Step 2 of 9

Memory. Add working headroom.

Choose the memory capacity for local AI services, containers, vector workloads, and responsive multitasking.

Step 3 of 9

GPU. Choose your local AI accelerator.

Every TeraNAS Cube ships with INTEL PRO 16GB graphics for accelerated local inference, with higher-capacity GPU tiers for heavier workloads.

Step 4 of 9

SSD storage. Choose your system drive.

Expand high-speed SSD capacity for faster model loading, vector search, scratch workloads, and sustained data-intensive AI workflows.

System SSD

Step 5 of 9

HDD storage. Choose capacity and drive count.

Choose None, 8TB, 12TB, 16TB, or 20TB. Selecting a capacity starts at 1 drive automatically; use the − / + control to adjust quantity from 0–8. None can also use the quantity control while HDD pricing remains CA$0.

Step 6 of 9

Networking. Move data without waiting.

Increase network bandwidth to move model files, datasets, backups, and shared AI workloads with less transfer overhead.

Step 7 of 9

Network switch.

Choose a network switch or None to continue to warranty.

Step 8 of 9

Warranty and upgrade service.

Choose your warranty and upgrade coverage.

Step 9 of 9

Agentic AI training. Choose coverage and agent count.

Add dedicated training for AI agents (CA$120 per agent). Selecting a training starts at one agent; use − / + to adjust the quantity from 0–5.