Open-source advocates warn EU faces twin deficits of scale and control as US dominates global infrastructure spending
Europe’s ambitions for sovereign artificial intelligence infrastructure face a fundamental mismatch between available resources and the investment required to build at scale, industry executives warned at a Brussels conference last month, as geopolitical tensions push global AI spending towards a projected $758bn by 2029.
The warnings came at OneNext 2026 in Brussels, hosted by OpenNebula, the open-source cloud platform developer, which brought together executives from Fujitsu, Nvidia, ARM, Supermicro, SUSE, HashiCorp and a cluster of European technology companies to debate the future of AI infrastructure on the continent.
Alberto P. Martí, vice-president of open-source innovation at OpenNebula Systems, set the tone in his opening keynote, noting that the United States still accounts for 76 per cent of global AI infrastructure spending. Europe, he said, confronts two structural problems.
“Europe is investing billions in AI infrastructure, but hyperscale AI infrastructure requires tens of billions per deployment,” he told delegates, describing this as a “scale gap.” The second challenge, which he termed the “control gap,” was equally pointed: “Many AI factories are being built on proprietary software platforms, recreating dependency instead of sovereignty.”
