Your model in Europe
One supported catalogue model, one European GPU environment and one API endpoint.
- Repeatable catalogue deployment
- Standard endpoint access configuration
- No mandatory monthly expert retainer
Bring your AI product to European infrastructure. We supply GPU capacity, deploy your approved software and help implement agreed data-handling controls. You keep your brand and customers.
Choose a starting point. See a model, a European GPU setup and an estimated budget for a private API endpoint.
Cloud hours mean allocated hours, including idle time. Dedicated servers are charged until cancelled.
No Kalkola setup fee for a standard catalogue model and API endpoint. GPU service rates include the standard deployment. Connecting your company data, documents, business tools or a custom interface is separately quoted. Activation follows confirmed capacity, scope and prepaid funding.
Planning examples, not live model availability. Memory estimates include weights, request cache and runtime headroom; speed and final fit require a benchmark. 4-bit quality and engine compatibility need validation. All prices exclude VAT, storage, licences and extras. Supplier price examples checked 9 October 2026.
Pay for GPU capacity at Kalkola service rates. Add customer data integration or expert help only when you need a tailored project.
One supported catalogue model, one European GPU environment and one API endpoint.
A bounded project to connect your approved data source or business workflow.
Technical evidence and a practical action list for one product deployment.
Need a tailored architecture? A custom technical review is €390 for up to 4 hours. Standard catalogue sizing does not require this service.
Ongoing customer data work? Optional expert support is from €390/month for up to 2 hours, with a 3-month initial term. Czech business hours; initial response within 2 business days. Extra work needs an agreed budget.
Additional agreed engineering is €95/hour. All prices exclude VAT. GPU rates include the standard deployment; human data work, storage, licences and specialist services are separate. Cloud usage requires a minimum €25 initial funded budget, applied to usage rather than charged as a setup fee. Dedicated capacity and supplier activation are funded before ordering. A written quote confirms scope, extras, availability and terms.
GEX45 GPU service. Standard model deployment included.
One-off supplier activation: €209. Kalkola standard setup: €0. Storage, licences and customer data work are extra.GEX63 GPU service. Standard model deployment included.
One-off supplier activation: €999. Kalkola standard setup: €0. Model fit and supplier availability must be confirmed.Kalkola service rates include our platform margin. Supplier costs checked 9 October 2026; dedicated rates include IPv4. Supplier activation is charged separately. We confirm the complete quote before taking any order.
GPU service rates include standard model deployment. Choose the GPU after checking model rights, memory, expected traffic and the European region.
For smaller models, embeddings and your next proof of concept.
Room for generative AI, richer visuals and more demanding models.
More memory for demanding inference, fine-tuning and research.
Kalkola GPU service rates, excluding VAT, based on supplier costs checked 9 October 2026. Cloud examples are for Paris. Storage, network and other extras are separately quoted. Capacity and final costs are confirmed before ordering.
View cloud price sourceA whole physical server for your project. Explore Hetzner’s GEX line for ongoing inference, fine-tuning and GPU-intensive work.
A focused starting point.
More room for your models.
A larger system around your GPU.
Kalkola dedicated service rates include IPv4 and standard deployment; VAT is excluded. Servers are charged while allocated, including idle time. Supplier activation is separate. Model fit, region and stock are confirmed before ordering.
View hardware sourceCloud or dedicated? Cloud instances suit individual jobs and changing requirements. A dedicated server keeps the whole machine allocated to your project between jobs.
Start with one product and an agreed European environment. Measure performance and costs before expanding capacity.
Build a company assistant, power a search experience or serve your own language model. Start with the memory and compute your model actually needs.
We agree one workload, benchmark the endpoint and document the setup. Your team approves the software, model rights and acceptance criteria.
Kalkola GPU service rates include our platform margin and standard catalogue deployment. Customer data projects, optional expert support and readiness work have separate, agreed scopes and fees.
We map data flows and implement agreed access, retention and logging controls. Legal advice, regulatory representation and certifications require a separately appointed specialist.
Cloud GPU price examples cover Paris. Dedicated GPU options include Helsinki and Germany. We confirm the exact site, backups, subprocessors, telemetry and support access for your deployment.
Let’s find your regionEuropean hosting is one part of the deployment. We help you prepare the technical evidence your team and European buyers need to review.
Record inference, storage, logs, backups, telemetry and support access. An EU-only data boundary is promised only after the complete architecture and access paths are checked.
Configure access, secrets, retention and deletion procedures. Gather the supplier and operational evidence needed for your procurement review.
Contracts define each party’s processing and operational role. Legal assessment, representation and regulatory approval require separate specialist work. Hosting and documentation do not certify the product.
You own the product and customer relationship. Kalkola operates the agreed deployment; your team retains responsibility for its software, model rights and product obligations.
Share your product, deployment rights and the European customer requirement you need to meet. We check technical fit and budget together.
Approve the region, data flows, hardware and scope. We deploy one endpoint and measure latency, throughput and cost with your test workload.
Review allocated usage and your spending budget before expanding capacity. Add paid expert work only when your data or integration needs it.
Kalkola, a product by Crowdechain, supplies European GPU capacity, deploys approved AI software and operates agreed endpoints. We also offer practical data-flow and documentation readiness work. You retain your product, brand and customer relationships.
Start with your model’s memory needs, expected workload and budget. Our cloud options have 24–80 GB GPU memory. Dedicated GEX45 servers have 24 GB; GEX63 and GEX131 have 96 GB each. System RAM is separate from GPU memory. Model size, precision and batch size affect what fits.
Hetzner’s underlying server billing is hourly with a monthly cap and a one-time setup fee. Charges continue while the server remains allocated, even when powered off. Your Kalkola quote confirms the customer price, setup charges, included services and cancellation terms before you commit.
Standard catalogue deployment has no Kalkola setup fee or mandatory expert retainer. GPU service starts at €0.99/hour or €279/month. GEX45 supplier activation is €209; cloud usage needs a minimum €25 initial funded budget applied to usage. Customer data integration, expert support, storage and licences are separate. All amounts exclude VAT.
We agree a named European deployment region before ordering. Published cloud prices cover Paris; dedicated GPU options include Helsinki and Germany. We also map where logs, backups and support access occur. A listed configuration is not a live stock guarantee.
Yes, we welcome enquiries from international teams building on European infrastructure. We confirm customer eligibility, workload requirements and supplier availability before agreeing a deployment.
We’re starting with a guided service. Share what you need and our team will confirm the configuration, price and delivery details with you. This page does not provision servers or charge your account.
Tell us what your European customers need, what you can deploy and your expected usage. We will scope a practical pilot with you.
Your product. Your customers. A clear deployment plan.