Pricing shaped around your capacity needs
Every engagement is structured around a contract model and a hardware generation. Tell us what you're building and we'll shape a quote around it.
Hardware platforms
One deployment standard. Two generations of NVIDIA hardware.
NVIDIA — Blackwell Ultra
GB300 NVL72
Dedicated, liquid-cooled GB300 NVL72 clusters for enterprise training and inference workloads.
- Architecture
- Blackwell Ultra
- Availability
- Dedicated capacity
- Cooling
- Direct Liquid Cooled
- Fabric
- High-bandwidth
NVIDIA — Next generation
Vera Rubin
Reserve your allocation on the next NVIDIA platform, while supply is being finalized.
- Architecture
- Vera Rubin
- Availability
- 2027 deployment
- Cooling
- Direct Liquid Cooled
- Fabric
- High-bandwidth
Get a quote built for your roadmap.
Rates are structured per contract model and hardware generation — request pricing and we'll respond with a tailored proposal.