Verda raises $189M Series B to expand AI cloud infrastructure

The round drew additional investors including MUFG Innovation Partners, Supermicro, Varma Mutual Pension Insurance Company, Lifeline Ventures, 6 Degrees Capital, byFounders and Tesi. Angel investors included Ola Tørudbakken, director of AI systems at Meta Platforms, and Mark Saroufim, cofounder of Core Automation. The funding brings Verda's total raised to more than $450 million across equity and debt.

Verda positions itself as a neocloud provider focused on large-scale managed AI inference for enterprise customers, aiming to supply capacity where and when it is needed with high optimization and reliability. The company builds its own compilers and serving software and handles its own performance and reliability engineering.

Its internal AI lab runs real research workloads on the platform, addressing engineering problems such as improving graphics processing unit utilization, inference optimization and kernel engineering, Verda said. The company says that work means customers get more out of their infrastructure whenever a new model arrives, and that it shapes the product roadmap.

"There's a window right now to build one of the defining compute companies of this generation, and to do so from Europe," founder and CEO Ruben Byron said. "It won't be open for long."

Verda plans to operate more than 250 megawatts in 2027, with data center capacity live in Finland and more coming online across Europe, the United Kingdom, the United States and Asia. The company said it is preparing early deployments of Nvidia VR200 NVL72 rack-scale supercomputers in the coming months.

The company already powers AI workloads for organizations in more than 50 countries, ranging from early-stage startups to enterprises. It supports sovereign AI development firm Aleph Alpha GmbH's research and development infrastructure and provides inference for AI-assisted visual content company Magnific's media generation at millions of requests per day.

Epsilon Health trains bespoke radiology AI models on a dedicated Verda cluster using native image resolution. Arjun Karpur, Epsilon Health's head of machine learning, said the unexpected benefit was a team of engineers who could help design its own data streaming and cluster management, rather than simply handing over a cluster and walking away.

Verda said it would use the funding to follow customer needs. Today's workloads differ from those of a few years ago, with longer sessions, context accumulation over many turns and agentic use cases. Those challenges require faster provisioning and setup times for instant clusters, so Verda is investing in its AI lab to work with other AI labs, open-source projects and developer tool companies on co-research.

Through that work, the company aims to learn how compute is used, what model architectures are emerging and which optimization techniques are shaping real workloads. Verda also said it will focus on developer experience, shipping new features and capabilities intended to build enterprise trust, enable new tools and readiness, and strengthen the platform's robustness.

Sifted reported that Verda, whose offering spans data centers, hardware and cloud infrastructure, has become Europe's latest unicorn after the raise.