AI startup Mistral has closed a substantial €600 million (about $644 million) funding round, the Paris-based firm announced on Tuesday. The fundraise was reportedly based on a company valuation of $6 billion—a nearly threefold increase from its $2 billion valuation in less than six months.
“Mistral AI has grown about 3x in valuation since December and about 25x in the last 12 months,” Mistral AI researcher Devendra Chaplot wrote on LinkedIn—adding that the company is hiring.
The Series B fundraising round was led by General Catalyst, with participation from notable investors like Lightspeed Venture Partners, Andreessen Horowitz, Nvidia, Samsung Venture Investment Corporation, Salesforce Ventures, Cisco, and IBM.
“We are grateful to our new and existing investors for their continued confidence and support for our global expansion,” Arthur Mensch, CEO and cofounder of Mistral AI tweeted. “This will accelerate our roadmap as we continue to bring frontier AI into everyone’s hands.”
We are announcing €600M in Series B funding for our first anniversary. We are grateful to our new and existing investors for their continued confidence and support for our global expansion. This will accelerate our roadmap as we continue to bring frontier AI into everyone’s…
— Arthur Mensch (@arthurmensch) June 11, 2024
The cash infusion could make Mistral the world’s most valuable open-source AI company and the fourth largest AI company in general, behind only OpenAI, Anthropic and Databricks. The company was founded in April 2023 by a team that included former employees of Meta and Google’s DeepMind AI project.
In announcing the fundraise, Mensch emphasized the capital-intensive nature of the industry, citing the need for increased computing power and talent acquisition to “break the barriers of artificial intelligence.” The capital will accelerate the company’s roadmap, enhancing computing capacity, recruitment efforts, and global expansion plans—particularly in the United States, he said.
Mistral AI has been focused on commercializing its proprietary models through API-based products tailored for corporate clients with specific business needs, largely aimed at privacy or personalization. Simultaneously, the company has developed and released open-source models like Mistral 7B, Mixtral 8x7B, and Mixtral 8x22B under the Apache 2.0 license, making them accessible to developers and researchers.
This open-source focus is a key factor in Mistral AI’s success. By comparison, Stability AI—which develops Stable Diffusion, the world’s leading open-source image generator with virtually no competition in the space—is valued at over $1 billion.
Both firms are still dwarfed by top AI firm OpenAI, which also tripled its valuation with a February deal that put its worth at a whopping $80 billion.
At a technical level, the highly efficient models from Mistral can outperform larger offerings on various benchmarks, making open-source models competitive with top commercial AI models like GPT-3.5 Turbo from OpenAI, Claude 2 from Anthropic, or Gemini Pro from Google on the LMSys Chatbot Arena Leaderboard.
Although Mistral AI offers a free chatbot, revenue streams include enterprise plans, licensing proprietary models on a pay-per-token basis, and a forthcoming API that supports model fine-tuning for customers through usage fees or subscriptions. The company also plans to commercialize its models through a multi-year partnership with Microsoft’s Azure platform, which includes a $16.3 million investment and equity participation.
The AI boom has revitalized the open-source industry, with significant contributions from major tech companies. Meta leads the charge with 1909 open source models released to date. Google has over 800 contributions, Microsoft has shared 300 models, and Apple 47 contributions.
Beyond the open-source community, significant investments are being made in AI developments across various fields. Several AI startups, including Inflection.ai, Modal.com, Character.ai, and Perplexity.ai have recently secured substantial funding rounds to advance generative AI applications.
