European tech sovereignty discussions at TechBBQ shift from model capabilities to who controls decisions

The question dominating European technology conferences is no longer confined to what artificial intelligence algorithms can build, but has clearly shifted toward who holds actual control over this infrastructure. In the Danish capital, Copenhagen, digital sovereignty issues took centre stage at the TechBBQ conference, which concluded under the theme 'Emergence from Agency'. Panel discussions and side meetings repeatedly raised the question of how Europe can regain the initiative in the face of dependence on the United States and China, moving from being a tenant of computing power and smart models to a true owner of the tools of digital production.

This direction was driven by difficult practical experiences European companies faced earlier this year, when developers lost access to advanced models such as Mythos and Fable from Anthropic, directly disrupting software engineering teams and raising deep concerns over the reliability of full dependence on external platforms.The era of intelligent agents is not about machines possessing decision-making power as much as it is about what humans agree to relinquish to them.This was emphasised by Eileen De Bever, partnerships manager at the Novo Nordisk Foundation accelerator, who noted that discussions focused on keeping human judgement in the driver seat and defining the boundaries within which machines are permitted to act.

Concerns surrounding privacy and data protection were also prominent on the discussion stages. Meredith Whittaker, president of the Signal messaging app, warned against efforts to integrate agents into operating systems, describing the current phase as the construction of a massive data collection ecosystem disguised behind marketing campaigns that divert attention from indirect consequences. Whittaker noted that the market is seeing growing demand for privacy solutions that align directly with national sovereignty aspirations, amid rising warnings over the concentration of technological leverage within a small group of major labs.

In the same context, Emad Mostaque, co-founder of Stability AI and the Intelligent Internet platform, stressed that true sovereignty lies in possessing the capability to resist external power and pressure. Mostaque explained that control over generative models effectively means control over state governance as the economy increasingly relies on agents, warning against confining these capabilities to a monopolistic few. Mia Negro, head of engagement at the Live with Artificials organisation, supported this view, pointing out that having agents take over cognitive tasks and robots handle physical ones forces a thorough rethink of the concepts of ownership, labour, and economic and political participation in modern societies.

These shifts place technology leaders and entrepreneurs in the Gulf, Egypt, and the Levant before a stark strategic mirror. The operational risks stemming from outages in foreign models or volatility in their provisioning policies mean that building critical applications exclusively on cloud APIs hosted abroad poses a genuine risk to business continuity. Investing in hosting open models locally, securing proprietary computing infrastructure for enterprises and banks, and developing data flow governance frameworks become essential steps to protect operational autonomy and reduce exposure to abrupt decisions by major global corporations.