Complete system - not raw compute

An AI appliance should arrive with fewer unknowns, not bigger claims.

“AI appliance” describes a configured hardware and software system intended for a defined job. The phrase is useful only when the supplier makes the software, security, test and support boundary visible.

Decision in one minute

Buy an appliance when the value of a repeatable baseline and accountable handover outweighs the flexibility of a DIY build.

Three-quarter supplier render of a 4U OEM multi-GPU rack server — OEM platform reference render. It is not evidence of a completed customer build or final specification.

Buyer field note

Let the work and the room reject the wrong machine.

01 / Fit: Start with the accepted task, not a chassis or GPU headline.
02 / Site: Check power, cooling, noise, network and maintenance access.
03 / Alternative: Keep cloud, colocation or a smaller system in the decision.

Diagram showing an approved request, a local service, an approved store and a policy-controlled data path — A private deployment starts with the permitted data path, access policy and logging boundary. Original explanatory plate. It sets out a decision method, not a measured result.

The appliance is the operating evidence

A chassis becomes an appliance when the versions, model source, access route, monitoring, backup expectations, acceptance tests and owner are documented.

Opace proposes an open-source baseline rather than a proprietary hardware key. Portability is a strength, but maintenance and licence review still exist.

Model: Weights and quantisation.
Context: Working input and cache.
Concurrency: Simultaneous active work.
Headroom: Operations and change allowance.

The bands are a checklist, not a scale. Exact memory belongs to a named model and test. Decision aid for: The appliance is the operating evidence

Diagram of an evidence pack containing an asset schedule, burn-in record, health readings, workload test and admin guide — A credible handover records the supplied assets, checks, operating evidence, instructions and unresolved items. Original explanatory plate. It sets out a decision method, not a measured result.

GPU server remote management dashboard with system status, access logs and sensor monitoring panels — Supplier screenshot of the platform management interface. The final management features and access policy depend on the ordered system. OEM supplier reference image. Written reuse permission pending.

Configured for one agreed use, expandable later

The first release should solve a named problem: private document search, an internal coding endpoint, a speech queue or mixed GPU workers. Starting with “every AI workflow” makes acceptance impossible.

Once the baseline is stable, further models, integrations and users can be scoped against evidence.

Electrical Circuit, protection and UPS policy.
Thermal Airflow, room heat and cooling route.
Network Access, bandwidth and isolation.
Service Rack space, noise and maintenance access.

A technically suitable server is still the wrong purchase when the operating site cannot support it. Decision aid for: Configured for one agreed use, expandable later

Turnkey does not mean no technical owner

The customer still owns identity, network policy, backups, data governance, physical security and daily availability unless a separate managed service says otherwise.

Remote onboarding explains the system; it does not silently create a national installation or 24/7 support promise.

Fit brief Workload and alternatives recorded.
Reference test Conditions and results retained.
Exact build Bill of materials and substitutions visible.
Handover Owners, versions and exclusions signed off.

The evidence trail should explain why this platform was chosen and what remains outside the order. Decision aid for: Turnkey does not mean no technical owner

Transparent where other appliances are opaque

The proposed pages show platform, GPUs, per-GPU VRAM, aggregate VRAM, RAM, storage, power planning, price and known constraints.

That allows a commercial buyer to understand the outcome while giving a technical evaluator enough detail to challenge it.

Questions answered

Straight answers to common questions

Is an AI appliance the same as a GPU server?

The GPU server is the hardware platform. An appliance adds a defined software, security, test, documentation and support baseline around a particular use.

Is the software proprietary?

The proposed baseline is open-source-first with pinned versions, model and software licence records and a documented update procedure.

Does it work on day one?

The agreed acceptance workload should work at handover. Customer integrations, data preparation and organisational adoption remain separate unless they are explicitly included.

Continue the decision

Useful next steps

Put the claim to work

Turn this guidance into a testable requirement.

The brief asks about workload and operating conditions - not just budget.

Plan your requirements Compare proposed systems

An AI appliance should arrive with fewer unknowns, not bigger claims.

Let the work and the room reject the wrong machine.

The appliance is the operating evidence

Memory demand has more than one layer

Connect the category to the physical and operating system.