AI at the Edge: The Backbone of Digital Infrastructure
The symbiotic relationship between AI and edge computing has continued to evolve from emerging technology to essential infrastructure, fundamentally reshaping how industries can operate across the globe.
Back in 2024, we explored the challenges facing builders of digital infrastructure as they created the massive data engines powering the AI paradigm-shift – particularly the mega-data centres hosting training systems for generative AI platforms.
While attention remains fixed on these behemoths, the true revolution is happening at the edge, where AI meets real-world applications across every industrial and commercial sector.
Beyond Generative AI: The Diverse Ecosystem of Industrial Applications
The narrative around AI has matured significantly since 2024. While generative AI platforms continue to capture headlines, the practical implementation of AI technologies spans a far more diverse ecosystem.
As Duncan Clubb, Senior Partner at Cambridge Management Consulting, notes, "Too much emphasis has been given to prominent large language models, when in reality, the market requires a more diverse model for deploying infrastructure that supports real-world applications."
Today we're seeing this diversity play out across multiple sectors. Industrial and manufacturing applications leveraging AI for optimisation have moved beyond early adoption into mainstream implementation. Gartner now predicts that 75% of enterprise data will be processed at the edge by the end of this year, a dramatic increase from just 10% in 2018. The Financial Services industry is also beginning to use edge technology for real-time fraud detection. This shift reflects a fundamental transformation in how businesses approach data processing and analysis.
Consider the manufacturing sector, where AI systems connected to production line sensors now control processes in real-time, improving efficiency and reducing waste. One compelling example highlighted by Clubb involves the automobile industry's use of specialist adhesives: "An AI platform has been in use to reduce the amount of glue used without compromising the efficacy of the bond. This may sound trivial, but the quantities used globally mean that even small proportional savings can amount to huge monetary savings."
Similar applications exist across healthcare, transportation, retail, and agriculture, where edge-based AI is providing unprecedented operational insights and efficiencies. According to recent projections, spending on edge computing technologies will reach $378 billion by 2028, demonstrating the critical importance these technologies have assumed in our digital infrastructure.
The Evolution of Edge Solutions
While large enterprises have leveraged AI for process optimisation for years, many smaller businesses have been excluded due to high barriers to entry. The ideal solution would be cloud-like services for AI-driven applications, but traditional cloud offerings from providers like Amazon, Microsoft, or Google often prove unsuitable for industrial applications.
"For most applications, the reason is either the amount of data that needs to be processed – it costs a fortune to transport the humungous amounts of data generated in production lines to where it could be processed – or the network latency is too long," explains Clubb. "Real-time control of industrial processes requires extremely fast networks."
The solution lies in distributed cloud-like infrastructure, positioning processing power near the companies or users generating or consuming data. This defines the current direction and rationalisation for edge computing.
"For me, it's quite simple," Clubb states. "The edge is where data processing needs to happen. That need is either defined by the sheer quantity of data that needs to be processed or the latency requirements."
The Edge Landscape: Smaller, Distributed, and Powerful
The infrastructure requirements for edge AI differ significantly from those of large language model training platforms. Edge AI data centres can be smaller and easier to build than their hyperscale counterparts, and they typically don't require the extreme power densities that training systems demand.
However, the key difference is quantity – we need many more of them. As Clubb observes, "Having access to an edge data centre within 20-40 km will normally be sufficient for many applications, but this means that we will need to build possibly hundreds of new (small) data centres to cover a country the size of the UK."
This presents both challenges and opportunities. According to research from Vertiv, the densification of computing workloads is driving innovation in power and cooling solutions. "AI applications demand increasingly efficient systems to manage the power requirements and thermal output of high-performance computing. Inferencing at the edge provides critical benefits such as reduced latency and enhanced security, making it an essential strategy for managing AI workloads efficiently in edge deployments."
The integration of 5G networks has further accelerated this trend, with edge deployments now capable of reducing latency to under 2 milliseconds – crucial for applications like autonomous vehicles and smart healthcare systems.
Meeting the Power Challenge
The exponential growth in data centre deployment brings with it significant energy demands. In 2025, the data centre industry faces intensifying power transmission challenges that threaten to delay development timelines.
