Mistral acquires SkipLabs to automate code verification

person using macbook pro on black table — Myriam Jessier / Unsplash

Mistral has acquired SkipLabs to enhance its capabilities in automating code verification. The team of six engineers from SkipLabs will now form the new "Language" team at Mistral. Previously, code verification tools targeted human developers through feedback loops and error messages. However, with the increasing use of AI agents in code generation, as reported by Oracle, which states that 70% of its own code is now generated by AI, there is a need for compilers and analysis systems that can evaluate code in real-time during its writing process. The integration of SkipLabs' technology into Mistral AI aims to create languages that allow models to reason and self-correct natively, from the training phase through to final software verification.

Mistral is also positioning its own software range, including the Codestral, Devstral, and Mistral Code models. The French publisher bases its competitive advantage on its hosting architecture, which allows solutions to be installed on local workstations, private clouds, or directly on enterprise servers. This approach meets the requirements of regulated industries in terms of intellectual property protection. Clients such as SNCF and Capgemini have validated this integration model for their development teams.

Valued at 25 billion dollars, Mistral is accelerating its vertical integration program to capture niche competencies. The Paris-based company is increasing its merger and acquisition activities to deploy industry-specific models. Recent acquisitions cover the entire digital value chain: Pimento for marketing automation, Emmi AI for industrial processes, New Material for product design, and Koyeb for cloud hosting infrastructures. The absorption of SkipLabs provides Mistral with an internal engineering foundation (open source framework) to ultimately systematize the production of reliable code, a prerequisite for deploying autonomous software agents in production environments.