Argyll launches UK sovereign AI cloud for regulated firms

Argyll Data Development has launched a sovereign AI inference cloud in the UK for organisations that want AI workloads to remain under UK control.

Built with SambaNova, the platform is designed to let customers run production AI systems without relying on foreign-owned hyperscale cloud infrastructure. It is aimed at sectors with strict regulatory or security requirements, including defence, healthcare and finance.

The launch comes as more companies move artificial intelligence projects from trial stages into day-to-day operations. That shift has increased scrutiny of where data is held, who controls the underlying systems and how much it costs to run AI services at scale.

Argyll said its platform keeps infrastructure, models and operations within UK jurisdiction, using SambaNova hardware and software as the basis for the service.

Sovereignty focus

Argyll is positioning the launch around sovereignty rather than raw computing scale. In practice, customers are being offered a UK-based environment for inference, the stage at which trained AI models process live requests and generate outputs.

Many organisations currently rely on large international cloud providers for that work. Argyll argues that approach can raise concerns about data sovereignty and accountability, particularly for institutions handling sensitive information or operating under tight compliance rules.

The company also highlighted the costs of conventional AI infrastructure. GPU-based deployments are widely used for AI processing, but they can require significant power and cooling, making them harder to install in existing data centre estates.

At the centre of the platform is SambaNova's Reconfigurable Data Unit architecture running SambaManaged software. The companies said the system operates at about 10kW per rack, allowing it to be deployed in existing UK data centres without the cooling demands often associated with traditional GPU systems.

Argyll said the service hosts open-source models including Minimax and can run at speeds of up to 400 tokens per second in a UK-resident environment. It said that efficiency profile is intended to make live AI deployments more financially practical for uses such as customer operations and fraud detection.

Distributed design

The platform has been built as a disaggregated system, with compute, storage and networking able to operate across multiple UK locations while functioning as a single inference layer. Argyll said the design gives customers added resilience and flexibility, particularly in regulated or security-sensitive settings.

That structure reflects a broader concern in the UK technology sector about dependence on overseas infrastructure for strategic digital services. Debate over AI policy has increasingly expanded beyond model development and safety to include the ownership and location of the systems that run commercial and public sector workloads.

For businesses in sectors such as finance and healthcare, the issue is not simply where data is stored. It also includes who has operational control over the platform, which legal jurisdiction applies to the infrastructure, and whether systems can be audited to local regulatory standards.

Peter Griffiths set out Argyll's argument for a stricter definition of sovereignty.

"Sovereignty in AI is not a label you can apply to a contract or a colocation agreement. It is a condition that has to be demonstrated: who is accountable, where the infrastructure sits, who controls the intelligence layer, and whether all of that aligns with the expectations of the society being served. Our platform satisfies those conditions. We are building the standard that others should be measured against," said Peter Griffiths, chairman of Argyll Data Development.

SambaNova presented the partnership as an alternative to mainstream AI infrastructure choices.

"As organisations scale AI, many are defaulting to GPU infrastructure without fully accounting for long-term cost, energy and operational complexity. Our work with Argyll provides an alternative, enabling high-performance AI inference that is more efficient, deployable and aligned with sovereignty requirements," said Jude Sheeran, managing director, EMEA, SambaNova.

Argyll has previously outlined plans for renewable-powered digital infrastructure in Scotland, including its Killellan AI Growth Zone in Argyll. Its wider strategy is to support secure AI growth in the UK through domestic infrastructure tied to local energy sources.

The launch gives that strategy a commercial service that organisations can use now to keep AI operations within UK jurisdiction, at a time when the balance between access to advanced AI and control over the systems behind it is becoming a more pressing issue for both industry and government.