Pricing
Predictable pricing for dedicated AI infrastructure
One rate card, in rand, with everything included. Commit the capacity you know you need, burst for the nights that run hot, and pay a published rate that gets cheaper the longer you commit.
The platform
GPU compute
NVIDIA L40S accelerators for serving and fine-tuning your models.
General purpose
CPU capacity for everything that runs around the model.
Storage
All-flash NVMe fabric at one flat rate: full performance, no surcharges.
Data protection
Daily snapshots and a weekly backup cycle, included with the platform.
Networking
Public IPs, routers, load balancers, VPN tunnels and dedicated ports.
How it works
From first contact to production
Three steps, no procurement maze, with the people who own the hardware on the other side.
- 01
Tell us what you run
A short conversation about your workload: models, data constraints, capacity, and where it all needs to live.
- 02
A rand proposal in 48 hours
A rand-denominated proposal, valid for the term: committed capacity, on-demand rates and managed services, all priced up front.
- 03
Production onboarding
Dedicated capacity with a direct line to the engineers who run it, alongside the self-service tools to manage it day to day.
Get a quote
Tell us what you run
We'll size it against your real numbers and return a scoped, rand-denominated proposal within 48 hours.