Mistral ai, a Paris startup founded seven months ago by researchers at Meta and Google, has raised 385 million euros, or about $415 million, in yet another sign of the feverish interest in a new type of artificial intelligence that drives online chatbots.
The deal values the 22-person company at about $2 billion, two people familiar with the deal said. Investors include Silicon Valley venture capital firms Andreessen Horowitz and Lightspeed Venture Partners.
The value of the new company has multiplied by more than seven in just six months. In the summer, it raised a seed funding round of €105 million (about $113 million) that valued the company at about $260 million.
Mistral creates technology that other companies can use to implement chatbots, search engines, online tutors, and other ai-based products. It is part of a small group of companies (including tech industry giants and a handful of startups) that are developing ai that could rival the technology being developed at OpenAI, the San Francisco startup that started the ai boom last fall with the launch of the ChatGPT chatbot.
Mistral is also among the companies that believe in sharing this technology as open source software (computer code that can be freely copied, modified and reused) giving outsiders everything they need to quickly build their own chatbots. Rival companies such as OpenAI and Google argue that the open source approach is dangerous and that the raw technology could be used to spread disinformation and other harmful material.
Mistral's fate has taken on considerable importance in France, where leaders such as Bruno Le Maire, the finance minister, have signaled that the company gives the nation an opportunity to challenge American tech giants. Europe has not produced many significant technology companies dating back to the dot-com boom and sees artificial intelligence as a field in which it can gain ground.
Investors are pouring money into other startups that believe in the open source approach. Perplexity, founded last year by another group of top researchers, has raised a new $70 million funding round that values the company at $500 million, a person familiar with the deal said. Investors include IVP and Bessemer Venture Partners.
“We simply believe that ai should be open,” said Anjney Midha, general partner at Andreessen Horowitz, who led the new investment in Mistral. Most of the major technologies that drive modern computing are open source, she added, including operating systems, programming languages and databases.
Mistral was founded by Timothée Lacroix and Guillaume Lample, who previously worked as researchers at the artificial intelligence laboratory of Meta, owner of Facebook and Instagram, in Paris, and Arthur Mensch, who was a researcher at DeepMind, an artificial intelligence laboratory run by Google. acquired in 2014 for $650 million.
Company employees like to joke that the first letters of the founders' last names are “LLM,” which is also short for large language model, the artificial intelligence technology the company is developing.
Companies like OpenAI, Microsoft and Google are leading the ai race, after spending hundreds of billions of dollars on this type of technology. By analyzing huge amounts of digital text scraped from the Internet, a large language model can learn to generate text on its own. This means you can answer questions, write poetry, and even generate computer code.
Companies like OpenAI and Google believe this technology is so powerful that they can only release it to the public in the form of an online chatbot after spending months applying digital barriers that prevent it from spewing disinformation, hate speech, and other toxic materials.
But many ai researchers, technology executives and venture capitalists believe that the ai race will be won by companies that develop the same technology and then give it away for free, without guardrails.
Meta, the former home of two Mistral founders, has been at the forefront of companies promoting this open source approach. This year, the tech giant created a great language model called LLaMA and essentially gave it away as open source software.
On Sunday, Mistral also released its latest technology as open source software, saying it works at an equivalent level to Meta's technology.
Widely sharing the underlying code of ai, Midha said, is the safest path because more people can review the technology, find its flaws, and work to eliminate or mitigate them.
“Now a single engineering team can find all the bugs,” he said. “Large communities of people create cheaper, faster, better and more secure software.”
Adam Satariano contributed reporting from London.
