techUK welcomes the publication of the Compute Roadmap
19 Jul 2025 04:26 PM
techUK welcomes the publication of the Compute Roadmap as a crucial step in delivering the £1 billion pledged in the Spending Review to boost the UK’s compute infrastructure and AI capabilities.
This ambitious roadmap, underpinned by delivery timelines beyond 2030, shows that UK Government is serious in its ambition to prioritise transformative research and deliver innovation with real-world impact in the UK.
It acknowledges that the current compute ecosystem is fragmented and facing a capacity overload when the many systems reach end of life over the next 18 months. This is an issue further exacerbated by inconsistent compute policy over the last 12 months that saw projects such as the exascale machine in Edinburgh halted.
Addressing this challenge, the roadmap seeks to establish a single national compute ecosystem, supported by centres of excellence, robust user support, and coordinated programmes for data, software, and skills.
It also attempts to connect other key aspects of the UK’s innovation landscapes – and several of the 50 recommendations in the AI Opportunities Action plan. Most notable are the announcements on the AI Growth Zones and Sovereign AI, which you can read more about below.
There are also numerous announcements techUK has summarised below. These include initiatives such as National Supercomputing Centres (NSCs) to open access, and changes to how public compute will be allocated.
However, while the direction is welcome, there remains a lack of detail on how the roadmap will bring together the dispersed talent, infrastructure, technologies, and governance needed to make the UK globally competitive in AI.
As often repeated through the outpouring of policy announcements this year: Further clarity and action is needed.
What does the roadmap say?
The 10-point plan is underpinned by four strategic objectives. These are:
Building a modern public compute ecosystem
- Invest up to £2 billion to deliver a diverse, joined-up and user centred compute ecosystem including over £1 billion to expand the AI Research Resource (AIRR) 20x by 2030 and up to £750 million for a new national supercomputer service in Edinburgh.
- Establish National Supercomputing Centres to curate datasets, build software assets and deliver a skills pipeline to support a broad range of users to access and utilise computing power.
- Partner with like-minded countries and computing centres to expand access to a broader range of research infrastructure, facilitating collaboration and skills and knowledge exchange.
Putting compute to use, powering innovation across the economy
- Introduce a refreshed allocation model that targets compute toward the UK’s highest-impact research and innovation priorities, backing mission-driven research and strategic national goals.
- Guarantee dedicated compute access for the Sovereign AI Unit and the AI Security Institute, ensuring the UK’s core AI functions have the capacity they need to lead.
Building AI infrastructure to keep the UK at the cutting edge of AI development
- Deliver large scale AI Infrastructure via AI Growth Zones (AIGZs) across the UK, ensuring the UK has the capability to support AI training and inference workloads.
- Explore new models for delivering the energy infrastructure that powers AI, including through sustainable solutions such as renewables, advanced nuclear, and innovative grid solutions.
- Harness AIGZs to deliver both national and local benefits across the economy, serving as platforms for innovation, adoption and collaboration through which AI capabilities are developed, trialled, deployed and scaled.
Creating sovereign, secure and sustainable capability
- Support British companies to develop sustainable and secure compute technologies, using the AIRR and AI Growth Zones to provide access, testbeds, and opportunities to scale.
- Establish compute as a priority area for the UK Sovereign AI Unit, supporting research and innovation in new computing paradigms to help British companies grow into global leaders.
techUK Summary
Acknowledging diversity in compute needs
- For AI compute, it does not take a stance on whether the UK will prioritise training or inference. Procurement for AIRR will follow a blended model, combining both purpose-built AI supercomputers and cloud-based compute to deliver flexibility, performance, and resilience. However, at this stage it remains unsure if this will be an extension of the current AIRFED project.
- A focus on interoperability and portability, with systems designed with a focus on user flexibility, ensuring that workloads can move easily between on-premise, cloud, and hybrid environments
- Commitment to public-private delivery and supplier diversity fostering resilience and competitiveness in the UK ecosystem.
AIRR proves the testbed for some of the most innovative ideas to join up compute.
- Launch a programme to identify opportunities for Public-Private Partnerships that could coordinate the delivery and operation of new AIRR supercomputing capacity
- Commence procurement of cloud compute to complement our dedicated AI supercomputers ensuring the AIRR offers both high-performance, purpose-built infrastructure and flexible, on-demand resources to meet the full range of user needs.
- Deliver ‘AIRRPORT’, which will provide a single front door to the AIRR service, enabling users to submit jobs, move workloads, access tools, and manage data across AIRR systems ensuring accessibility and interoperability.
The way compute will be allocated is also changing
-
They will introduce a refreshed allocation model that targets compute toward the UK’s highest-impact research and innovation priorities, backing mission-driven research and strategic national goals. This includes allocation for AI Security Institute (AISI) and Sovereign AI Unit. However, more detail is needed.
