Ericsson has partnered with France-based startup Mistral AI to develop advanced AI tools tailored specifically for telecom networks, with a long-term focus on automation, resilience and the evolution towards 6G.
The collaboration will combine Ericsson’s deep expertise in network infrastructure with Mistral AI’s large language models to co-develop AI agents designed to address telecom-specific challenges. These include automating legacy code translation, accelerating 6G research and development, and deploying custom AI agents to optimise network operations.
In practical terms, the companies aim to create AI software capable of improving decision-making in product development and deployment by bringing data closer to AI models. The partners said this approach will enable faster innovation cycles and set new benchmarks for performance and resilience in telecom infrastructure.
Marjorie Janiewicz, Chief Revenue Officer at Mistral AI, said the partnership goes beyond simply layering AI onto existing systems.
“This partnership with Ericsson isn’t just about applying AI to telecom, it’s about transforming networks from the ground up,” she said. “By combining our frontier models with Ericsson’s expertise in radio, cloud and carrier-grade systems, we’re customising models with deep domain knowledge that accelerate modernisation. Together, we’re setting a new standard for what AI can achieve in telecom.”
Dag Lindbo, Head of AI & Emerging Technologies in Ericsson’s Business Area Networks, emphasised the practical focus of the initiative.
“At Ericsson, AI for networks is about precision, not hype,” he said. “With Mistral AI, we’re applying advanced models where they matter most — from accelerating code migration and supporting 6G research to building trustworthy AI agents. This collaboration helps improve time to value for customers while strengthening network performance and resilience.”
The move reflects a broader industry push to embed AI more deeply into telecom infrastructure as operators look to automate operations, reduce costs and prepare networks for the demands of future connectivity.
