From creating models to integrating them
Mistral AI, one of the main European bets to compete in artificial intelligence with companies like OpenAI, Google, or Anthropic, is modifying its strategy. The French company has begun to give more weight to integration and deployment services of AI, in addition to developing its own models.
The change means that Mistral no longer wants to limit itself to developing its own model that competes directly with ChatGPT. Its platform also allows companies and institutions to use third-party models, including open models, within Mistral's infrastructure. In August, it announced, for example, the incorporation of the Chinese GLM model from Z.ai.
What changes for users
For the end user, the change may be less visible than what the strategic shift suggests. Mistral continues to develop its own models and offer its AI assistant, while expanding its commitment to enterprise solutions.
The company has also recently strengthened its presence among consumers through an agreement with Mozilla to integrate its models into Firefox Smart Window. Mistral Vibe will be one of the AI options available alongside other assistants like ChatGPT or Gemini.
And for companies
The move is more significant for corporate clients. Mistral seeks for companies to choose between different models and deploy them under conditions of greater control over data, infrastructure, and the location of processing.
The company presented this strategy in August as part of its commitment to what is called sovereign AI, with infrastructure and computing capacity located in Europe and models that can adapt to the needs of each organization.
The strategy is already being reflected in its client portfolio. According to La Vanguardia, Mistral is increasing its service contracts with companies like HSBC, BMW, Airbus, or SNCF while slowing down the launch of new models.
Does this mean that Europe is abandoning its own AI?
No. Mistral's change does not mean that Europe gives up developing its own models. In fact, the company has just closed a round of 3 billion euros that places it at a valuation of over 21 billion and maintains its investment in open models and European infrastructure.
What changes is the role that Mistral aspires to play in the market: from trying to become primarily a European competitor of the large American laboratories to also positioning itself as intermediary and integrator of different AI models.
This movement responds, in part, to a market in which developing cutting-edge models requires increasingly larger investments and in which companies can value both the model used and its integration, security, control, and deployment. Mistral itself maintains that organizations must be able to decide which models they use, where they run, and how they manage their infrastructure.