Finland’s Verda Becomes a Unicorn With €161 Million for Europe’s Next Neocloud
Set Trending Topics as a preferred source on Google.
Helsinki-based AI cloud provider Verda has just closed a Series B round of 189 million dollars (about 161 million euros), giving the company a valuation of more than one billion dollars. The Finnish firm joins a growing group of European neoclouds that sell compute capacity for AI training and inference and take on the US hyperscalers on price and data residency.
Verda was founded in 2020 by Ruben Bryon under the name DataCrunch and adopted its current name in late 2025. It runs its own data centers in Finland, powered by renewable energy according to the company, and sells GPU capacity along with a software platform on top. The round was oversubscribed, Verda says.
Who Is Investing How Much
The Series B is led by Silicon Valley firm Emergence Capital. Also participating are MUFG Innovation Partners, server maker Supermicro, Finnish pension insurer Varma, Lifeline Ventures, 6 Degrees Capital, byFounders and state investor Tesi (Finnish Industry Investment). A group of angel investors joins them, among them Ola Tørudbakken and Mark Saroufim. How the money splits across individual backers has not been disclosed.
The round brings total capital raised across equity and debt to more than 450 million dollars (around 385 million euros). Earlier this year Verda raised 100 million euros in a round led by Lifeline Ventures, with debt financing from Nordic banks. Before that came a 12 million euro seed round and a 55 million euro Series A.
What the Money Is For
Verda plans to spend on three layers: its own data centers, the hardware inside them and the platform running on top. The company says it will have more than 250 megawatts in operation in 2027. Capacity is live in Finland today, with further sites planned in Europe, the UK, the US and Asia. On the hardware side, Verda expects early deployments of Nvidia’s upcoming VR200 NVL72 generation in the coming months.
On the product side, the focus is inference, meaning the day-to-day serving of models, where longer sessions and agentic workloads create bursty demand. Verda also wants to cut provisioning times for instances and clusters, add S3-compatible object storage and a managed Kubernetes offering, and expand enterprise features such as audit logs, single sign-on and confidential computing. SOC Type II and C5 certifications are meant to address procurement teams at larger corporations.
A dedicated AI Lab will continue work on compilers, serving software and kernel optimization, and collaborate with research teams and open-source projects.
€141 Million in Annualized Revenue
Verda puts its annualized revenue run rate for July 2026 at 165 million dollars (roughly 141 million euros), up from about 100 million dollars at the start of the year. The company employs more than 250 people in Helsinki, London, Taipei and San Francisco, and says it serves customers in over 50 countries.
Named references include German AI company Aleph Alpha, which runs its research infrastructure on Verda GPU clusters, image generation service Magnific, and Epsilon Health, which trains radiology models on a dedicated cluster.
Europe’s Neocloud Field Is Getting Crowded
Verda is far from alone in this positioning. Nebius is already a listed European neocloud: the Amsterdam-based company trades on the Nasdaq and reported revenue growth of 454 percent year over year for the second quarter.
British provider Nscale filed for a listing on the New York Stock Exchange last Friday, targeting a valuation in the region of 30 billion dollars. Its prospectus also shows the flip side of the business model: Nscale currently sells compute below direct cost, reports an operating loss of 492 million dollars, and depends on just two customers, Microsoft and Anthropic, for roughly 88 of the 103.4 billion dollars in its order book.
France’s Mistral AI is moving in the same direction. The model maker is building out its neocloud business, now hosts third-party models such as GLM-5.2 from Chinese provider Z.ai through Mistral Compute, offers regional endpoints with data residency guarantees, and has anchor customers including Amadeus, ASML, Capgemini and CMA CGM prefinance capacity. For its own data center south of Paris, Mistral took on 830 million dollars in debt financing.
The Bigger Competition Sits in the US
The real benchmark for European providers lies across the Atlantic. CoreWeave is the largest listed player in the category, joined by Crusoe, which pairs data centers with cheap and often remote energy sources, and Lambda, which grew out of renting GPUs to research teams.
SpaceX has entered the market from an entirely different industry. The company resells compute capacity in xAI’s Colossus data centers, including to Google, and has filed plans for AI data centers in orbit. For Europe’s neoclouds that means competing for the same Nvidia supply allocations, the same grid connections and the same large AI customers.
What the European providers share is the data residency argument and a price advantage over AWS, Azure and Google Cloud. Verda advertises savings of up to 90 percent. They equally share the capital intensity of the business: data centers, GPUs and power connections have to be financed long before utilization is certain. How durable the model is will show up in the quarterly numbers of the listed players.
