TECH & AI
For decades, the data centre was a model of predictable growth. Its purpose was clear: to provide reliable uptime for enterprise applications, web hosting and cloud storage – workloads that, while growing, did so along a relatively stable curve. Now, that model is broken.
The industry is in the midst of a rapid and costly transformation driven by AI, demanding a capital deployment estimated at US $ 1.8tn between last year and 2030. Some forecasts place the total cost of necessary compute infrastructure closer to US $ 7tn.
This surge in investment highlights a fundamental conflict between the exponential demands of software and the linear progress of the physical world. Vivek Swaminathan, Director of Products and Solutions for Digital Workplace Solutions at Unisys, sees this gap regularly. He describes a frequent client situation that has become a defining issue of mid-2025:“ We have this real estate. We have this data centre that’ s all up, but there’ s no plumbing, there’ s no piping, there’ s no facilities for anything liquid cooling enabled.”
This gap between what exists and what is needed has created three deeply interconnected challenges that now dominate industry strategy: securing a sufficient power supply, engineering systems to manage immense heat and finding a workforce with the right skills.
120 August 2025