USA

Mistral to explore designing own chips, CEO Arthur Mensch says

French initiative Mistral AI is exploring designing its own chips and may eventually develop them, Mistral AI CEO Arthur Mensch told CNBC.

This is Mensch’s first comment on Mistral’s semiconductor ambitions, underscoring how the company wants to control more of its infrastructure as it competes with US heavyweights OpenAI and Anthropic.

“Of course it’s interesting,” Mensch said of the possibility of Mistral developing its own chips, adding that the company had not ruled it out..

Custom chips allow a company to “reduce the cost of token distribution to meaningful levels,” Mensch said. Tokens are units of data processed by artificial intelligence models.

“Having chips may come, I think it has to come at some point, but for now Nvidia“It’s a great partner for us and we’re testing a few things here and there,” Mensch told CNBC.

Mistral, valued at around 12 billion euros, develops artificial intelligence models but is also investing in building data centers with Nvidia chips. Headquartered in Paris, the firm is often seen as Europe’s answer to OpenAI and Anthropic.

The company is enterprise-focused and counts companies like the chip equipment giant. ASML is among its most important customers.

If Mistral were to develop its own chips, it would follow in the footsteps of the major American hyperscalers. Amazon And Googlehas designed its own semiconductors and deployed them in data centers.

Custom chips, also known as application-specific integrated circuits, are seen as a way for hyperscalers to gain more control over their hardware and software integrations, potentially offering a different product from competitors.

Data center expansion

Mistral on Thursday announced a new data center in France designed specifically for inference, the actual process of running artificial intelligence models.

Mistral has invested 4 billion euros in data centers in France and Sweden as it plans to increase its computing capacity.

“Europe is lagging behind in this regard” [the] “We’re investing in building the infrastructure and that’s why we’re investing in closing that gap,” Mensch told CNBC. he said.

He said that the bloc is facing not only a technological problem but also a macroeconomic problem, adding that Europe has now started to see artificial intelligence as a strategic asset, just as it looks at gas.

“If you really want to stay competitive in the race, you can’t afford a trade deficit of a trillion, and I think that’s something people realize that we’re talking about something that should be concerning for all of us.”

The additional computing capacity in France will be used to serve Mistral’s customers and other AI labs, Mensch said, without mentioning specific companies.

“AI labs need a lot of computation, and we have some of it, and some of them actually want a lot of computation from us today,” Mensch said.

Mistral’s CEO said the company should “prioritize access” for computing, with some going to AI labs and “more importantly” to its customers.

agent focus

On Thursday, Mistral also unveiled a new agency platform for businesses as it plans to compete with U.S. rivals including Anthropic and OpenAI, which have recently increased their own offerings in this space.

Mistral’s new corporate agent platform, which it calls “Vibe,” can handle tasks such as drafting and coding, the company said.

AI companies are increasingly focusing on agency AI, which are systems that can autonomously perform tasks on behalf of users.

“Vibe is an intermediary platform for the tasks at hand that engages artificial intelligence at the edge,” Timothée Lacroix, Mistral’s chief technology officer, said in a statement. he said.

“Users can set the brief and move on, with Vibe thinking, drafting, and delivering the finished work all from a single conversation. Vibe Code writes, tests, and deploys code across codebases.”

Mistral’s latest statements underscore its efforts to accelerate revenue growth. The company targets revenues of 1 billion euros in 2026. While this figure is a step higher than the 200 million euros it achieved the previous year, it is dwarfed by the balance sheets of OpenAI and Anthropic.