Azur Datacenters has completed a 240MW AI campus in Béthune, Northern France, for neocloud operator Nebius, in one of Europe's largest dedicated AI infrastructure transactions. The deal, facilitated by digital infrastructure marketplace Inflect, concluded in late 2025. The project highlights France's growing position in European AI infrastructure, benefiting from nuclear power capacity and government support. Structure Research identified multiple large-scale developments across the country, including Data4's proposed 1GW campus in Cambrai and various projects in Grenoble, Châteauroux and Calais. Azur Datacenters' Emmanuel Vannier emphasised that AI facilities require fundamentally different engineering approaches than traditional datacentres, particularly for cooling infrastructure. The company is sourcing equipment from oil and gas manufacturers rather than conventional datacenter vendors. The Béthune campus will be delivered in three phases, with first capacity available in July 2026.

AI's Impact on Global Market Expansion Patterns: how artificial intelligence is redefining the future of Global infrastructure. At infra/STRUCTURE Summit 2025, industry leaders from Inflect, NTT and NextDC explored how AI is accelerating development timelines, reshaping deal structures, and redrawing the global data center map. The infra/STRUCTURE Summit 2025, held at The Wynn Las Vegas from October 15 - 16, 2025 convened the brightest minds in digital infrastructure to explore the seismic shifts underway in the age of artificial intelligence. Among the most forward-looking sessions was "AI Impact on Global Market Expansion Patterns," a discussion that unpacked how AI is transforming where and how data centers are developed, financed, and operated worldwide. Moderated by Swapna Subramani, Research Director, IMEA, for Structure Research, the panel featured leading executives including Mike Nguyen, CEO, Inflect; Steve Lim, SVP, Marketing & GTM, NTT Global Data Centers; Craig Scroggie, CEO and Managing Director, NEXTDC. Together, they examined how the explosive demand for AI compute power is pushing developers to rethink long-held assumptions about geography, energy, and risk. AI Is Rewriting the Rules of Global Expansion For decades, site selection decisions revolved around a handful of core variables: power cost, connectivity, and proximity to major user populations. But in 2025, those rules are being rewritten by the unprecedented scale of AI workloads. Regions once considered secondary are suddenly front-runners. Scroggie noted how saturation in markets like Singapore and Hong Kong has forced expansion across Thailand, Indonesia, Malaysia, and India, each now racing to deliver power, land, and permitting capacity fast enough to attract global hyperscalers. "You can't build large campuses in Singapore anymore," Scroggie said. "But throughout Southeast Asia, we're seeing rapid acceleration as operators balance scale, sustainability, and access to emerging population centers." The panelists agreed that energy constraints, not capital, are now the primary limiting factor. "The short term is about finding locations where power exists at scale," explained Scroggie. "The longer-term challenge is developing new storage and generation models to make that power sustainable."