Join the event trusted by enterprise leaders for nearly two decades. VB Transform brings together the people building real enterprise AI strategy. Learn more
Mistral AI, the French artificial intelligence startup, announced Wednesday a sweeping expansion into AI infrastructure that positions the company as Europe’s answer to American cloud computing giants, while simultaneously unveiling new reasoning models that rival OpenAI’s most advanced systems.
The Paris-based company revealed Mistral Compute, a comprehensive AI infrastructure platform built in partnership with Nvidia, designed to give European enterprises and governments an alternative to relying on U.S.-based cloud providers like Amazon Web Services, Microsoft Azure, and Google Cloud. The move represents a significant strategic shift for Mistral from purely developing AI models to controlling the entire technology stack.
“This move into AI infrastructure marks a transformative step for Mistral AI, as it allows us to address a critical vertical of the AI value chain,” said Arthur Mensch, CEO and co-founder of Mistral AI. “With this shift comes the responsibility to ensure that our solutions not only drive innovation and AI adoption, but also uphold Europe’s technological autonomy and contribute to its sustainability leadership.”
How Mistral built reasoning models that think in any language
Alongside the infrastructure announcement, Mistral unveiled its Magistral series of reasoning models — AI systems capable of step-by-step logical thinking similar to OpenAI’s o1 model and China’s DeepSeek R1. But Guillaume Lample, Mistral’s chief scientist, says the company’s approach differs from competitors in crucial ways.
“We did everything from scratch, basically because we wanted to learn the expertise we have, like, flexibility in what we do,” Lample told me in an exclusive interview. “We actually managed to be, like, a really, very efficient on the stronger online reinforcement learning pipeline.”
Unlike competitors that often hide their reasoning processes, Mistral’s models display their full chain of thought to users — and crucially, in the user’s native language rather than defaulting to English. “Here we have like the full chain of thought which is given to the user, but in their own language, so they can actually read through it, see if it makes sense,” Lample explained.
The company released two versions: Magistral Small, a 24-billion parameter open-source model, and Magistral Medium, a more powerful proprietary system available through Mistral’s API.
Why Mistral’s AI models gained unexpected superpowers during training
The models demonstrated surprising capabilities that emerged during training. Most notably, Magistral Medium retained multimodal reasoning abilities — the capacity to analyze images — even though the training process focused solely on text-based mathematical and coding problems.
“Something we realized, not exactly by mistake, but something we absolutely did not expect, is that if at the end of the reinforcement learning training, you plug back the initial vision encoder, then you suddenly, kind of out of nowhere, see the model being able to do reasoning over images,” Lample said.
The models also gained sophisticated function-calling abilities, automatically performing multi-step internet searches and code execution to answer complex queries. “What you will see is a model doing this, thinking, then realizing, okay, this information might be updated. Let me do like a web search,” Lample explained. “It will search on like internet, and then it will actually pass the results, and it will result over it, and it will say, maybe, maybe the answer is not in this results. Let me search again.”
This behavior emerged naturally without specific training. “It’s something that whether or not on things to do next, but we found that it’s actually happening kind of naturally. So it was a very nice surprise for us,” Lample noted.
The engineering breakthrough that makes Mistral’s training faster than competitors
Mistral’s technical team overcame significant engineering challenges to create what Lample describes as a breakthrough in training infrastructure. The company developed a system for “online reinforcement learning” that allows AI models to continuously improve while generating responses, rather than relying on pre-existing training data.
The key innovation involved synchronizing model updates across hundreds of graphics processing units (GPUs) in real-time. “What we did is that we found a way to just unscrew the model through GPUs. I mean, from GPU to GPU,” Lample explained. This allows the system to update model weights across different GPU clusters within seconds rather than the hours typically required.
“There is no like open source infrastructure that will do this properly,” Lample noted. “Typically, there are a lot of like open source attempts to do this, but it’s extremely slow. Here, we focused a lot on the efficiency.”
The training process proved much faster and cheaper than traditional pre-training. “It was much cheaper than regular pre training. Pre training is something that would take weeks or months on other GPUs. Here, we are nowhere close to this. It was like, I depend on how many people we put on this. But it was more like, it was like, fairly less than one week,” Lample said.
