Mistral AI, a French AI startup unicorn hailed as the "European version of OpenAI," recently announced that it has secured an $8.3 million syndicated loan from banks. This massive investment will be directly used to build AI data centers in the suburbs of Paris and across Europe, with plans to deploy up to 13800 of NVIDIA's latest GB300 chips. This move not only signifies Mistral AI's official transition from a "model developer" to a "bottom-level infrastructure provider," but also echoes the strong European push for "sovereign AI" in recent years, attempting to carve out a path to data autonomy amidst competition from US tech giants.
1.3 GB300 chips arrive in Paris
According toFinancial TimesAndReuters NewsThe report states that Mistral AI's expansion plans are extremely ambitious. This $8.3 million funding round marks the first time since the company's inception that it has been provided as a bank loan by seven major financial institutions, including BNP Paribas and HSBC.
The primary purpose of this funding is to build a large AI data center near Paris, which is expected to be operational this year. The data center will deploy approximately 13800 of NVIDIA's latest GB300 chips, specifically designed to support the massive training and high-load inference computation requirements of large language models (LLMs).
In addition to deepening its presence in France, Mistral AI also plans to extend its physical infrastructure to Sweden and has set an ambitious goal of deploying approximately 200MW of computing power across Europe by 2027.
Embracing "Sovereign AI" to Break the US Cloud Monopoly
Founded in 2023 by researchers from DeepMind and Meta AI, Mistral AI has gained significant favor from European capital and governments because it perfectly aligns with Europe's "sovereign AI" strategy.
In recent years, European businesses and governments have become increasingly uneasy about entrusting sensitive data to US cloud giants (such as AWS, Microsoft Azure, and Google Cloud). To strengthen data governance and privacy protection, the European market is eager to establish a fully controllable AI computing environment "on-premises".
Among the many European startups, Mistral AI is one of the very few that simultaneously possesses the capability to develop world-class large-scale language models in its own right, and now it has joined the ranks of comprehensive European companies that provide local physical infrastructure services.
Why choose to borrow money? A smart leverage operation that doesn't dilute equity.
Last September, Mistral AI completed a massive €17 billion Series C funding round, valuing the company at €11.7 billion. Given its deep pockets, why did it choose to borrow from a bank this time?
Market analysts point out that, unlike the intangible assets of software development, data center facilities and NVIDIA AI servers are currently considered highly valuable "heavy assets." This type of hardware is ideally suited as direct collateral, and by raising funds through project financing or bank loans, Mistral AI can not only obtain the substantial capital needed for expansion, but more importantly, the founding team and existing investors will not face the risk of further equity dilution.
