Right now, Mistral's data centers run on Nvidia chips. So does almost everyone else in AI.
Its CEO just said that might not stay that way.
In a new interview, Arthur Mensch shared something new. Europe's biggest AI hope is exploring designing its own chips.
Mistral has been building its own data centers. They run the company's AI models. The goal is to own more of what Mensch calls the "technology stack." That means everything from the software down to the computing power underneath it.
Designing its own chips would be the next step. It would also cut Mistral's reliance on Nvidia. Nvidia sells the chips nearly every AI firm is fighting to buy.
This is a page from the U.S. playbook. Google and Amazon already design their own chips. They do not lean fully on outside suppliers.
Mensch was careful with his words. He framed this as exploring, not a promise. He said Nvidia is still a great partner for now.
There is a money angle, too. Custom chips can lower the cost of running AI. Mensch says owning the chips should bring those costs down over time.
The race to break free of Nvidia is one of the biggest stories in tech, and we cover it in Market Briefs - five minutes a day, plus a free investing masterclass when you sign up.
Mistral wants to be Europe's answer to U.S. labs like OpenAI and Anthropic. The startup turned heads back in 2023. It raised more than $100 million as a four-week-old company with no product yet.
"Europe is starting to be looking at AI as a strategic asset," Mensch said.
Its big bet now is on "agentic AI." That is software that can do long, complex tasks for you. Think of a digital assistant that actually does the work.
Mistral's version is called Vibe. It pairs the company's chatbot with a coding tool. Coding is where this kind of AI has caught on fastest so far.
Mistral is also pushing hard into big companies. It wants firms across many industries to build its AI into daily work. Mensch thinks that shift will change how teams are run. Each firm has to decide which tasks to automate and where people should stay in the loop.
Owning the chips, the data centers, and the software is like a restaurant that grows its own food. It is more work up front. But it means more control later.
Mensch thinks we are still early. He pointed to "a lot of viscosity in adoption" inside big companies. That is a clunky way of saying firms are still slow to use this stuff.
To him, that slowness is the opportunity. The value has not been captured yet.
Europe wants its own AI champion. Right now, Mistral is the clear front-runner in that race.
For the AI moves that actually matter to investors, join Market Briefs and get a 45-minute investing course thrown in.
