commercial and regulatory conditions are suitable.
T-Systems, Deutsche Telekom’s enterprise technology arm, disclosed the plan after chief executive Tim Höttges said the group was working intensively on an application for the EU-backed programme. The company has not yet disclosed a proposed site, electrical capacity, computing configuration, or construction budget.
The Gigafactory programme is intended to push Europe’s computing infrastructure beyond the smaller AI Factory model and into facilities capable of supporting the training, fine-tuning, inference, and deployment of frontier-scale artificial intelligence systems.
Deutsche Telekom already operates an Industrial AI Cloud in Munich, which T-Systems says is largely booked. That provides a commercial basis for the company to argue that a larger facility would respond to existing capacity demand rather than simply anticipated future use.
The European Commission plans to mobilise €20bn for several AI Gigafactories through a public-private investment structure. An earlier expression-of-interest exercise produced 77 proposals across 60 sites in 16 member states, and the formal EuroHPC Joint Undertaking call was launched on 30 July 2026.
Compute policy becomes an infrastructure programme
The Commission’s preliminary timetable envisages construction of the first Gigafactory beginning in 2027. At that point, the initiative moves from technology policy into the same physical constraints facing the rest of the European data-centre market: land, grid capacity, cooling systems, electrical equipment, planning, supply chains, and the time needed to commission large blocks of compute.
Facilities designed for frontier AI systems also concentrate loads differently from conventional enterprise data centres. High-density accelerator deployments can push more power into smaller technical footprints, increasing the importance of liquid cooling, heat rejection, power-distribution architecture, and resilient electrical systems.
The Commission presents the Gigafactory programme as part of a wider attempt to reduce Europe’s dependence on computing infrastructure located or controlled elsewhere. It also expects large domestic compute projects to create demand for European semiconductor design and, over time, potentially manufacturing.
That industrial-policy objective does not remove the delivery problem. A project large enough to compete with leading AI infrastructure in the US or elsewhere must secure substantial electricity capacity at a point when several European markets are already tightening rules for large grid connections.
Site selection will therefore carry as much weight as the computing specification. A Gigafactory with public backing but a distant connection date would not close Europe’s compute gap quickly, while a site with available power but weak network connectivity, limited cooling options, or difficult planning conditions creates a different set of constraints.
Deutsche Telekom’s existing infrastructure gives it an operational base from which to make a bid. The Munich Industrial AI Cloud provides a smaller-scale reference point for sovereign European compute, and the company can combine telecommunications networks, cloud operations, data-centre engineering, and enterprise customers inside one proposal.
The next material step is the application itself. Telekom has not named the location it intends to put forward or whether it will bid alone or as part of a consortium. The EU programme is designed around private-led facilities supported by public financing, which means the eventual project structures are likely to combine operators, technology suppliers, investors, utilities, and public bodies.
For the European data-centre market, the Gigafactory programme is becoming another large source of future power demand alongside hyperscale cloud and colocation expansion. The policy goal may be AI sovereignty, but the build-out still begins with substations, cooling plant, grid agreements, and construction schedules.
