Berlin, GermanyWhat young candidates need to see in a defence job ad
3d ago
Back to Companies
Helsinki, Finland35+more jobs on
the careers site35+more jobs on
the careers siteHelsinki, Finland
Products & Solutions
GPU Instances
On-demand virtual machines backed by NVIDIA GPUs from V100 through the latest Blackwell Ultra B300, with pay-as-you-go and spot pricing. Instances spin up in 1-3 minutes and come with pre-installed NVIDIA drivers and CUDA toolkit. Storage options include NVMe block volumes and a shared filesystem that can be detached and reused across instances.
Instant Clusters
Self-service multi-node GPU clusters giving teams access to 16 to 128 GPUs linked by InfiniBand interconnect for distributed training. Clusters are provisioned on demand without sales involvement, making large-scale runs accessible without long lead times or enterprise contracts.
Bare-metal Clusters
Custom-built GPU clusters designed to a customer's exact specifications and serviced by Verda's infrastructure engineers. Suited to organisations with sustained large-scale compute needs who want dedicated hardware without the overhead of managing it themselves.
Serverless Containers
Auto-scaling endpoints for containerised AI models with pay-per-request pricing. The service scales to zero when idle and scales up automatically under load, making it practical for inference workloads with variable or unpredictable traffic.
Managed Endpoints
Pre-configured API endpoints for running state-of-the-art open models at scale. Designed for teams that want inference access to leading models without provisioning or managing their own GPU infrastructure.
Co-development
Custom full-stack AI solutions built and maintained by Verda's engineers for customers with specific use cases. Covers both the software stack and the underlying GPU infrastructure, working directly with customer teams.
About
DataCrunch (now Verda) is a Helsinki-based GPU cloud provider founded in 2020. It gives AI teams on-demand access to NVIDIA GPU instances and multi-node clusters for training, fine-tuning and serving machine learning models, with data centers in Finland and Iceland powered entirely by renewable energy.
Offices
Helsinki, FinlandBacked by
byFoundersJ12 VenturesSkaalaVarma Mutual Pension Insurance CompanyTesi
Customers & Partners
SimliNVIDIA (Preferred Partner)1X TechnologiesSGLang Project
Timeline
2025-11
Rebranded from DataCrunch to Verda.
2025-09
Closed a $64 million Series A round led by byFounders, with participation from Skaala, Varma, Tesi and J12 Ventures.
2024-10
Raised $13 million in seed funding, led by byFounders.
2020
Founded in Helsinki, Finland.
Open Positions (10)
the careers site
Products & Solutions
GPU Instances
On-demand virtual machines backed by NVIDIA GPUs from V100 through the latest Blackwell Ultra B300, with pay-as-you-go and spot pricing. Instances spin up in 1-3 minutes and come with pre-installed NVIDIA drivers and CUDA toolkit. Storage options include NVMe block volumes and a shared filesystem that can be detached and reused across instances.
Instant Clusters
Self-service multi-node GPU clusters giving teams access to 16 to 128 GPUs linked by InfiniBand interconnect for distributed training. Clusters are provisioned on demand without sales involvement, making large-scale runs accessible without long lead times or enterprise contracts.
Bare-metal Clusters
Custom-built GPU clusters designed to a customer's exact specifications and serviced by Verda's infrastructure engineers. Suited to organisations with sustained large-scale compute needs who want dedicated hardware without the overhead of managing it themselves.
Serverless Containers
Auto-scaling endpoints for containerised AI models with pay-per-request pricing. The service scales to zero when idle and scales up automatically under load, making it practical for inference workloads with variable or unpredictable traffic.
Managed Endpoints
Pre-configured API endpoints for running state-of-the-art open models at scale. Designed for teams that want inference access to leading models without provisioning or managing their own GPU infrastructure.
Co-development
Custom full-stack AI solutions built and maintained by Verda's engineers for customers with specific use cases. Covers both the software stack and the underlying GPU infrastructure, working directly with customer teams.
Open Positions (10)
the careers site
About
DataCrunch (now Verda) is a Helsinki-based GPU cloud provider founded in 2020. It gives AI teams on-demand access to NVIDIA GPU instances and multi-node clusters for training, fine-tuning and serving machine learning models, with data centers in Finland and Iceland powered entirely by renewable energy.
Offices
Helsinki, FinlandBacked by
byFoundersJ12 VenturesSkaalaVarma Mutual Pension Insurance CompanyTesi
Customers & Partners
SimliNVIDIA (Preferred Partner)1X TechnologiesSGLang Project
Timeline
2025-11
Rebranded from DataCrunch to Verda.
2025-09
Closed a $64 million Series A round led by byFounders, with participation from Skaala, Varma, Tesi and J12 Ventures.
2024-10
Raised $13 million in seed funding, led by byFounders.
2020
Founded in Helsinki, Finland.
Latest reports
Explore more
Explore more
Similar Companies
Faculty
Applies decision intelligence to government, defence and critical infrastructure, combining simulation, digital twins and AI agents in a platform called Faculty Frontier.
5 roles
hema electronic
Designs and manufactures FPGA-based embedded electronics and optronics hardware for military vehicles, submarines and frigates, as well as industrial embedded vision and intelligent camera systems.
4 roles
Blackshark.ai
Builds AI infrastructure that converts satellite and aerial imagery into 3D world models and operational intelligence for defence simulation, autonomous systems and disaster response.
3 roles
S
Stendr
Fuses data from multiple sensors into one real-time view so defence operators can detect and track drone swarms faster than legacy systems allow.
DataCrunch owns its name, logo and all referenced trademarks — used here for identification only. Claim, update or remove this profile.