Local AI Engine. Your own intelligence, running in your own server room.
5ac installs and hands over an on-prem AI cluster for your business. Your data never leaves the building. You own the entire stack.
Why businesses choose Local AI Engine
Full ownership
You buy the hardware, you own the hardware. No cloud rental, no per-call API fees. A clear, one-time cost.
Data never leaves the building
Customer records, drawings, formulas — they stay on your server. Built for finance, healthcare, legal and manufacturing.
Worry-free operations
5ac installs, 5ac operates, 5ac updates. You just use it. A monthly operations care plan is included.
Six-layer architecture
Local AI Engine is not just a model running on a server. It is six operating layers, all closed on site — that is how the promise “data never leaves the building” is guaranteed by architecture, not by commitment alone.
Identity — Sign-in & authorisation
Every query goes through centralised sign-in (AD/SSO) and group-based authorisation, matching each user's role in the plant. There is no separate AI account store outside your own identity system.
Orchestration — Agent coordination
Agents coordinate the work and keep read actions separate from write actions; sensitive actions always go through a human approval.
Model router — Model routing
Routing policy decides which request runs on which model; external APIs are disabled by default.
Private RAG — Internal knowledge base
Internal documents (SOPs, work orders, drawings) are indexed privately and retrieved strictly within ownership permissions.
Runtime — Inference infrastructure
The model runs directly on hardware on site (DGX Spark); data never leaves the building.
Audit — Logging & SIEM
Every question, decision and action is fully logged and exported to your security monitoring system.
Three tiers, transparent cost
| Basic | Advanced | Enterprise | |
|---|---|---|---|
| Capacity | 1 on-prem AI server | 2 servers + high-speed interconnect | 2–4 servers + HA, custom |
| Context window | 1M tokens | ~1.1M tokens, shared | 2.49M+ tokens |
| Concurrent users | ~6 | ~6–12 | 12–24+ |
| Estimated hardware price* | VND 175M | VND 345M | VND 600M+ |
| Ops care (per month) | VND 1.5M | VND 3M | VND 5M |
*Estimated hardware prices in Vietnam, passed through directly — no hidden markup. Installation, training and operational handover are quoted separately during consultation, based on each business's setup.
Our handover process
Consult on needs
We map your data volume, user count and industry, then pick the right tier.
Install & configure
5ac receives the hardware, installs the stack and tunes it for your workloads.
Ops training
On-site training for your IT team, plus Vietnamese operations documentation.
Handover & care
We verify, then hand over. The monthly Care plan handles the rest.
Standard timeline: Basic tier ≤5 working days, Advanced ≤8 (counted from when we receive the hardware).
Frequently asked questions
Does the price include hardware?
Yes. Hardware is passed through transparently in the quote — no hidden fees.
What if a machine fails?
Hardware warranty follows the vendor and is stated in the contract. The Care plan monitors and resolves operational issues.
Do I need my own IT team?
Not required. 5ac trains your existing team, and the Care plan handles monitoring, updates and security hotfixes.
Can I move from cloud to on-prem?
Yes. An advisor assesses your existing data and maps a migration path.
How long does handover take?
Basic ≤5 working days, Advanced ≤8, counted from when we receive the hardware.
Ready to own your own intelligence?
Contact us for a free consultation on your needs and a rollout plan for an on-prem AI cluster.
If you're a builder who wants to build on open source yourself: follow 5acAI-Lab →