Start on your own box. Scale into your own cloud.
Adam AI is containerised from day one. When a department outgrows its on-premises machine, the same containers move to private GPU hosting in your Microsoft Azure, Amazon Web Services or Google Cloud tenancy. Your data, your region, your keys. Nothing is rebuilt and nothing is handed to a third-party AI vendor.
One platform, your choice of where it lives
On your premises
A dedicated AI workstation or desk station on your network. Fastest route to production, lowest running cost, and the data never leaves the building. Where every deployment starts.
Private GPU cloud
The same containers on GPU instances inside your own Azure, AWS or Google Cloud subscription, in a UK or EU region you choose. You hold the account, the keys and the bill.
Hybrid
Sensitive workloads and the 2nd brain stay on site; burst compute, large models or 24/7 agent teams run in your cloud. The model layer routes each request by how sensitive the data is.
Private GPU hosting on the three major clouds
| Platform | How Adam AI runs there |
|---|---|
| Microsoft Azure | GPU virtual machines or Azure Kubernetes Service in your own tenancy, alongside the Microsoft 365 estate most of our clients already run. Entra ID for sign-in, your existing Azure agreement for billing. |
| Amazon Web Services | GPU instances or EKS inside your own VPC, with private endpoints to your data and no public exposure. Suits clients already standardised on AWS or with US and global footprints. |
| Google Cloud | GPU instances or GKE in your project, with Google Workspace sign-in where that is your productivity suite. Strong fit for data-heavy and analytics workloads. |
In every case the account is yours, the region is yours and the models run privately. Adam AI is deployed and operated by Total Group under the same fixed monthly fee; cloud consumption is billed to you by the provider with no mark-up.
Containerised, so nothing has to be rebuilt
Six components, stable interfaces
Dashboard, 2nd brain, agent framework, skills depository, model layer and integration broker each run as containers. Any one can move or scale on its own.
Pull-only fleet updates
Every Adam AI box or cluster pulls signed updates from the control centre. The same release runs on premises and in the cloud.
Agents hold no keys
Credentials live in the vault, agents receive short-lived single-job permissions, and every call is logged. The security model is identical wherever it runs.
Backed up and portable
Nightly encrypted backups of knowledge, configuration and skills. Move provider or come back on premises without losing a day of training.
Let the numbers decide, not the vendor
Stay on the box while
One or two departments are live, token use is steady, and the workstation has headroom. This is the cheapest and most private way to run AI, and most clients stay here for the first year.
Add cloud when
Agent teams need to run around the clock across sites, a larger model is worth the cost for a specific job, or a department’s volume outgrows local GPU memory. We show you the crossover in your own usage data.
Never
Hand your training, knowledge and agent IP to a platform you rent. Adam AI scales inside accounts you control, and you can bring it back on premises at any time.
The technology we build on, and the people who deliver it
Adam AI runs on your own hardware first. When clients scale, we deploy the same platform into enterprise compute from the providers below, in the client’s own account.
Technology partners
- Claude by Anthropic
Frontier reasoning and the orchestrator behind Adam AI’s agent teams, with training off and your data held locally. - NVIDIA
GPU compute on the box and in the cloud, from RTX workstations and GB10 desk stations to data-centre instances. - Microsoft Azure
Private GPU hosting in your Azure tenancy, alongside your Microsoft 365 estate and Entra ID. - Amazon Web Services
GPU instances inside your own AWS account and VPC, with private endpoints to your data. - Google Cloud
GPU instances in your own Google Cloud project, with Workspace sign-in where that is your suite.
Ecosystem partners
- Managed service providers
Deliver Adam AI to the businesses whose IT you already run, with our team behind you. - Reseller partners
Resell, white-label or co-deliver private AI in your market. - Hardware and system builders
Build and supply Adam AI workstations and desk stations to our specification. - Become a partner
Tell us about your business and your clients: book a consultation.
Product names are the trademarks of their owners and are listed to show the platforms Adam AI is built on and deployed into.
The controls travel with the platform
Total Group holds Cyber Essentials Plus and IASME Gold and offers CREST-approved penetration testing. Deployments are built for UK GDPR with UK or EU data residency, and the principal consultant is an IASME-certified Data Protection Officer who can act as your fractional DPO. Clients working towards ISO 27001 get the evidence trail the platform already produces.