It is not just about infrastructure
Intention to connect emerging ‘novel’ technologies – but light on details
- The roadmap promises support for novel compute technologies through strategic programmes and testbeds in NSCs.
- AIRR capacity will be reserved for emerging areas including quantum, new chip architectures, and next-gen software stacks, and linking to the frontier technologies in the Industrial Strategy
- National Supercomputing Centres to act as a trusted proving ground to test, validate, benchmark, and deploy emerging technologies.
- The most successful UK-developed technologies will be turned into commercial, at-scale deployments of AI Growth Zones – this boldly sets a new role for the AIGZ to play a key role in helping UK emerging tech businesses scale
- However, concrete plans and timelines remain sparse. More information is needed on how emerging UK technologies will be pulled through to scale.
What does it say about AIGZ?
- AIGZs will be the primary delivery mechanism for AI infrastructure, with zones expected in Scotland and Wales, and decided, at least in part, on energy capacity and innovation needs.
- AIGZs look to support UK capability by providing platforms for secure, scalable compute “physically located in the UK” that is “not just rented from abroad” to help British firms access infrastructure and markets.
- They will also play a key role in energy innovation, with a focus on advanced nuclear, renewables, and new grid models. The AI Energy Council is referenced in playing a role here – but light on details.
- There is a key role for AIGZs helping emerging tech businesses scale, where the most successful UK-developed technologies will be turned into commercial, at-scale deployments within these zones. This process is referred to as a “compute bridge”
What Does it Say about Sovereign AI
- Sovereign AI is a critical piece of this roadmap, described as part of the "governments core strategic AI functions."
- It will prioritise targeted public intervention that can support UK leadership, scale national champions, and secure long-term strategic advantage - this means novel, IP-rich capability advancements, emerging technical and market trends in AI.
- Compute will be allocated as a priority area: guaranteeing dedicated compute access for the unit to ensure the “UK’s core AI functions have the capacity they need to lead.”
- It will support key priority areas “whether in emerging paradigms, high-impact sectors like healthcare or defence or through backing the most promised UK-based startups.”
- Sovereign AI already has initiatives up and running: OpenBind (a project to create the world’s largest open protein-ligand dataset for AI-native drug discovery), partnerships with companies to “gain a UK ‘stake’ in frontier AI development” and ensure “strategically important companies are invested in the UK” and Encode (a programme to get top global talent in UK labs as part of AI for Science).
Role of commercial compute
- As techUK has argued, the UK must recognise the central role of private sector infrastructure in meeting demand.
- The roadmap rightly acknowledges that most of the UK’s compute capacity will remain privately owned, with public infrastructure supporting national strategic needs.
techUK concluding remarks - more to be done
techUK is supportive of the bold ambition set out in this roadmap to deliver the compute necessary for both scientific excellence and AI capability in the UK. We also welcome that this is underpinned by appropriate funding to build the necessary infrastructure. And critically, it is backed by a timeline to create momentum and stability. But clarity and details on how the recommendations will be delivered are still lacking – surprising, since the role of a roadmap needs to be clarifying delivery.
This is obvious in its ambition to recognise the future of compute as a holistic ambition beyond AI – something that techUK has previously called for. We previously emphasised that compute needs go well beyond AI, spanning scientific research, health, and incorporating the next generation of compute such as quantum. The roadmap does recognise this, and the creation of a “Compute Bridge” in the AIGZ to deliver alongside existing initiatives is a bold commitment to deliver the UK’s leadership on emerging and transformative technologies beyond AI. But specifics are sparse, with the timeline stating “integration of novel computing technologies – including quantum, neuromorphic, and AI-hybrid systems – into public infrastructure” by 2027.
There is still more to be done to help develop the role of compute beyond public infrastructure, which will be critical to meet demand. For example, it is envisioned that commercial, cloud-based compute will be part of the AIGZ, and needed to support researchers and businesses alike, but the role of Government in helping secure or potentially subsidise that access is limited in this roadmap. Furthermore, there is little said in timelines and delivery for international partnerships for compute access.
It recognises the role of novel chip architectures but fails to deliver any plan to ensure reliable chip supply planning to avoid bottlenecks.
Importantly, it is still not clear where the Sovereign AI Unit will focus its programming on next, or where exactly the AIGZs will be in Scotland and Wales, though these will likely be announced in due course.
It is also undeniable that digital sovereignty and the AIGZs are connected: They state that, in an ideal end state, AI Growth Zone will be powered by a full UK-designed compute stack – from chip to system to software – demonstrating what sovereign capability truly looks like. As such, this is an area where further join up between government and UK industry would be welcome.
With this in mind, techUK remains open and willing to work with members and government to explore how to turn this roadmap into action, keeping innovation and research excellence at the heart of delivery.