The privately-funded French AI lab Kyutai aims to conduct open-source AI research. It already has a total budget of $330 million.
Kyutai is a recently launched Paris-based AI lab funded by French billionaires, including Xavier Niel, CEO of Iliad. At the ai-PULSE conference, Niel gave an update: the lab now has a budget of nearly €300 million or $330 million, raised through contributions from Rodolphe Saadé, CEO of logistics giant CMA CGM, and the Eric Schmidt Foundation, among others.
The conference was organized by Scaleway, Iliad’s cloud division. According to Niel, Scaleway will support Kyutai’s research projects with Nvidia H100 GPUs at cost.
Kyutai has already begun hiring its core scientific team, including former researchers from Meta’s FAIR and Google’s DeepMind, and Patrick Perez, formerly of Valeo, will lead the research lab. As scientific advisors, Kyutai has assembled a team of renowned AI researchers, including Meta’s Head of AI Yann LeCun and German Leibniz Prize winner and AI researcher Bernhard Schölkopf.
Kyutai bets on open-source AI research
One of Kyutai’s main arguments for attracting researchers is its commitment to open science and the ability to publish their work:
All of Kyutai’s models will be open source, and researchers plan to publish not only the models but also the training source code and documentation of the training process.
“When it comes to the timeline, I don’t think our aim is necessarily to go as fast as Mistral, because our ambition is to provide a scientific purpose, an understanding and a code base to explain the results” said Alexandre Defossez, part of the core team. French AI startup Mistral AI, which received funding in June, released its first AI model, Mistral 7B, in September. Kyutai hopes to release its first model within a year.
President Macron defends France’s position on EU AI law
During the AI-PULSE conference, French President Emmanuel Macron expressed his support for open source and emphasized the need to make it a strength of France. He also defended France’s position on the European AI Act, arguing that use cases, not model makers, should be regulated. France has been lobbying to weaken the AI law in the trilogue negotiations between the three main European institutions – the Parliament, the Commission, and the Council.
Xavier Niel supported France’s position on the issue, arguing that Europe is currently lagging behind in terms of AI innovation and that regulation could slow down European newcomers and reduce their chances of catching up.