Verda Raises $189M to Expand European AI Cloud Capacity
Helsinki-based Verda secures $189M to build managed AI infrastructure, targeting 250 megawatts of capacity by 2027.
Key points
- Verda raised $189 million in a Series B round to expand its AI cloud infrastructure globally.
- The company plans to reach 250 megawatts of operational capacity by 2027 across multiple regions.
- Verda differentiates itself by building custom compilers and software to optimize AI inference performance.
Verda, a Helsinki-based cloud provider formerly known as DataCrunch, has secured $189 million in new funding to expand its artificial intelligence infrastructure. The Series B round, led by Emergence Capital, is designed to help the company scale its managed AI services for enterprise clients across multiple continents.
This investment brings Verda’s total capital raised to over $450 million, combining equity and debt. The company aims to provide high-performance AI compute and platform services, differentiating itself by developing its own compilers and serving software to optimize reliability for large-scale inference workloads.
Investors back European cloud ambition
The funding round includes participation from notable entities such as MUFG Innovation Partners, Supermicro, and Varma Mutual Pension Insurance Company. Angel investors also joined the deal, including Meta Platforms AI Systems Director Ola Tørudbakken. SiliconANGLE notes that the involvement of these major players signals strong confidence in Verda’s strategy to build a competitive compute platform from Europe.
Founder and CEO Ruben Byron stated that there is a limited window to establish a defining compute company in this era. By building from a European base, Verda seeks to offer an alternative to dominant US-based providers, leveraging local engineering talent to create infrastructure that meets specific regional and global enterprise needs.
Scaling power and hardware capacity
Verda plans to operate more than 250 megawatts of data center capacity by 2027. While its initial live capacity is located in Finland, the company is expanding its footprint to include sites in the United Kingdom, the United States, and Asia. This geographic spread is intended to reduce latency and provide diverse options for global enterprises.
The company is preparing to deploy Nvidia VR200 NVL72 rack-scale supercomputers in the coming months. These high-end systems are critical for handling the heavy computational loads required for modern AI training and inference. The trade-off for this rapid expansion is significant capital expenditure, which the new funding is specifically earmarked to cover.
Custom engineering for real workloads
Unlike standard cloud providers, Verda maintains an internal AI lab focused on real engineering problems. The team works on improving graphics processing unit utilization and kernel engineering to ensure that customers extract maximum value from their hardware. This approach means that when new models are released, Verda’s platform is already optimized to handle them efficiently.
Clients like Epsilon Health have praised the hands-on engineering support, noting that Verda helps design data streaming and cluster management rather than just providing raw hardware. However, this deep customization comes with the requirement for close collaboration between the client’s technical teams and Verda’s engineers, which may demand more time and integration effort than standard cloud services.
