The honest answer to the question who’s going to make the most money out of AI agents is: nobody knows.
But that’s not stopping a mad scramble of AI startups, backed by trigger-happy investors, rushing to figure it out. Just under half of the startups (67 out of 144) accepted into YCombinator’s first-ever spring batch are AI agent companies of one form or another. All business processes seem up for grabs, ranging from onboarding and payments to customer service and grant applications. “This year is a turning point: AI graduated from an answer engine to an action engine in the workplace,” Sequoia’s Konstantine Buhler wrote recently, explaining the potential of agentic AI.
The sector is just as hot in Europe as anywhere else, with scores of startups looking to build AI agents in what is being hailed as the next big business transformation. Although definitions vary, Sifted’s take is that an AI agent is a tool that acts autonomously to complete a wide range of tasks without human intervention (beyond prompting and supervision). By that definition, 88 European agentic AI companies raised €1.6bn in equity last year, according to Sifted data.
The enthusiasm has spilled over into this year. Everyone in Europe (and beyond) seems to have fallen in love with Lovable, Sweden’s poster child for agentic AI and one of the region’s fastest-growing startups in history. The company, which enables users to create websites using plain text prompts, raised $15m of pre-Series A funding earlier this year.
And Lovable appears to love Europe in return. “We have many of our customers in the US, so we need to be close to them, but I believe the core of Lovable will stay European,” Anton Osika, Lovable’s cofounder, told Sifted.
European startups are also building agentic AI tools for many other specific industries, including finance, marketing, education and healthcare. Last week Paris-based Nabla raised a $70m Series C to build AI agents for doctors, helping them with note-taking, report-writing, diagnosis and prescription. The fresh funding will also enable Nabla to accelerate its expansion in the US, where it already generates 90% of its revenue.
More sceptical investors, however, view the investment frenzy into agentic AI startups as the next big bubble, about to burst. One of the features of agentic AI startups has been how quickly and cheaply they can scale and generate meaningful annual recurring revenue. That should make them highly attractive to investors.
However, few agentic AI startups have sufficiently deep domain expertise or customer relationships to enable them to build moats around their business. They will always be at risk of being outcompeted by the next startup on the block armed with an even more capable AI model. Moreover, the giant US tech companies, including OpenAI, Google, Microsoft and Amazon are also building AI agents and often integrating them into their existing suites of services.
Even so, Mathias Strasser, an Austrian serial entrepreneur who has founded WSD, Scissero and Team Suzie — which all use AI agents — reckons that Europe is excellently positioned to take advantage of the latest technological wave.
The giant US tech companies may have spent billions developing their state-of-the-art proprietary AI foundation models, but Europe still has the “dumb luck” to be able to run commoditised open-source models at very low cost on desktop computers. With strong manufacturing and service companies, and a regulatory environment that should inspire more consumer trust, Europe has every chance of exploiting the full benefits of AI agents. “We should be the ones who really drive the AI transformation of our traditional industries,” Strasser says.
But dumb luck must still be married with smart judgment. Knowing how, when and where to deploy AI agents in meaningful, durable and profitable ways is going to require a lot more trial-and-error experimentation.
