Comparison
Sovereign AI vs AWS
AWS is strong on infrastructure sovereignty. But the real test is whether you own the model, not just where it runs.
By James Drayson
In short
AWS is excellent for infrastructure sovereignty, regions, residency, and its European Sovereign Cloud, but with Bedrock the model itself is still rented from a third party. Locai is sovereign at the model layer: you own the weights, the IP, and the update cadence, not just the data centre.
Infrastructure sovereignty vs model sovereignty
AWS genuinely solves a hard problem: keeping your data and compute in the right jurisdiction with strong controls. For infrastructure sovereignty it is a credible, mature option, and Locai models can run inside your AWS tenant.
But infrastructure is only one layer. With a hosted model service the model you actually use is owned by someone else, frozen between releases, and changeable or withdrawable under you. Running a rented model in a sovereign region does not make the model sovereign.
Model sovereignty is the missing layer: holding the weights, controlling the training data, and deciding the update cadence. That is what Locai adds, and it is the layer that determines whether the capability is truly yours.
Own with Locai vs rent the model via AWS
| Own with Locai | Rent the model via AWS | |
|---|---|---|
| Infrastructure residency | Your choice, incl. inside AWS | Strong (Sovereign Cloud) |
| Own the model weights | Yes | No, rented via the model service |
| Model can be deprecated | Never, it's yours | Yes, by the model provider |
| Trained on your domain | Yes, post-trained on your data | General-purpose foundation models |
| Improves continually | Yes | Frozen between releases |
| Cost model | Fixed, owned asset | Per-token / per-call |
| Full weight & data audit | Yes | No |
| Air-gapped option | Yes | No |
What this looks like with Locai
Sovereignty stops being a contractual promise when the machine is standing in your own building. Here is what that looks like in practice.
Locai Labs builds Locai One, an on-prem AI appliance. It is one machine that arrives with everything already in it: the hardware, our open-weight Locai Juno models, and Locai OS, the operating system that serves the models and handles users, access and monitoring. You plug it into a mains socket and your network, and your team is working in about 15 minutes. No cloud account, no per-token bill, and nothing leaving the building.
The reason a data-centre-class model fits in a box on your floor is SPACE, our compression algorithm. Instead of asking how much of a model can be cut while keeping it broadly similar, SPACE asks what the model needs to be good at, preserves the subnetworks behind those capabilities and strips back the rest. The result is a smaller specialist rather than a shrunken generalist, tuned to the exact hardware it ships on.
Locai One starts at £29,950 for a team, and Locai One Pro at £49,950 for an organisation, bought once and owned outright. Both run air-cooled on standard mains power and work fully air-gapped. Any compatible open-weight model runs alongside Juno, and if you need a model trained on your own proprietary data we can post-train one and deploy it on the same machine.
Frequently asked questions
Can I run Locai inside my existing AWS environment?
Yes. Locai models can be deployed inside your own AWS tenant, so you keep your cloud relationship and residency while gaining model ownership on top.
Isn't AWS already 'sovereign'?
AWS provides strong infrastructure sovereignty. The gap is model sovereignty: with a hosted model service you still don't own the weights or control the update cadence. Locai closes that gap.
Does this mean I should leave AWS?
No. This isn't AWS versus Locai at the infrastructure layer, you can keep AWS and add model ownership with Locai. The point is to be sovereign at the model layer, not just the data-centre layer.
Book a sovereign AI briefing
A 30-minute session on owning your model: deployment options, the data path, and a clear cost range for your use case.