Samsung leads $21B valuation round for Mistral AI

The funding round saw participation from several other notable investors, including Scaleup Europe Fund, which is managed by EQT and supported by the European Union. The fund's portfolio includes entities like the European Commission, Novo Holdings, and Santander. Existing investor PSG Equity also joined in. With this new financial injection, Mistral's valuation has nearly doubled within a single year.

Arthur Mensch, CEO of Mistral, plans to invest a significant portion of the fresh capital into expanding infrastructure. While the company currently relies on rented computing capacity alongside its own data centers, the long-term strategy is to reduce dependence on external resources. "In the long term, it is planned to focus entirely on capacities that we build ourselves," Mensch stated in an interview with CNBC. The goal is to double Mistral's self-held computing capacity over the next five years.

Parallel to infrastructure expansion, Mistral intends to train "larger and faster models." However, the Paris-based company pursues a distinct business approach compared to US giants like OpenAI and Anthropic. Rather than focusing solely on developing proprietary AI models, Mistral specializes in custom AI tools designed to embed directly into existing corporate business processes. The company already maintains such partnerships with several major corporations.

A key driver of Samsung's high level of interest is a specific collaboration strategy. Just as ASML, the Dutch semiconductor equipment manufacturer, has integrated Mistral's AI specialist into its production processes, Mensch now aims for a similar cooperation with Samsung. This alignment explains the significant attention from the South Korean conglomerate.

Mistral remains optimistic about its revenue trajectory. The company expects its recurring annual revenue to surpass the $1 billion mark later this year. Given the new funding and additional partnerships, Mistral anticipates exceeding this figure significantly if current trends continue, though Mensch declined to provide a specific updated revenue forecast.

Strategically, Mistral aims to position itself as an alternative to both American and Chinese AI providers while establishing itself as "sovereign AI" for Europe and other regions. A key differentiator is its focus on open-weight models rather than exclusively offering proprietary closed systems. Mensch announced that new, highly competitive models would be released very soon.

Despite the availability of alternatives, Mistral insists on developing its own models to ensure it remains a reliable partner for enterprise customers. "They want from us the security that they will have access to better models in a year than today," Mensch explained. "We can only give this guarantee by continuing to train our models ourselves." This commitment addresses concerns about long-term model availability and trust.

Regarding competition, while Chinese AI models are technically capable, Mistral's value proposition lies in its role as a trusted partner for Western enterprises that may hesitate to collaborate with Chinese providers. There remains uncertainty regarding whether Chinese models will be maintained and developed long-term or if future export restrictions could impact their usage. For European companies, this creates an additional layer of uncertainty, reinforcing the need for a reliable local alternative.

Ultimately, Mistral's goal is to provide customers with the assurance that they will receive superior models in the coming year. This guarantee can only be delivered by maintaining control over model training and infrastructure, a path now backed by substantial investment from Samsung and other key partners.