The Cambridge city region’s economy depends increasingly on access to secure, reliable and scalable compute capacity. Demand is growing across AI, life sciences, semiconductors and other data-intensive activities, bringing questions that once sat largely within specialist digital infrastructure discussions into much wider view.
What sort of compute capacity will Cambridge need over the coming decade? Where should it be located? What will it mean for land, power, water and connectivity? And how should these questions be approached in a place where infrastructure pressures, environmental constraints and quality of life are already central to public debate?
These issues are part of the wider challenge of how our city region sustains its position as a globally significant centre for science, technology and innovation; and they act as the premise for a piece of work we at Cambridge Ahead are kicking off on the topic.
A growing need, but an incomplete picture
One of the clearest conclusions from a round table convened by Cambridge Ahead of Members and key stakeholders is that future demand for compute infrastructure is very probably underappreciated, particularly in planning and infrastructure policy.
Demand is rising rapidly. What is less clear is how much of that demand will need to be met close to places like Cambridge, how much can be served regionally or nationally, and whether existing or planned capacity will be sufficient – with the prevailing view being that it will not. That uncertainty presents a practical challenge in its own right. It makes planning and infrastructure decisions harder, particularly where local authorities are being asked to weigh proposals against competing pressures on land, energy, water and community impact.
This is about understanding demand better from the user side: how compute, storage and network requirements are evolving, and what those changes imply in practice.
Different users, different needs
A useful way of framing the issue is to recognise that demand is not uniform.
Some local organisations require advanced compute capacity for research and development, large datasets and specialist applications. Others are regional users drawing on wider infrastructure networks. More general cloud demand is likely to be served elsewhere, provided that the connectivity is strong enough.
This matters because it suggests that a single model of provision is unlikely to be sufficient. The Cambridge city region may need a mix of local, near-local and remote capacity.
Low latency is one example. It is essential in some sectors, but not in all. For many research and knowledge-intensive users, the more pressing issues may be data sovereignty, security, resilience, the cost and speed of moving very large datasets, and assured access to high-end compute. The policy and planning response therefore needs to be more nuanced than a simple question of whether capacity should sit inside or outside Cambridge.
A planning issue as well as an infrastructure issue
These questions land in a planning context that is already complex. Cambridge faces well-established pressures around housing, employment land, transport, utilities and environmental capacity. Data centres add another layer to navigate.
They require land and infrastructure, but do not fit neatly into traditional planning categories. They may create fewer direct jobs than some competing land uses, which can make them harder to justify politically in areas where land is scarce. At the same time, their enabling value can be considerable, suppotying the wider innovation economy.
This points to the need for a clearer local framework for understanding their role. Data centres cannot sensibly be considered only on a site-by-site basis. Their relationship to fibre routes, grid access, water availability, transport infrastructure and strategic employment locations means they need to be considered as part of a wider spatial and infrastructure picture.
There is also a broader local and national question. If compute infrastructure is increasingly strategic to economic growth, then planning systems need a stronger evidence base and clearer policy signals to support decisions on where capacity should sit and on what terms.
Power, water and connectivity
If there is one issue that stands out most clearly, it is power. Electricity availability and connection timescales are widely seen as the main practical bottleneck.
That challenge is not unique to Cambridge, but it is particularly acute here given existing infrastructure constraints. It means that future discussions about compute provision cannot be separated from wider questions about grid capacity, energy planning and system coordination.
Water is also an important issue, especially in a water-stressed region. But here, too, the debate needs to be grounded in current evidence rather than outdated assumptions. Data centre design has evolved, including greater use of closed-loop and liquid cooling systems. However, that does not remove the need for careful scrutiny of resource impacts and trade-offs.
Connectivity is the third essential part of the picture. Fibre and long-haul links help determine what can be hosted locally, what can be accessed remotely, and how resilient the region is to wider infrastructure constraints. If there are gaps in fibre provision to science parks, employment sites or growth areas, those can become economic constraints in their own right.
Taken together, this suggests that compute capacity should be understood as part of a wider enabling infrastructure system rather than as a stand-alone land use issue.
A sector that is changing quickly
Another reason for caution in policymaking is that the sector itself is evolving at pace. The traditional mid-sized data centre model appears to be giving way to a more polarised landscape, with smaller on-premises or edge facilities on one side and larger campus-scale or hyperscale solutions on the other.
Co-location remains important, while modular approaches are also gaining attention. The shell of a building may remain in place for many years while the technology inside it is refreshed on a much shorter cycle. That has implications for planning, design and long-term adaptability.
Data centres should not be treated as static assets. Their operational profile and technical requirements are likely to evolve, and planning frameworks need to be capable of responding.
Quality of life and public legitimacy
In Cambridge, these issues cannot be discussed without taking quality of life seriously – particularly here at Cambridge Ahead, where quality of life runs throughout our work. Environmental performance, visual impact, noise, servicing, water use and wider public acceptability all matter hugely.
Public understanding of data centres does not always reflect the reality of modern facilities. Some of the sector’s reputation still draws on older models of development and more general anxieties around AI and digital growth. Yet contemporary facilities are often more efficient and less intrusive than is commonly assumed.
Embedding better design, greater transparency and a clearer account of how data centres fit within local priorities is key. Heat reuse is one example. It will not be appropriate everywhere, but in some locations it may offer real value if considered early enough. More broadly, the principle is that environmental performance and community benefit need to be designed in from the outset.
An opportunity for the Cambridge city region
Cambridge city region is well placed to contribute to a more informed discussion about compute infrastructure. It combines world-leading science, advanced industry, major compute users and a strong public interest in how growth is managed. That gives it both a stake in the outcome and a distinctive perspective to offer.
The challenge now is to develop a clearer local view: a stronger evidence base on future demand; a more explicit planning and infrastructure framework for weighing compute capacity alongside competing pressures; and a more confident public narrative about the role this infrastructure plays in supporting research, innovation and growth.
These are focused questions rather than abstract ones. What sort of demand is likely to arise over the next five to ten years? Which needs require local or near-local provision, and which can be met elsewhere? How should data centres, fibre, grid capacity and water be reflected in planning and infrastructure strategies? And what does a good model look like for Cambridge in terms of sustainability, design quality and local benefit?
Cambridge Ahead is convening planners, compute users, infrastructure experts and the wider public and private sector to tackle these questions from a Cambridge perspective, and develop practical insights and data that can help the city region – and wider UK – prepare to meet the data demands of the future in a scalable, sustainable and inclusive way which can keep our city region in the vanguard of innovation. We will share more updates on this work in the coming weeks; if you would like to learn more, please get in touch.
