22/10/2025
AI looks so 'hot' and Jet Engines, so 'cool'
Artificial Intelligence is undoubtedly reshaping the future – but with it comes a steep surge in energy consumption and heat generation.
The energy world has been closely observing how the rise in AI workloads is driving energy infrastructure to its limits. The latest data shows global data centers consumed approximately 460 TWh in 2022 alone. By 2026, this is projected to nearly double. That’s almost 1,000 TWh — more than the current electricity usage of Germany.
What’s often overlooked in this conversation is the thermal load AI infrastructures are now producing. A single advanced AI chip such as NVIDIA’s H100 can consume up to 700 watts. Scale this across massive server farms, and the heat output is not unlike what we manage in aerospace propulsion systems.
This is where aerospace engineering offers a surprising, but practical solution.
Jet turbines and rocket engines face—and solve—thermal challenges that far exceed the demands of data centers. Features like closed-loop Brayton cycle systems and regenerative cooling, using supercritical CO₂ or advanced heat exchange fluids, have long delivered reliable cooling in extreme conditions. These same principles can now be applied to AI infrastructure, allowing significantly higher heat removal efficiency (up to 95%), while potentially regenerating waste heat into usable energy.
We’re witnessing early research and pilot studies exploring the integration of miniature turbine-based thermal systems within server racks. These approaches move beyond conventional water or air cooling and could enable compact, scalable solutions for high-density compute centers.
With the AI industry projected to reach over $400 billion by 2027, the call for sustainable, high-performance thermal management is not optional — it’s urgent.
From managing 1,400°C turbine exhausts to cooling hypersonic surfaces at Mach speeds, aerospace has been solving heat problems for decades. It’s time to transfer that knowledge into the infrastructures powering our digital future.
AI is accelerating. Our cooling technologies must keep pace.
