The story begins with an admission. Europe, a commentary in Luxembourg’s Delano argued at the start of 2026, has fallen decisively behind in the race to build foundational AI. The frontier now sits in the United States, with OpenAI, Google and Anthropic, and increasingly in China, where open-source models are doing impressive work.
Even France’s Mistral AI, the best-funded AI lab in Europe, remains invisible on the global stage, the author wrote, and Europe has not rallied behind it in the name of sovereignty. Usage data tells the same story, since no European platform has made a dent globally. In October 2025 the European Commission unveiled a €1bn plan to encourage AI across industries. The piece called it a drop in the bucket of global AI financing. The Draghi Report, a year on, has little to show for itself, with structural reforms and capital market integration largely unrealized.
Adoption as the real opportunity
At its heart, the argument is that losing the frontier race does not mean surrender. The path forward is adoption: putting AI to work in public services, businesses and everyday life. Estonia’s e-government is held up as proof that practical implementation can bring lasting benefits. Europe has world-class talent, robust industries and a commitment to ethical, human-centered technology, and the author sees that as a moral compass worth offering.
Two moves are proposed. The first is vertical AI companies, which build industry-specific products on top of frontier models developed abroad. They are far cheaper to build, and the author names Helsing, Luminance and Synthesia as examples. Incumbents can then focus on their core mission. Hospitals, the piece says, should concentrate on patient care, not on developing internal AI.
The second is national AI agendas rather than one driven from Brussels. Member states know their own economies, public services and citizens’ needs.
Luxembourg and the case for health care
The author, writing from Luxembourg, points to the country’s record of strategic bets, including being the first EU member state to put part of its sovereign wealth fund into Bitcoin. Finance is the obvious target, and Finance Minister Gilles Roth created an advisory committee on AI in finance in January 2026. But finance depends on private banks and market actors, so the author argues a national agenda should also reach sectors under direct government control, starting with health care.
The case rests on a hospital’s daily grind. Clinicians spend one hour on administration for every five hours of care. A randomized trial covering more than 100,000 Swedish women found that radiologists supported by AI beat standard double reading without it, detecting 29% more cancers and cutting radiologists’ workload by 44%.
