Summary
- Brookfield is raising its planned French AI infrastructure investment from €20bn to €30bn.
- The increase sits alongside Data4 campus plans and Campus AI activity in France.
- France is using power availability and industrial policy to attract AI data centre capital, with delivery dependent on grid, permitting, and supply chains.
Brookfield is increasing its planned AI infrastructure investment in France to €30bn, up from the €20bn programme announced in 2025.
The increase emerged around the latest Choose France summit, where France announced a wider package of foreign investment commitments heavily weighted towards AI, data centres, energy, and industrial infrastructure. Brookfield’s earlier French programme targeted data centres and associated AI infrastructure, with a large share expected to be led through Data4.
Data4 is expected to build and operate major campuses at Cambrai E-Valley and Escaudain, while Brookfield-backed Campus AI is close to selecting a second site. Separately, Nebius is developing a 240MW data centre at Béthune in Hauts-de-France, reinforcing northern France’s growing role in AI infrastructure siting.
The Brookfield increase follows the investor’s broader global move into AI infrastructure, spanning energy, land, data centres, and compute. The company has previously described AI infrastructure as a value chain requiring capital at several points rather than a narrow data centre real estate play.
Capital is following power
France is becoming a test case for whether Europe can turn low-carbon electricity and industrial land into large-scale AI infrastructure. Brookfield’s additional commitment sits alongside SoftBank’s much larger French programme and other investment pledges made during the Choose France cycle. The shared logic is the search for sites where power availability, permitting, and political support can align quickly enough to justify major capital deployment.
Institutional investors are drawn to AI infrastructure because demand is rising across hyperscalers, AI labs, enterprises, and public-sector users. AI data centres are more complex than conventional colocation growth, however. They require more power per rack, more capital per megawatt, shorter delivery windows, and stronger integration with energy infrastructure.
Brookfield’s advantage is its ability to bring infrastructure capital, power-sector experience, and data centre platforms together. That combination is valuable in markets where the main constraint is not tenant interest, but the assembly of land, grid access, equipment, and operating capability. Data4 gives Brookfield a European development platform with existing market knowledge; the extra capital increases the pressure to convert pipeline into energised capacity.
Investment announcements can still create a capacity narrative faster than projects can be built. A €30bn programme implies multiple sites, large power requirements, complex procurement, and extensive planning work. Even with French state support, projects will need to pass through environmental assessment, grid studies, local consultation, construction mobilisation, and equipment lead-time constraints.
France’s commercial opportunity is clear. Data centre capital can support industrial regions, create construction work, anchor supply chains, and position the country as a European AI compute hub. The policy test will be whether that energy-intensive infrastructure delivers enough domestic value in return for grid capacity and land.
Brookfield’s increased plan is therefore both a vote of confidence in France’s AI infrastructure offer and a delivery challenge. The next phase will depend less on investment pledges than on powered sites, permitted campuses, and equipment that arrives when construction programmes need it.
