Reference builds · Perth

AI workstation specs

Curated starting points for local development and governed pilots — we do not hold inventory or dropship. Links are reference only; confirm warranty and delivery with the vendor. Need implementation? See AI implementation Perth.

Tier A — Developer / 7B–13B models

32 GB RAM, 8 GB+ VRAM or Apple Silicon 24 GB unified, 1 TB NVMe. Suitable for coding assistants and small local models.

Tier B — Team inference 32B+

64 GB RAM, 24 GB VRAM (e.g. RTX 4090 class) or dual-GPU workstation. For on-prem experimentation with human-in-the-loop review.

Tier C — Edge / NUC for demos

Low-noise demo machines for boardrooms — not production HA. Use cloud burst for heavy training.

Disclosure

Dent Consulting may earn affiliate or referral fees where programs exist; links are provided for convenience. We are not the seller of record. For procurement and Entra-hardened rollout, request a scoped implementation quote via Contact.