The full AI stack, under one roof.
Add Data operates across the entire AI infrastructure stack — hardware, data, and strategy. Whatever stage of the journey your organisation is on, we have the expertise and the infrastructure to move you forward.
We design, manufacture, and deploy the systems organisations rely on to train, run, and scale AI — and we stand behind every one of them, from the first pilot to the production floor.
Principles we engineer by.
Engineered in-house
Designed, assembled, and validated under one roof. No reference designs, no outsourced quality.
Sovereignty by default
Your models and your data stay on your infrastructure. Full control, no third-party exposure.
Accountable end to end
One team owns the outcome — from architecture to commissioning to years of on-site support.
Sustainable by design
Direct-liquid cooling and dense, efficient builds — measurably lower energy per petaflop.
From one workstation to the world.
Founded — the first AD-4 ships
Add Data is founded to put data-centre-class compute under the desk. The first AD-4 workstation ships to a research lab.
Rack-scale, with the AD-8
The AD-8 server brings direct-liquid cooling to dense, multi-user training — the building block of every cluster since.
First turnkey AI factory
We design, build, and commission our first full AD-X deployment — power, cooling, fabric, and orchestration, handed over live.
Private models, 10+ countries
Add Data now hosts frontier-class models privately on customer infrastructure, with direct support across more than ten countries.
Engineers, not middlemen.
Add Data is run by the people who design the boards, route the fabric, and commission the floor — not a layer of management between you and the work.
Our founders built the first AD-4 by hand, and that hands-on culture still runs through every team — thermal, systems, manufacturing, and strategy. When you work with us, you talk directly to the engineer accountable for your system, from the first architecture call to years of on-site support.
We're deliberately flat and deliberately small. The people who lead our hardware, data, and strategy work are the same ones doing it — which is why we stand behind every machine long after it ships.