Finnish AI infrastructure startup Verda has raised $189 million in an oversubscribed Series B round, pushing its valuation above $1 billion as it expands GPU capacity across Europe.
Emergence Capital led the round, joined by Supermicro, MUFG Innovation Partners, Varma, Lifeline Ventures, 6 Degrees Capital, byFounders, Tesi, and several angel investors. The financing brings Verda’s total equity and debt funding to more than $450 million.
Verda, founded in 2020 by Ruben Bryon as DataCrunch, builds and runs full-stack artificial intelligence cloud infrastructure. By operating its own facilities in Finland and Iceland with Nordic renewable energy, the company provides bare-metal GPU clusters, custom compilers, and serving software tailored for AI model training and inference.
Verda reported its annualized revenue run rate reached $165 million in July, serving clients across 50 countries, including Aleph Alpha, Magnific, and Epsilon Health.
“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 Bryon said in the company’s funding announcement. “It won’t be open for long.”
Verda plans to expand operations to more than 250 megawatts in 2027 and deploy Nvidia VR200 NVL72 rack-scale systems. The company also secured hardware maker Supermicro as a strategic investor, easing its access to scarce accelerator supply.
Verda Is Betting on Vertical Integration
Verda’s primary differentiator is its refusal to operate as a passive compute landlord. Instead of merely renting raw silicon, Verda runs an internal AI Lab alongside its custom compiler and kernel engineering teams.
This vertical integration creates a feedback loop: Verda tunes its operating stack to hardware nuances before releasing capacity to external customers. For enterprise engineering leads, this provides performance optimizations that generic infrastructure rarely delivers.
Yet this structure introduces capital exposure. Because the company builds dedicated data centers and develops custom software, front-loaded capital expenditures are massive. Verda is aiming to secure up to $1.5 billion in funding this year and $10 billion by 2027, according to an interview Bryon gave to Bloomberg.
That scale creates execution risk as well as opportunity. Competing with heavily funded neocloud providers such as CoreWeave and Lambda requires sustained spending on GPUs, power, data centers, and software, making utilization rates and continued AI compute demand important factors as Verda expands.
How this affects developers and enterprise users
Verda’s ongoing expansion offers direct relief to machine learning researchers and engineering teams facing ongoing compute shortages.
The company offers GPU clusters with Kubernetes and Slurm support, along with engineering assistance for areas such as data streaming and cluster optimization. That could appeal to teams that want more infrastructure support than a basic GPU rental service provides.
Verda may also be attractive to European organizations that want workloads hosted within regional infrastructure. For companies in regulated sectors such as healthcare, finance, or defense, keeping sensitive workloads in European facilities can be relevant to data-residency and compliance requirements.
The larger question is whether Verda can scale fast enough to compete with larger neocloud providers while preserving the software and support advantages it is using to differentiate itself. With a 250-megawatt target for 2027 and billions more in funding potentially required, the next phase will depend as much on capital execution as customer demand.
Other news: SoftBank is seeking more than $11 billion through dollar- and euro-denominated bonds to help fund its next $10 billion OpenAI investment, in a deal that could become the largest APAC nonfinancial corporate bond offering on record.