Public sector prepares cloud expansion to strengthen national AI research
A new cloud procurement aims to expand national AI computing capacity, signalling a shift towards shared digital infrastructure across UK public services.
A new cloud procurement aims to expand national AI computing capacity, signalling a shift towards shared digital infrastructure across UK public services.
Follow Tenderlake on LinkedIn for concise insights on public-sector tenders and emerging procurement signals.
In December 2025, the Department for Science, Innovation & Technology signalled plans to expand the UK’s shared AI infrastructure. Through the AI Cloud Compute Expansion project, it is engaging the market for a scalable cloud solution to increase the AI Research Resource’s computing capacity. The aim is to strengthen national AI infrastructure, offer flexible access to compute and support innovation across a range of critical sectors.
Published on 10th December 2025, the notice sets out a concise ambition: to procure a scalable cloud environment that boosts the computing power available to the AI Research Resource and underpins national AI infrastructure.
According to the notice, the planned cloud solution is intended to:
Expanding compute for a shared research resource is about more than raw processing power. Cloud environments can match capacity to demand, support varied workloads and open access for organisations that could not fund large systems alone. For the public sector, that can mean faster experimentation and shorter routes from research to service design.
As a Prior Information Notice focused on market engagement, the document does not yet specify architectures, volumes or pricing. It instead signals a shift towards flexible cloud compute as a core tool for government‑backed AI research.
The move sits alongside other UK efforts to underpin research with cloud‑based platforms. On 10th December 2025, UK Biobank issued a Prior Information Notice for its Cloud Infrastructure Procurement, covering cloud infrastructure and related services to build a new health research analysis platform, RAP2.0, and to improve data accessibility and security for researchers.
In October 2025, UK Research and Innovation set out plans for the Next National Supercomputing Service, focusing on supercomputing hardware at the University of Edinburgh. It underlines how large on‑premises systems and flexible cloud services are being developed in parallel as complementary parts of the research landscape.
In September 2025, the Advanced Research and Invention Agency sought Managed IT Service Proposals to deliver secure, scalable IT support, strengthen a cloud‑based technology ecosystem and align with government security frameworks. Together, these moves point to shared, cloud‑hosted platforms becoming the backbone for UK research and innovation.
Other public bodies are buying in AI expertise and tools for operational use. In defence, a July 2025 Prior Information Notice from the Defence AI Centre for AI Technical and Data Science Support seeks leadership and technical support across AI technology, assurance, generative AI and battlespace advantage, plus data science services to enhance defence capabilities.
Frontline services are exploring generative AI and guidance for staff. Dublin City Council’s November 2025 consultation on AI Solutions for Local Government centres on generative AI expertise and integration with Microsoft technologies for customer service and local administration. In Wales, Social Care Wales is commissioning Digital Resources for AI in Social Care to create concise English and Welsh materials promoting safe, responsible use of AI by professionals.
Infrastructure operators are reshaping how they manage assets and cyber risk. Network Rail’s October 2025 Digital Twin Platform Procurement seeks a software‑as‑a‑service platform that integrates rail asset data, supports advanced analytics and complies with security standards, accessible to staff and authorised contractors. In August 2025, National Grid Electricity Transmission started market engagement on Cyber‑Attack Path Analysis products and services, emphasising automation and rich data sources for proactive cyber defence.
The UK is not alone in treating compute as strategic public infrastructure. In October 2025, Germany’s federal government launched an AI Platform Procurement to put in place a framework for platform‑as‑a‑service offerings on a secure cloud that supports AI applications and integrates with existing systems. Denmark’s Syddansk Universitet is establishing a GPU as a Service framework to provide secure, scalable GPU resources for data‑intensive research via its UCloud environment.
Research organisations are also investing in dedicated clusters. A December 2025 framework from Fraunhofer IIS for HPC and AI Cluster Components will deliver a directly water‑cooled GPU computing cluster for training generative AI, plus associated infrastructure and maintenance. In the same month, Vilniaus Gedimino technikos universitetas began a Data Processing Platform Expansion to extend a high‑performance data platform with CPU and GPU subsystems, AI computing nodes and extra storage.
For now, the AI Cloud Compute Expansion notice offers only a high‑level sketch of the UK’s next step on AI infrastructure. The forthcoming specification – covering performance, security and how access will be managed – will determine how far a shared cloud resource can widen participation in advanced AI research and how it complements supercomputing and sector‑specific projects. Suppliers, researchers and policymakers will be watching the next stage of the procurement to see how that balance is struck.
Follow Tenderlake on LinkedIn for concise insights on public-sector tenders and emerging procurement signals.