The leading French artificial intelligence company Mistral AI officially launched the Mistral Forge platform at the NVIDIA GTC conference, aimed at helping enterprises and government agencies build highly customized AI models using their proprietary data, and fully focusing on the enterprise services market.

This platform not only supports fine-tuning based on Mistral's open-source model library (such as the latest Mistral Small4), but also breaks through the limitations of traditional Retrieval-Augmented Generation (RAG), allowing customers to train models with deeper business insights from scratch, thereby solving the pain points where general models struggle to understand specific industry logic and institutional knowledge.

AI Education Speech Artificial Intelligence (3)

Mistral's recent strategic move directly targets OpenAI and Anthropic. Unlike its competitors who have a broad presence in the consumer market, Mistral has always focused on enterprise customers. CEO Arthur Mensch revealed that thanks to the steady implementation of this strategy, the company's annual recurring revenue (ARR) is expected to exceed 1 billion dollars this year.

Currently, the Forge platform has attracted several major partners, including Ericsson, the European Space Agency, and Dutch chip giant ASML. To lower the technical barrier, Mistral also introduced a team of Frontline Deployment Engineers (FDEs) to provide professional guidance for enterprises lacking experience in synthetic data creation and model evaluation.

As enterprises increasingly demand control over AI systems, compliance, and domain expertise, the launch of Mistral Forge marks a shift in the large model competition from "parameter scale competition" to "business value implementation." By unleashing the potential of small models in specific fields and ensuring data sovereignty, Mistral is redefining the standards for enterprise-level AI infrastructure. Its valuation of 13.8 billion dollars