French AI Startup Raises 105 Million Euros from Investors in Four Weeks

Some do not accept it, some fear it, some find it helpful in business, but one thing is certain – everyone is talking about artificial intelligence these days, which is becoming an increasingly frequent topic among schoolchildren, students, and entrepreneurs alike. As usual, it was only a matter of time before various investors sniffed fertile ground for planting new millions in new and exotic AI projects.

Mistral AI, a French startup based in Paris, has raised as much as 105 million euros in a seed funding round in just four weeks as a new competitor to OpenAI, as it will compete in building, training, and applying large language models and generative AI. This is, in fact, the largest capital raising in a seed funding round in European history.

The company’s co-founders are former employees of Google’s DeepMind and Meta, and they will focus on open-source solutions targeting businesses to create what CEO Arthur Mensch believes is currently the biggest challenge: ‘Making artificial intelligence useful.’ Mistral AI plans to release its first AI language models in 2024.

Lightspeed Venture Partners leads this funding round, along with Redpoint, Index Ventures, Xavier Niel, JCDecaux Holding, Rodolphe Saadé, and Motier Ventures in France, La Famiglia and Headline in Germany, Exor Ventures in Italy, Sofina in Belgium, and First Minute Capital and LocalGlobe in the UK, TechCrunch reported.

There is capital, but no product

Interestingly, among the shareholders are the French investment bank Bpifrance and former Google CEO Eric Schmidt. Sources close to the company confirm that the 105 million euros in funding values Mistral AI at 240 million euros. Mensch worked at DeepMind in Paris, while Timothée Lacroix and Guillaume Lample were at Meta’s Paris AI branch. Mensch stated that they began discussing the direction in which they could see the development of artificial intelligence as early as last year.

It is still early to speculate on what exactly the newly formed startup will do, given that it is only a few weeks old, but according to what Mensch said, the plan is to build language models using only publicly available data to avoid legal issues that others have faced during the training of chatbots.

And while some believe that open source has created a tricky landscape (and minefield) when it comes to areas like application security,’we believe that the benefits of using open source can outweigh the potential for abuse,’ he added.

It is also too early to think about how their future products will resonate in the market. However, what is interesting is the unique focus of the startup on enterprise clients rather than consumers, as well as the idea that there is a gap in the market to help these customers understand what they need to do and how they can do it.

– We currently have proof that artificial intelligence is useful in some cases, but there are still too many workers in various fields who are asked to be creative (with artificial intelligence), and we need to figure that out instead of them. We want to give them tools that are easy to use to create their own products – said Mensch.

Will the French AI newborn survive without a concrete product (which may not come out until next year) and with a huge amount of money, or will it be acquired by larger players in the meantime? That remains to be seen.