TIME magazine has revealed this year's selection of the 100 most influential people in artificial intelligence - TIME100 AI 2025. Alongside the obligatory Silicon Valley faces, some very European priorities come to the fore: technological sovereignty, security and infrastructure realpolitik. Meanwhile, DeepSeek, with its Chinese "open-weight" approach, is simultaneously inspiring and triggering bans.
TIME 100 AI 2025 It's not a competition of egos, but a thermometer of power: who sets the standards, who regulates, who runs the infrastructure, and who captures the imagination. The list has been around since 2023, and is divided into four roles — leaders, innovators, shapers, thinkers — and thereby communicates that AI is not just code, but an ecosystem of politics, capital, culture, and security. So – TIME100 AI 2025.
Europe draws a “third pole”: Virkkunen, Chappaz, Ilott
Henna Virkkunen (European Commission)
The new Executive Vice-President for Technological Sovereignty, Security and Democracy has taken political “root” access to the European AI strategy. Under her auspices, the Commission launched AI Continent Action Plan (gigafactories, public/private supercomputing network, data factories), with the idea of Europe accelerating from inventions to industrial moonshotsAt the same time, it came to life in the summer voluntary Code of Conduct for GPAI, as a bridge to the full use of the AI Act. This is not just PR; it is a set of mechanisms that try to “unwind” Europe’s dependence on foreign clouds and chips.
Clara Chappaz (French Minister for AI and Digital Affairs)
France has stepped up in the last year: it has established INESSIA (National Institute for AI Evaluation and Safety) and presented a plan Dare AI, which aims to embed AI into the fabric of the French economy by 2030. Chappaz openly criticizes “sovereignty washing” — when American hyperscalers create a “sovereign” impression with European intermediaries. Her mantra? Europe must “act in a pack” against “predatory” hyperscalers and offer “third way” — ethical, sustainable and trustworthy.
Oliver Ilott (UK AI Safety Institute)
British AISI has grown from a task force into a lab that others are emulating in two years. Ilott now leads a team that tests advanced models and shares findings with partners — a kind of “technical brake system” for the hyper-fast AI race.
DeepSeek: an open-weight challenger — and a regulator magnet
Anyone who thought that the only competitors in AI were the USA and USA-2, found out on January 20th DeepSeek‑R1: the first open-weight model to seriously challenge current closed premiums. On the same day, the founder Liang Wenfeng spoke at a closed-door meeting with Premier Li Qiang. In a symbolic sense: Beijing is showing that it wants to build AI as a state-industrial project, not just a startup race.
At the usage level, DeepSeek has also become a “main character” in consumer statistics. According to the a16z report, on the Web lost since February peak >40 % traffic, on mobile growth flattened (–22 % from the peak), while Perplexity and Claude are climbing the curve. It still remains at the top of global AI applications, with Google Gemini consolidates second place for ChatGPT.
Regulatoryly, the story is even more "hot":
- Italy blocked access to the application and initiated proceedings with the privacy ombudsman.
- The Netherlands it is banned its use among government employees and opened an investigation.
- Germany has officially called on Apple and Google to DeepSeek withdrawn from stores due to illegal data transfers to servers in China.
- Czech Republic is using DeepSeek banned in public administration.
European investigations (France, Netherlands, Belgium, Ireland) mostly revolve around GDPR, cross-border data transfers and transparency. Countries are cautious: open doesn't mean yet safe use.
Old acquaintances — and a new pop culture force
The list is expected Sam Altman, Elon Musk, Jensen Huang and Mark Zuckerberg — people who move the curves of capitalization, chips and users. In the categories of leaders there is also Matthew Prince (Cloudflare), which indicates that network (security, latency, computational edge layer) are as crucial as the models themselves.
Among the innovators, there is a surprise that is not at all "small": an actress Natasha Lyonne came on the list as a co-founder Asteria Film Co. and with the movie Uncanny Valley, emerging with AI tools. Filmmaking enthusiasts are raising eyebrows, but that’s exactly the point — AI won’t just optimize cultural industries; it will transform them.
What this means for Europe (and everyone who creates with AI)
- Infrastructure as a policy: Gigafactories and supercomputers are not "luxuries", but a condition for European models to even have a chance in the competition with hyperscalers. AI Continent Action Plan This is literally industrial policy 2.0.
- Rules + sandboxes: GPAI Code is a practical guide to how to “fasten your seatbelt” for models before the AI Act comes into full force. Good for startups, essential for giants.
- The European “Third Way”: Chappaza's "pack hunting" is not nationalism, but the coordination of public money, standards and talent so that European companies do not remain eternal tenants in foreign clouds.
- Open ≠ unregulated: DeepSeek’s rise proves that “open weights” can democratize research, but data is still sensitive. GDPR remains a tough nut to crack even for the “cheapest” models.
TIME100 AI 2025
TIME100 AI 2025 is a kind of atlas of power: models and platforms reign in the US, and in Europe, institutional sovereignty, and in China, a national strategy is intertwined with “open-source” projects that irritate the market and regulators. If you’re looking for a signal amidst the noise, this is it: AI is not just a game of models, but a game of infrastructure, rules and trust — and Europe is rapidly closing the gap in all three.