Nvidia commits 18,000 chips to European AI independence
The Mistral Compute platform will run on 18,000 of Nvidia’s newest Grace Blackwell chips, housed initially in a data center in Essonne, France, with plans for expansion across Europe. Nvidia CEO Jensen Huang described the partnership as crucial for European technological independence.
“Every country should build AI for their own nation, in their nation,” Huang said at a joint announcement in Paris. “With Mistral AI, we are developing models and AI factories that serve as sovereign platforms for enterprises across Europe to scale intelligence across industries.”
Huang projected that Europe’s AI computing capacity would increase tenfold over the next two years, with more than 20 “AI factories” planned across the continent. Several of these facilities will have more than a gigawatt of capacity, potentially ranking among the world’s largest data centers.
The partnership extends beyond infrastructure to include Nvidia’s work with other European AI companies and Perplexity, the search company, to develop reasoning models in various European languages where training data is often limited.
How Mistral plans to solve AI’s environmental and sovereignty problems
Mistral Compute addresses two major concerns about AI development: environmental impact and data sovereignty. The platform ensures that European customers can keep their information within EU borders and under European jurisdiction.
The company has partnered with France’s national agency for ecological transition and Carbone 4, a leading climate consultancy, to assess and minimize the carbon footprint of its AI models throughout their lifecycle. Mistral plans to power its data centers with decarbonized energy sources.
“By choosing Europe for the location of our sites, we give ourselves the ability to benefit from largely decarbonized energy sources,” the company stated in its announcement.
Speed advantage gives Mistral’s reasoning models practical edge
Early testing suggests Mistral’s reasoning models deliver competitive performance while addressing a common criticism of existing systems — speed. Current reasoning models from OpenAI and others can take minutes to respond to complex queries, limiting their practical utility.
“One of the things that people usually don’t like about this reasoning model is that even though it’s smart, sometimes it’s taking a lot of time,” Lample noted. “Here you really see the output in just a few seconds, sometimes less than five seconds, sometimes even less than this. And it changes the experience.”
The speed advantage could prove crucial for business adoption, where waiting minutes for AI responses creates workflow bottlenecks.
What Mistral’s infrastructure bet means for global AI competition
Mistral’s move into infrastructure puts it in direct competition with technology giants that have dominated the cloud computing market. Amazon Web Services, Microsoft Azure, and Google Cloud currently control the majority of cloud infrastructure globally, while newer players like CoreWeave have gained ground specifically in AI workloads.
The company’s approach differs from competitors by offering a complete, vertically integrated solution — from hardware infrastructure to AI models to software services. This includes Mistral AI Studio for developers, Le Chat for enterprise productivity, and Mistral Code for programming assistance.
Industry analysts see Mistral’s strategy as part of a broader trend toward regional AI development. “Europe urgently needs to scale up its AI infrastructure if it wants to stay competitive globally,” Huang observed, echoing concerns voiced by European policymakers.
The announcement comes as European governments increasingly worry about their dependence on American technology companies for critical AI infrastructure. The European Union has committed €20 billion to building AI “gigafactories” across the continent, and Mistral’s partnership with Nvidia could help accelerate those plans.
Mistral’s dual announcement of infrastructure and model capabilities signals the company’s ambition to become a comprehensive AI platform rather than just another model provider. With backing from Microsoft and other investors, the company has raised over $1 billion and continues to seek additional funding to support its expanded scope.
But Lample sees even bigger possibilities ahead for reasoning models. “I think when I look at the progress internally, and I think on some benchmarks, the model was getting a plus 5% accuracy every week for like, maybe like, six weeks in all,” he said. “So it it’s improving very fast on, there are many, many, I mean, ton of tons of like, you know, small ideas that you can think of that will improve the performance.”
The success of this European challenge to American AI dominance may ultimately depend on whether customers value sovereignty and sustainability enough to switch from established providers. For now, at least, they have a choice.
Daily insights on business use cases with VB Daily
If you want to impress your boss, VB Daily has you covered. We give you the inside scoop on what companies are doing with generative AI, from regulatory shifts to practical deployments, so you can share insights for maximum ROI.
Thanks for subscribing. Check out more VB newsletters here.
An error occured.
