AI Now Institute proposes four policy directions to counter the dependence of European AI on US companies

An analysis published in collaboration with AI Now Institute, authored by Frederike Kaltheuner (an adviser to governments and institutions on AI industrial policy and sovereignty, formerly a special adviser to the Vice-President of the European Commission) and Leevi Saari (a doctoral candidate at University of Amsterdam and EU Policy Fellow at AI Now Institute), builds on earlier installments in a series on the European AI ecosystem. According to the authors, the European AI ecosystem remains exposed to the “gravitational pull” of dominant US players – even where European AI companies succeed, they argue that value still flows toward US AI labs and ultimately to hyperscalers, which are both indispensable infrastructure providers for Europe and its competitors.

The authors criticize the fact that the response from Brussels has so far relied primarily on market instruments – co-funding so-called gigafactories, supporting consortia building models, channeling institutional savings into venture capital and growth equity, alongside deregulation and the easing of labor law restrictions. According to the text, this strategy aims to replicate the dynamics of Sillicon Valley, where successful startups generate exits and those exits feed capital and experienced people back into the ecosystem.

The authors say they do not have “one magic solution”, but propose four distinct directions for thinking: take market uncertainty seriously, actively shape the market toward greater openness, competition and interoperability, address cloud dependence and AI sovereignty as one and the same problem, and finally clearly define what AI is actually meant to serve.

The text available for analysis is part of a longer three-part series and does not contain a detailed elaboration of each of the four proposals. Details can be found in the source article.