Mistral bets on ‘build-your-own AI’ as it takes on OpenAI, Anthropic in the enterprise

In the world of enterprise AI, the failure of many projects often stems not from a lack of technology but from models that do not resonate with specific business needs. This gap, where generic models trained on widespread internet data miss the nuances of internal workflows and institutional knowledge, presents a significant opportunity for innovation. French startup Mistral aims to capitalize on this with the launch of Mistral Forge, a groundbreaking platform enabling businesses to develop tailored AI models using their proprietary data. Unveiled during Nvidia's annual GTC conference, which this year focuses heavily on AI advancements for enterprises, Mistral Forge represents a strategic move for the company. While competitors like OpenAI and Anthropic have gained traction with consumer-facing applications, Mistral has maintained a strong commitment to corporate clients. CEO Arthur Mensch reports that this focus is paying off, with the company on track to exceed $1 billion in annual recurring revenue. Mistral Forge offers enterprises the autonomy to customize AI systems to meet their unique requirements. According to Elisa Salamanca, Mistral's head of product, the platform allows organizations to create AI models fine-tuned for specific applications. Although some companies in the enterprise AI sector provide similar functionalities, their approaches typically involve only fine-tuning pre-existing models or layering proprietary data through methods like retrieval augmented generation. Mistral differentiates itself by allowing companies to train models from the ground up, potentially overcoming limitations associated with traditional methods, such as handling non-English data or highly specialized domains. This capability not only enhances the precision of AI systems but also reduces reliance on third-party model providers, mitigating risks associated with model changes or deprecation. Mistral's extensive library of open-weight AI models is at the disposal of Forge users, enabling them to create customized solutions. Timothée Lacroix, Mistral's co-founder and chief technologist, noted that while smaller models may not perform as universally well as larger ones, their customization allows businesses to prioritize strengths and address weaknesses. Mistral offers guidance on model selection and infrastructure while empowering customers to make final decisions. For those requiring additional support, Forge includes access to Mistral's team of forward-deployed engineers, who collaborate with clients to identify the right data and adapt solutions to specific needs. This hands-on approach draws inspiration from established firms like IBM and Palantir. Additionally, Mistral Forge is already in use by several notable partners, such as Ericsson, the European Space Agency, and Italian consulting firm Reply, with early adopters including ASML, the Dutch semiconductor giant that valued Mistral at €11.7 billion during its Series C funding last fall. Mistral anticipates that Forge will serve a variety of sectors, from government entities requiring tailored language models to financial institutions with stringent compliance needs and manufacturers looking for custom solutions.