In just a few days, France will host the Artificial Intelligence Action Summit, with heads of state flocking to Paris to meet global tech leaders. They’ll most likely announce some big investments and diplomatic agreements focused on safety or the environmental impact of artificial intelligence.
Ahead of the summit, early-stage VC firm Galion.exe, growth investment firm Revaia, and advisory firm Chausson Partners teamed up to create the French AI Report, which looks at the current trends in the tech ecosystem.
While all eyes are currently on the U.S. and China with OpenAI looking to raise tens of billions of dollars and DeepSeek capturing everyone’s attention, there has been a boom in AI startups in Europe, too. In 2024 alone, AI companies represented around 20% of all VC funding in the region.
In total, that represents around $8 billion in funding for AI startups in 2024. That metric is most likely going to grow rapidly, as AI startups are still relatively young. Seventy percent of the capital raised by AI startups in 2024 was for a seed to Series B round.
European countries that tend to attract VC funding in general have also become the main AI investment hubs, with the U.K. leading the group, France and Germany following suit, and the Nordics punching above its demographic weight. Here’s the breakdown from 2020 to 2024:
Interestingly, as AI companies become bigger, they tend to attract international investors, with U.S. VC firms accounting for around 50% of money invested in AI companies at the Series C round and later.
In France more specifically, there are “more than 750 startups that have created 35,000 jobs and operate in all areas that are transforming today’s society,” Minister Delegate for artificial intelligence and digital technologies Clara Chappaz said at a press conference.
She also mentioned that there are 2,000 scientists focused on AI research and 600 doctorate students working on artificial intelligence. And you may have noticed that there are quite a few French engineers and researchers working for AI companies in the U.S., too.
The team behind the French AI Report looked more closely at the top 400 AI startups in France and tried to identify the rising stars. While Mistral AI and Poolside are already some familiar names for readers who have followed the AI industry, the vast majority of AI startups aren’t working on the next foundation model.
On the infrastructure front, some companies are optimizing the data workflows and pipelines, such as Linkup and Kestra, or improving inference performance, such as ZML, or developing agents that can sift through large data sets and improve productivity. Dust is a good example of that.
But the reality is that most AI startups in France are focusing on applications for specific verticals. Based on this report, two important areas for AI startups in France are health and climate.
Owkin and its spin-off biotech company Bioptimus have been leading the pack on the health tech front, but it’s a surprisingly diverse group of companies with three big areas of interest: imaging tools, drug discovery, and medical treatment improvement.
Similarly, while a large chunk of the AI industry is focused on productivity gains for office workers, artificial intelligence is also actively being used to build the next-generation of climate startups. In addition to agritech, carbon and energy management — two topics that are related — seem to be a big focus. There are also a handful of promising new materials companies that are emerging (Altrove, for instance).
All the companies included in the list can be found here. You’ll also find plenty of AI companies working on one job function — sales, customer care, HR or legal — and using AI to simplify the most common tasks.
Of course, some companies aren’t going to be there in five years. But many of them are currently growing at a rapid pace. We’re still at the beginning of the AI revolution, and though it’s easy to think about the AI industry as a zero-sum game with one country or one company “winning” over the others, it seems like the AI boom is more distributed than expected.
