Mistral AI launches privacy-focussed AI tool Forge to target enterprise clients

Paris-headquartered artificial intelligence (AI) startup Mistral AI made waves at Nvidia’s annual GTC 2026 conference with the unveiling of its new platform Forge, which enables companies to build powerful custom AI models using only their own private data, on their own systems.

Forge is positioned as a more privacy-focussed tool, unlike those from its rivals, including OpenAI, which rely on cloud services. Forge is designed for industries where data privacy is important, such as finance, defence, and manufacturing. It simplifies the creation of tailored AI by starting with Mistral’s proven open models and customising them to fit specific business needs, such as handling multiple languages or industry-specific tasks.

For Mistral, the new tool marks a shift from sharing AI models to powering enterprise AI strategies. CEO Arthur Mensch said the company is on track to cross $1 billion in annual revenue by 2026, moving beyond free models to high-value enterprise deals.


Early users of the tool include chip equipment maker ASML (one of Mistral’s backers), telecom leader Ericsson, the European Space Agency, Singapore’s defence agencies, and Italian consultancy Reply.

The launch comes as Nvidia pushes enterprises to adopt its powerful new GPUs. Mistral positions itself as the top choice for companies wanting customisable AI that runs on Nvidia hardware without relying on US cloud giants. This will drive demand for Nvidia GPUs for on-premises training.

The French company raised over $2 billion in Series C funding in September 2025, taking its valuation to $14 billion and making it the most valuable European AI company. Mensch had told ET in an interaction earlier this year that Mistral is looking to partner with sovereign cloud providers in India.