"Energy availability is becoming a major concern as the demand for compute capacity grows and power densities increase," notes Vertiv's assessment of 2025 trends. "At the edge, this challenge is particularly pronounced due to distributed locations with varying access to power. AI applications further complicate this issue, as they require consistent and scalable energy sources to maintain operations."
These challenges are driving innovation across the sector. Companies are increasingly adopting renewable energy integrations, high-efficiency power systems, and alternative energy sources. For edge deployments specifically, local energy solutions like microgrids and battery storage systems are gaining traction to ensure uninterrupted operations even in remote or underserved areas.
Notably, 2025 has seen an acceleration of small modular reactor (SMR) announcements, with the total gigawatt capacity likely to double by year-end. "Nuclear power is emerging as a preferred solution to meet growing energy demand. As traditional power grids struggle to keep pace, the sector is exploring both traditional large-scale nuclear power and small modular reactors (SMRs)."
Building for a Distributed Future
The construction sector has responded robustly to these evolving demands. What started as a relatively niche sector a few years ago has transformed into a powerhouse of construction activity. Data centres continue to drive significant growth in nonresidential construction planning, with these projects contributing to a 19% increase in planning activity since December 2023.

The growth shows no signs of slowing. J.P. Morgan estimates that spending on data centres could add between 10 to 20 basis points to U.S. economic growth in 2025 and 2026. This growth extends beyond the United States, with global players making substantial investments in core and edge infrastructure.
The major tech giants companies are likewise doubling down on their data centre investments and exploring new energy solutions. Amazon is planning to invest $150bn over the next 15 years on infrastructure to handle the expected demand for artificial intelligence and other cloud computing needs.
Recently, it was reported that Google has signed its first corporate geothermal energy agreement in Taiwan with Baseload Capital, adding 10 megawatts of continuous power to support its local data centres.
Announcements such as these show market-wide confidence in the projected growth in demand, as well as growing interest in renewable energy sources.
Conclusion: The Dual Future of Data Infrastructure
As we prepare for Cambridge Tech Week 2025 (September 15-19), where ‘Seizing the AI Advantage’ will be a central theme, it's clear that the future of digital infrastructure depends on both centralised and distributed computing models working in harmony.
Duncan Clubb summarises this dual approach eloquently: "For the data centre industry, I think that is just as exciting as the need to build the behemoth data centres at the core – the truth is, we need both."
The AI industry is now a reality demanding immediate attention and action. As businesses navigate this rapidly evolving landscape, those who understand and leverage the symbiotic relationship between edge computing and artificial intelligence will be best positioned to thrive.
Cambridge's deep-tech ecosystem is at the forefront of this revolution, developing innovative solutions to address the challenges of power, cooling, and infrastructure that accompany the growth of AI and edge computing. The conversation at Cambridge Tech Week 2025 will undoubtedly shape how industries approach these opportunities in the years to come.
For organisations looking to harness the power of edge AI, the message is clear: the future is distributed, the future is intelligent, and the future is now.
If you would like to know more about our Data Centre, Edge and Cloud services, please
get in touch with Duncan by email or use the Contact Form below.
Select References
[1] https://www.youtube.com/watch?v=rmGid83_Y-0 (Interview with Duncan Clubb for JSA TV at Datacloud USA in 2023)
[2] https://www.cambridgemc.com/ai-at-the-edge
[3] https://viso.ai/edge-ai/edge-intelligence-deep-learning-with-edge-computing/
[4] https://www.jll.com/en-us/insights/data-center-outlook
[5] https://www.youtube.com/watch?v=9DCsqTlcGEI
[6] https://www.edgeir.com/how-ai-2025-data-center-trends-impact-edge-computing-20250220
[8] https://www.constructiondive.com/news/data-center-projects-construction-2025/738160/
[9] https://shows.acast.com/camtechpod/episodes/deloitte-2025
[10] https://www.iotforall.com/edge-ai-2025-predictions-reality-check
[12]
https://www.scalecomputing.com/blog/5-predictions-edge-computing-virtualization-2025
Contact Form
Subscribe to our Newsletter
Blog Subscribe
SHARE CONTENT
