``` ARR revenue surpasses $400 million, "Europe's OpenAI" sees 20-fold increase in annual revenue. ```

```

French AI startup Mistral has achieved a remarkable growth leap, with annual recurring revenue surpassing $400 million, soaring twentyfold from a year earlier. This company, regarded as the "European OpenAI," is seizing a historic opportunity as European businesses and governments seek alternatives to American tech firms.

On February 11, the UK's Financial Times reported that Mistral's co-founder and CEO Arthur Mensch stated the company’s annual recurring revenue (ARR) has reached “over $400 million,” compared to just $20 million a year ago.

Mensch revealed that the Paris-based group, which was valued at nearly €12 billion last year, is on track to surpass $1 billion ARR by the end of this year. This growth is driven by Mistral’s aggressive expansion into large enterprise customers; it now serves more than 100 major clients.

The report states that as a key step in its expansion strategy, Mistral announced on Wednesday it will invest €1.2 billion to build a new AI data center in Sweden. This will be the company’s first such facility outside France, aimed at diversifying its layout to reduce dependence on external infrastructure. Previously, Mistral completed a €1.7 billion funding round in September, led by Dutch chip equipment maker ASML.

Mistral’s rise highlights a strategic shift in the European market amid geopolitical uncertainty. Mensch noted that Europe has realized its over-dependence on American digital services is "excessive and on the brink of collapse." Mistral is trying to capitalize on this regional demand by providing models, software, and computing power entirely independent from US vendors, offering crucial chips to customers seeking data sovereignty.

Vertical Integration and Infrastructure Expansion

According to reports, Mistral is implementing a “vertical integration” strategy—building and operating its own AI data centers, rather than simply relying on American hyperscale cloud computing providers like Amazon, Microsoft, and Google to bring products to market.

The company is partnering with EcoDataCenter to build the Swedish facility, which will provide 23MW of computing power and will go online next year.

Mensch explained that Sweden is an ideal location for hosting high-energy AI chips because the local energy is “both low-carbon and relatively cheap.”

This vertical integration model helps to fund chip training for the next generation of models—running client workloads during the day, and training new AI systems at night. It also gives European clients peace of mind knowing their data is stored on local servers.

Mensch expects that this infrastructure investment will generate over €2 billion in revenue within the next five years, describing it as a “fairly predictable business.”

“Sovereign AI” Demand Driven by Geopolitics

The report points out growing concerns among European boards and capitals that Trump’s foreign policy could force “technological decoupling.” Currently, over 80% of the EU’s digital services and infrastructure are provided by overseas firms, most of which are US companies.

Mensch bluntly stated that simply building data centers for American hyperscale companies isn’t of much use on a national level.

Although Mistral’s shareholders include Microsoft and Nvidia, and it has long claimed that its ambitions are global rather than limited to Europe, its position as Europe’s only local “cutting-edge” large language model developer gives it a favorable advantage.

Mistral’s clients currently include ASML, TotalEnergies, HSBC, as well as the governments of France, Germany, Luxembourg, Greece, and Estonia. About 60% of its revenue comes from Europe, the rest from the US and Asia.

On the capital front, while US rivals OpenAI and Anthropic are competing to launch IPOs, Mensch said Mistral does not need to go public this year.

He noted that readily available debt financing means the company is well-funded. As for IPO, Mensch said it is “absolutely something we will consider in the next few years,” to “ensure our future independence.”

Not a "Fairy Tale": Pragmatic Enterprise Applications

Even though OpenAI’s ChatGPT and Anthropic’s Claude have become the fastest-growing products in Silicon Valley’s history, Mensch offers a sober assessment of the current market.

He stated that many corporate clients feel “a bit disappointed” with “off-the-shelf chatbots,” as these products are hard to deliver a return on investment. He refuted a "fairy tale" view that a single system will eventually run all business processes.

Regarding Wall Street’s sell-off of traditional business software providers following new AI systems like Claude Code, Mensch believes this is not “very rational.” He thinks that because these traditional software firms control critical business data, they will not disappear.

However, he warned that for software startups merely building user interfaces for specific industries, “the value of this approach has greatly diminished today,” since AI can now understand intentions and generate the required interfaces.

Risk Warning and DisclaimerThere are risks in the market and investment should be done cautiously. This article does not constitute personal investment advice nor does it take into account the individual user's particular investment objectives, financial situation or needs. Users should consider whether any opinions, views, or conclusions in this article suit their specific circumstances. Investments made based on this article are at your own risk. ```