SOFIAH NICHOLE SALIVIO

News Editor

OneAdvanced has signed a Memorandum of Understanding with Cosine to join the Lumen Sovereign Coalition, which is working on what the companies describe as Britain’s first fully sovereign frontier AI model.

The agreement brings OneAdvanced into a group of companies and institutions helping to shape the model for regulated and essential service sectors. Its role will include contributing use cases, operational requirements, testing scenarios and deployment feedback from industries where its software is already in use.

Those sectors include healthcare, government, justice, education, care and business services. OneAdvanced will use that experience to help define how the model should operate in settings where security, governance and operational rules can limit the use of foreign-hosted AI systems.

Cosine, a London-based AI laboratory, is developing the Lumen Sovereign model with support from a wider industry coalition. Other members include BAE Systems, Babcock International Group, BT, HSBC, Lloyds Banking Group, LSEG, NatWest Group, PwC, The Alan Turing Institute and Vodafone.

The project is backed by the UK Government’s Sovereign AI Programme, which has committed GBP £500 million. The model is due to be trained using the Isambard supercomputer in Bristol.

OneAdvanced’s involvement follows its work on a healthcare large language model developed with NVIDIA and trained on NHS primary care data. That system was designed to improve primary care triage, and the company is positioning it as evidence of its experience in applying AI to sector-specific operational needs.

In healthcare, OneAdvanced says its software supports more than 40 million NHS patients across NHS 111, GP practices and trusts. It also works across local government, education, legal services, transport and supply chains.

Sector input

The tie-up reflects a broader push in the UK technology sector to develop AI systems that can be hosted, trained and audited within the country. For organisations operating in heavily regulated environments, questions about where a model runs, what data it is trained on and who can review its operation have become increasingly important in procurement and deployment decisions.

Cosine says the coalition is intended to ensure the model is shaped by organisations that will use it in real operating environments, rather than by technical teams alone. That means feeding in practical constraints from sectors such as healthcare, local government and legal services.

Cosine framed that goal around control and auditability.

“We’re building Lumen Sovereign because we think the UK needs frontier AI it can actually control: where the model runs, what data trained it, who can audit it. That only works if we build it alongside the organisations running the country’s real, regulated infrastructure. OneAdvanced helps us achieve that goal. Their experience with the NHS, local government, education, and thousands of legal practices means they bring insights into the operational constraints a sovereign model needs to actually be useful,” said Alistair Pullen, Co-founder & Chief Executive Officer of Cosine.

Practical focus

For OneAdvanced, the move also underlines its argument that the next stage of AI adoption in critical services will depend less on general model performance claims and more on whether systems can fit existing workflows, governance standards and data-handling requirements.

The company has been advocating for AI systems designed around UK data residency and domestic control over infrastructure. Its software footprint in public and regulated services gives it a route into organisations that may be cautious about deploying overseas-hosted models in sensitive settings.

Simon Walsh set out that position in his remarks on the coalition agreement.

“We have long advocated sovereign AI as the safest and most secure way for UK organisations to leverage the transformative power of AI in their flow of work. Through our work with NVIDIA to develop the UK’s first sovereign healthcare large language model, trained on NHS primary care data to improve patient triage outcomes, we have seen first-hand the importance of combining advanced technology with deep sector knowledge, robust governance and a clear understanding of the workflows organisations rely on every day. That is why we are delighted to join the Lumen Sovereign coalition.

“Organisations in the UK’s highly regulated and essential service sectors need AI that is not only powerful, but secure, trusted and designed around the realities of the environments in which it will operate. We look forward to working with Cosine and the wider coalition to help turn that ambition into practical capability,” said Walsh.

The club will continue working with AWBS, which will remain as the club’s Official Substitution Sponsor until the end of the 2027/28 season.

Lee Barton, head of revenue at Oxford United, said: “We’re thrilled to announce this enhanced multi-year agreement that will see AWBS continue as our Official Substitution Sponsor.

“We look forward to developing the relationship further during the 2026/27 season.”

AWBS is one of the largest suppliers of building and landscaping materials in Oxfordshire and Wiltshire, with three branches.

The company also has extensive showrooms, a fleet of delivery vehicles, and a dedicated landscape installation service.

SOFIAH NICHOLE SALIVIO

News Editor

SHI has joined the Tokenomics Foundation as a founding member. The group is part of the Linux Foundation.

The move makes SHI one of the early members of an industry body focused on standards and best practice for the economics of artificial intelligence infrastructure. The foundation is developing frameworks to help organisations track the cost and value of AI systems as use of generative AI and agentic tools expands.

Businesses are under growing pressure to show returns on AI spending while keeping control of usage, governance and operating costs. That has led to closer scrutiny of how computing resources are used and whether AI deployments are delivering measurable business value.

The Tokenomics Foundation was set up to advance open collaboration around token production, token consumption and AI value management. Its work draws on FinOps principles, linking product design, engineering, finance, operations and governance through a shared understanding of AI cost and value.

For SHI, the membership extends an existing role in adjacent areas of technology cost management. It is also a Premier member of the FinOps Foundation, which is also hosted by the Linux Foundation.

SHI said it would work with other members to help define frameworks and best practices for organisations adopting and scaling AI. The effort is intended to improve how AI-generated value is measured, distributed and governed.

One issue for companies deploying AI is that consumption-based pricing can make spending difficult to predict. As use spreads across departments, executives often face questions about whether systems are delivering enough value to justify the cost, especially when oversight is split across technical, financial and operational teams.

Shane Cronin outlined SHI’s view of that shift. “As organizations move from experimenting with AI to operating it at scale, the ability to understand how token consumption drives cost, efficiency, and business value is becoming critical,” said Shane Cronin, Head of FinOps & ITAM Services, SHI.

“SHI has long believed that strong governance requires collaboration across the disciplines of IT Asset Management, FinOps, and now AI economics. We are committed to helping shape the industry standards, frameworks, and best practices organizations need to manage technology and AI investments with confidence. Joining the Tokenomics Foundation strengthens our ability to help customers turn AI adoption into measurable business outcomes while contributing to the evolution of this rapidly emerging discipline,” said Cronin.

Industry standards

The foundation’s backers argue that common methods are needed because organisations often lack a consistent way to measure value from AI spending. That challenge has become more acute as AI tools move beyond pilots into broader operational use.

J.R. Storment, Executive Director of the Tokenomics Foundation, said SHI would add practical experience to the group’s work. “SHI has established itself as a key voice and member across the technology value community, including the FinOps Foundation, ITAM Forum, and now Tokenomics Foundation,” said J.R. Storment, Executive Director, Tokenomics Foundation.

“SHI brings a depth of customer and practitioner experience to the table. As organizations work to measure value from AI spend, Tokenomics Foundation gives them a neutral, community-built discipline to do it, and we look forward to the insight SHI will add as we define AI value management together,” said Storment.

SHI describes itself as a USD $16 billion technology solutions provider serving more than 17,000 corporate, public sector and academic customers worldwide. The business employs more than 7,000 people.

Its entry into the foundation reflects a wider effort across the technology sector to create AI governance models more closely tied to spending discipline. As companies seek to move from experimentation to routine use, industry groups are trying to establish shared definitions and measurement standards that can be applied across suppliers, customers and internal teams.

That work is likely to shape how organisations assess the financial case for AI projects, especially where usage is measured in tokens and infrastructure costs can shift rapidly with demand. The Tokenomics Foundation’s central aim is to build common frameworks and best practice that support responsible, efficient and business-aligned AI adoption.