The data centre industry stands at a critical juncture. AI has transformed data infrastructure from a background utility into a front-line force in the global economy. As demand for computational power surges, operators are racing to design facilities that can meet both performance and sustainability expectations (McKinsey & Company, 2025).
The scale of transformation ahead
Between 2025 and 2030, AI workloads will redefine how data centres are designed, built, and managed. McKinsey estimates $6.7 trillion in global investment will be required to meet demand, with around 70% driven by AI applications (McKinsey Quarterly, 2025). Average facility capacity has risen from 30 megawatts to more than 200 megawatts, and some developments are approaching gigawatt scale (BloombergNEF, 2025).
Electricity consumption from AI training and services is set to quadruple within a decade, reaching 1,600 terawatt-hours by 2035 - about 4.4% of global power use, equivalent to the electricity consumption of the world’s fourth-largest nation (BloombergNEF, 2025). While efficiency metrics such as Power Usage Effectiveness are improving, from 1.4 today to a projected 1.2 by 2030, those gains are being overtaken by escalating demand.
Managing power density and energy supply
The rapid rise of AI computing has made energy access a defining factor in infrastructure planning. Power density has more than doubled since 2023 and may increase fourfold by 2027 (McKinsey Quarterly, 2025). By the end of the decade, AI racks could exceed 200 kilowatts per rack, pushing current copper busbar and cabling beyond their physical limits and prompting a shift toward 1,000- to 1,500-volt direct-current systems (Forbes Technology Council, 2024).
Traditional power networks are struggling to keep pace. In the United States, natural gas is carrying much of the short-term load, while renewables are expanding to meet longer-term climate goals (BloombergNEF, 2025). Many operators are now building on-site microgrids and islanded power systems, blending natural gas with wind, solar, or nuclear energy to reduce dependence on external grids and secure stable supply (Forbes Technology Council, 2024).
Balancing growth with sustainability
AI growth is amplifying the industry’s environmental responsibilities. Data centres could see a 160-200% increase in electricity demand by 2030, driven by AI and cryptocurrency mining (Forbes Technology Council, 2024). In response, operators are accelerating investment in renewable sources and innovative cooling systems such as liquid and bioengineered cooling, which use genetically modified microorganisms to manage heat and convert waste energy into reusable power (Forbes Technology Council, 2024).
A more intelligent and secure future
The shift toward AI-ready infrastructure brings new operational and cyber risks. Workforce shortages in HVAC, power systems, and physical engineering threaten to delay projects (Forbes Technology Council, 2024). Meanwhile, the adoption of zero-trust architecture and intelligent hardware security is becoming essential to safeguard critical systems (Forbes Technology Council, 2024).
The next five years will test whether the data centre sector can grow responsibly while ensuring resilience and transparency. Those able to combine digital oversight, renewable integration, and secure operations will define the next generation of infrastructure.
To speak to one of our experts.
References
- McKinsey & Company, “What is a data center?” (2025). https://www.mckinsey.com/featured-insights/mckinsey-explainers/what-is-a-data-center
- McKinsey Quarterly, “The trillion-dollar question: How AI is transforming data infrastructure.” (2025). https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-cost-of-compute-a-7-trillion-dollar-race-to-scale-data-centers
- Forbes Technology Council, “Data Centers: 18 Challenges (And Solutions) On The Horizon.” (2024). https://www.forbes.com/sites/forbestechcouncil/2024/12/19/data-centers-18-challenges-and-solutions-on-the-horizon
- BloombergNEF, “AI Data Centers Fuel Quicker Growth in Power Demand.” (2025). https://about.bnef.com/blog/ai-data-centers-fuel-quicker-growth-in-power-demand
