Verda raises $189M for the GPU cloud investors once called too capital intensive

Why it matters

Verda now has more than $450M in equity and debt to build a European AI cloud, but its $1B-plus valuation depends on converting a self-reported $165M run rate into durable demand and high GPU utilization.

Ruben Bryon, founder and CEO of Verda, has raised $189 million in a Series B led by Emergence Capital, Bloomberg reported, giving his Helsinki-based AI cloud a valuation of at least $1 billion. Verda declined to give Bloomberg the exact figure.

The oversubscribed round brings Verda's total funding to more than $450 million across equity and debt, according to Verda. Participants included MUFG Innovation Partners, Supermicro, Varma, Lifeline Ventures, 6 Degrees Capital, byFounders and Tesi, along with angel investors Ola Torudbakken and Mark Saroufim.

Bryon started Verda, then called DataCrunch, from his Helsinki garage in 2020. Milosz Szewczak and Tamir Segal joined him during the first year to build the platform's foundations. Bryon had previously worked on GPU servers and cloud-rendering software, an unusually direct apprenticeship for the expensive infrastructure problem Verda would pursue.

From 300 investor calls to unicorn valuation

Bryon has said the founders needed roughly 300 investor calls to close Verda's first institutional round. Investors repeatedly told him the market was too capital intensive, too early and difficult to attack from Europe, he wrote earlier this year.

They were right about the capital requirement. Verda raised a $13 million seed round in 2024 and a $64 million Series A in 2025. In April 2026, Verda announced another $117 million equity-and-debt package, later expanded to $155 million with financing from Nordic Investment Bank. The new Series B takes the disclosed total above $450 million, according to Verda.

That pace reflects the economics of the AI cloud market. Verda must secure accelerators, data centers, networking equipment and power before customers can consume the resulting capacity. Software margins may arrive later. The hardware bill arrives first.

Bryon's original complaint about the large cloud platforms was practical: high prices, complicated systems and a procurement process poorly suited to developers who wanted GPUs immediately. Verda has turned that frustration into a vertically integrated product spanning GPU instances, InfiniBand clusters, serverless containers, storage, managed inference and an internal AI lab.

The Series B buys power and capacity

Verda says the new capital will expand all three layers it controls: data centers, the infrastructure inside them and the software platform customers use. Verda expects to have more than 250 megawatts of operations in 2027, with capacity in Finland and additional deployments planned across Europe, the UK, the US and Asia.

The plan also includes early deployments of Nvidia's VR200 NVL72 systems, faster provisioning and new services for large-scale inference. Verda recently added Slurm and Kubernetes configurations to its self-service GPU clusters, along with health checks, audit logs and a container registry located near customer compute.

Supermicro's participation is strategically relevant. Verda needs access to servers and the supply chain around Nvidia accelerators as much as it needs venture capital. Bringing an equipment supplier onto the cap table gives the round an operational dimension beyond its valuation.

Verda opened a San Francisco office in the first half of 2026, adding a US base to its European footprint. The company now has offices in Helsinki, London and San Francisco.

Revenue is growing, according to Verda

Verda says it reached a $165 million annualized revenue run rate in July, up from the $100 million run rate it reported in June. Verda also says it employs more than 250 people and runs workloads for customers in more than 50 countries.

Those figures are self-reported snapshots rather than audited annual revenue. A run rate annualizes recent performance, and it does not show contract duration, customer concentration, gross margins or how much revenue depends on capacity Verda has already financed. Those measures will decide whether rapid demand produces durable returns on hundreds of millions of dollars in infrastructure spending.

Verda has named Aleph Alpha, Magnific and Epsilon Health among the customers running production or research workloads on its systems. Earlier materials also listed 1X, ExpressVPN, Freepik and Nokia. Verda says its Finnish facilities use renewable electricity, pairing the sovereignty pitch with Nordic power and cooling economics.

A European cloud with global ambitions

Verda is raising into a market where the largest AI infrastructure operators measure their needs in billions. Crusoe announced a $3.9 billion Series F at a $30.9 billion valuation on September 17th. CoreWeave, which is publicly traded, recently expanded an AI infrastructure agreement with Meta to as much as $21 billion through 2032.

Verda remains smaller, but Bryon is making a distinct geographic bet. European AI developers and regulated enterprises want local jurisdiction, data residency and an alternative to US hyperscalers. Verda still has to prove that those preferences can support global scale, high equipment utilization and competitive pricing.

Bryon describes the present market as a narrow opening. "There's a window right now to build one of the defining compute companies of this generation," he said in Verda's funding announcement. The $189 million gives Verda additional time, hardware and power to take that shot. It also moves Bryon's garage-born cloud into the phase investors warned him about six years ago: scaling a capital-intensive infrastructure business before the market window closes.