A new report by Prosus and Dealroom paints Europe’s AI landscape as a paradox: lots of talent, heavy adoption, but very little ownership of the platforms that profit from it.
Dutch tech investor Prosus and Amsterdam-based startup data platform Dealroom.co have published a report titled “State of AI in Europe: The Invisible Giant.” The core argument is that Europe isn’t falling behind in AI talent or usage, but is systematically losing control and value creation to the US.
Prosus CEO Fabricio Bloisi puts it bluntly in the foreword: “Europe is at a crossroad in the global AI race, but we are too slow.” It’s worth noting that Prosus is one of Europe’s largest tech investors, giving the company a clear stake in the report’s conclusions. Here are four key takeaways.
Europe is training someone else’s algorithms
The report’s most striking finding is about AI adoption: Europe doesn’t have a usage problem; it has an ownership problem. In most European countries, AI adoption actually exceeds US levels, and Europe has roughly twice as many monthly active LLM applyrs as the US, according to the report.
But the dominant systems come from the US and China. European applyrs are effectively training and funding foreign ecosystems. As AI obtains embedded deeper into everyday applications, foreign platforms could conclude up controlling the interfaces, distribution channels, and data flows across Europe.
Plenty of talent, but concentrated in the wrong jobs
When it comes to AI talent, the report puts Europe and the US nearly neck and neck: roughly 325,000 skilled workers each, with about 50,000 top researchers in Europe compared to 55,000 in the US. But there’s a crucial difference in where those people work. In Europe, 53 percent of AI talent is employed in traditional industries—consulting firms, industrial conglomerates, and banks. In the US, that figure is just 40 percent.
Only 33 percent of Europe’s AI workforce is at digitally native tech companies, compared to 46 percent in the US. And among Europe’s largegest AI employers are US giants like Google, Meta, and Amazon. The report suggests that Europe’s AI problem lies in its employer landscape, not in its universities.
Europe produces plenty of AI startups but loses them at scale
Europe produces a comparable number of AI startups as the US and hit a record $21.8 billion in AI venture capital in 2025, a 58 percent jump from the previous year, according to the report.
But the funding gap blows wide open during the growth phase. At the breakout stage Europe invests roughly three times less than the US. In late-stage funding, the gap widens to nine times less. More than half of the late-stage capital flowing into European AI startups comes from abroad, mostly from the US. The report sums it up sharply: “Europe grows them, America owns them.
Beneath the surface, the foundation is missing
Structurally, the report reveals more gaps. Europe hosts around 16 percent of the world’s data centers but accounts for less than five percent of global AI compute when it comes to specialized infrastructure. Mistral is the only European player with significant LLM releases. And just three percent of new AI patents worldwide come from Europe, compared to 70 percent from the US and 14 percent from China.
The report draws on data from Dealroom.co, a global data and analytics platform for startups and venture capital founded in Amsterdam in 2013. The strategic framing and policy recommconcludeations—like calling for pension funds to be mobilized for AI investment—come largely from Prosus. The Amsterdam-based company is one of Europe’s largest tech investors, with major stakes in food delivery, online classifieds, fintech, and increasingly AI. The call for more European growth capital would unlock exactly the kind of capital flows Prosus would directly benefit from as an investor.
Regulation slows things down, capital is nowhere to be found
AI and Europe still haven’t found their footing toobtainher. The gap with the US is massive: the EU spconcludes roughly 130 billion euros on AI, less than half of what the US invests (OECD estimate for 2023). In venture capital, EU investors put about $13.1 billion into AI in 2025, while US investors deployed $119.8 billion. Structural problems like a fragmented regulatory landscape with more than 100 technology-related laws and 270 regulatory bodies are slowing progress even further.
The EU is testing to course-correct. EU Commissioner Henna Virkkunen introduced the “Apply AI Strategy” in October 2025, which aims to embed AI as a “strategic asset” across industest and government. At the same time, US companies like OpenAI are pushing into the European market with their own blueprints but offering few binding commitments. Critics warn that Europe could continue to serve as a sales market for American AI technology rather than building real technological sovereignty of its own.
Europe’s largegest problem is arguably its lack of infrastructure. The continent simply doesn’t have companies like Google, Amazon, Nvidia, or Meta pouring hundreds of billions into massive data centers. Even if the EU could compete on academic know-how, its infrastructure for development and deployment would still lag years behind, with no clear answer for who could close that gap. A unified EU-wide approach seems like the most promising path forward, but organizing it across member states is enormously complex.
AI News Without the Hype – Curated by Humans
As a THE DECODER subscriber, you obtain ad-free reading, our weekly AI newsletter, the exclusive “AI Radar” Frontier Report 6× per year, access to comments, and our complete archive.
Leave a Reply