Offline LLM Systems // Nemotron · Gemma

Intelligence that works when nothing else does.

AIOD design, build and install air-gapped LLM appliances — self-contained AI agents running entirely on hardware you own. No cloud. No uplink. No data ever leaving the room.

Powered by NVIDIA Nemotron & Google Gemma · Jetson Orin to DGX-class hardware
Installed on-site · Verified air-gapped · Maintained for life

Deployments

Three systems. One principle: the model stays with you.

Every AIOD unit is a sealed, self-contained inference appliance — Nemotron weights resident on local silicon, knowledge base curated for the mission, no external dependency of any kind.

UNIT F-1 // PORTABLE

AIOD Field

Ruggedised inference in a sealed transit case. Carried to the places where help can't call out — and answers anyway.

computeJetson Orin
modelNemotron Nano · Gemma
powerbattery / solar
runtime18 hr autonomous

Deployed at: mountain rescue stations, vessels at sea, expeditions, remote clinics.

FIELD SPEC →

UNIT C-1 // RESILIENCE NODE

AIOD Civic

One node, one town. A community knowledge server that runs from a library or emergency ops centre when the grid — and the internet — are gone.

computeDGX-Spark class
modelNemotron Super
accesslocal Wi-Fi mesh
corpusmedicine · water · power

Deployed at: libraries, town halls, councils, civil-defence sites.

CIVIC SPEC →

UNIT P-1 // AIR-GAPPED

AIOD Private

Full LLM capability with zero exfiltration risk. Your documents, your model, your building — confidentiality enforced by physics, not policy.

computeRTX / DGX rack
modelNemotron / Gemma, tuned
RAGyour private corpus
networknone. ever.

Deployed at: law firms, hospitals, defence contractors, R&D labs.

PRIVATE SPEC →

The deployment protocol

From threat model to a machine humming in your building.

01 / ASSESS

Threat-model the mission

We map your knowledge domains, privacy constraints, power envelope and failure scenarios before a single part is ordered.

02 / BUILD

Spec, tune, red-team

Hardware sized to the workload, Nemotron variant selected, retrieval built over your corpus — then stress-tested offline.

03 / INSTALL

Delivered and verified on-site

We install, verify the air gap, and train your people. We leave with nothing — no keys, no copies, no telemetry.

04 / SUSTAIN

Knowledge packs, offline

Quarterly model and corpus updates arrive on signed, encrypted drives. Maintenance that never requires a connection.

Why offline

The threat model speaks for itself.

Cloud AI
AIOD appliance
Where do your prompts and documents travel?
third-party servers
they don't. ever.
Works during outages, disasters, blackouts?
no — dead with the link
yes — needs only power
Can the vendor read your data?
technically, yes
physically impossible
Subscription dependency?
service ends, AI ends
you own the machine
Attack surface
the entire internet
the door to the room

FULL SECURITY MODEL & NEMOTRON STACK →

The demo

Pull the cable. It keeps answering.

Every AIOD demonstration ends the same way: mid-conversation, we unplug the network. The agent doesn't notice. That isn't a feature — it's the entire point.

Max Fontana-Reval, founder of AIOD

Founder

Built by an operator, governed like a board matter.

AIOD is founded by Max Fontana-Reval — former CEO of a £10m+ UK manufacturer at 27, portfolio chair, certified NED and technologist. Founded in February 2023 as an offshoot of MAXFR LTD, AIOD treats resilience and privacy not as features but as the founding thesis. It started with a poem →

MEET THE FOUNDER →

Begin deployment

Deploy intelligence you actually own.

Tell us the mission — the mountain, the courtroom, the town, the lab. We'll spec the machine.

DEPLOY@AIOD.APP →

response < 24h · NDAs welcome · on-site demos available