Coming soon: AI in space
Sign up for The Key Point of View, our weekly newsletter of blogs, podcasts, and videos
With Earth Day happening last month, I’ve been thinking about a conversation I had over Easter weekend. I had mentioned the recent discourse about how asking an artificial intelligence (AI) program a single question could use as much as one bottle of water for the cooling required in data centers.
My cousin Jonathan is the CTO and Co-Founder of Boosted.ai and has a doctorate in theoretical chemistry. He just looked at me and said: “That’s not how thermodynamics works.”
We are truly left- vs. right-brained individuals. He excels in the mathematic and scientific fields and I’ve built my entire career around editing and writing. So, the dichotomy of our comments wasn’t surprising. My instinct, as always, was to argue with him...until I realized I didn’t really know the answer to a very basic question:
What is the actual environmental impact of AI?
The Scale Behind AI Infrastructure
AI systems rely on physical infrastructure, primarily large-scale data centers that require significant amounts of electricity and cooling. Recent reporting highlights growing concern among researchers and environmental experts about how quickly this demand is increasing. Training and running advanced AI models can consume large amounts of energy, particularly as models become more complex and widely adopted. Water usage is also part of the equation, since many facilities rely on water-based cooling systems to prevent overheating.
At the same time, the exact environmental cost is difficult to quantify. Estimates vary depending on system design, geographic location, and energy sources. A single AI query does not carry a fixed resource cost; it depends on the infrastructure supporting it. As a result, widely shared claims such as “one bottle of water per query” are better understood as simplified representations rather than precise measurements.
Energy Use Is Rising, but So Is Efficiency
Data centers have become more efficient over time. Many operators are investing in renewable energy, improving cooling systems, and designing hardware that requires less power. However, overall demand continues to rise, and AI is accelerating that growth. Some projections suggest that data centers could account for a larger share of global electricity consumption in the coming years if expansion continues at its current pace.
This doesn't make the impact inherently negative, but it does make it significant. There is a clear tension: AI has the potential to improve efficiency across industries, while also requiring substantial resources to operate.
Two things can be true, and both realities can exist simultaneously.
The More Extreme Idea: Moving Data Centers to Space
Some proposals extend beyond improving existing infrastructure and instead rethink it entirely. One of the more widely discussed ideas is placing data centers in orbit. The concept is built on a few theoretical advantages. Satellites in certain orbits can access near-continuous sunlight, which means a steady source of solar energy without the interruptions caused by weather or nighttime on Earth.
There is also the matter of cooling. The vacuum of space and its naturally low temperatures could help dissipate heat more efficiently than traditional, water-intensive cooling systems used in terrestrial data centers, but this idea is far from concrete.
Experts continue to point out significant challenges, including the cost of launching and maintaining infrastructure in orbit, the complexity of transmitting data efficiently back to Earth, and the technical limitations of operating sensitive hardware in a high-radiation environment. There are also broader questions about scalability and whether space-based systems could realistically match the capacity of ground-based data centers in the near term.
Even so, interest in orbital data centers continues to grow, driven in part by the increasing demand for AI computing power and the difficulty of expanding infrastructure on Earth.
Where That Leaves the Conversation
Is AI ruining the environment? Is space actually the final frontier? That’s difficult to answer in any sort of absolute. The Easter comments turned out to be less about one being right and more about asking better questions.
From a technical perspective, the details matter. From a broader perspective, the concern itself reflects a growing awareness of how digital infrastructure affects the physical world. Somewhere between those viewpoints is where a more accurate understanding begins.
And in some cases, it may even lead to solutions that sound as unconventional as launching data centers into orbit.
Stay ahead in the ever-evolving print industry by browsing our Industry Reports page for the latest insights. Log in to the InfoCenter to view research and studies through our Workplace- and Production-based Advisory Services. Log in to bliQ for product-level research, reports, and specs. Not a subscriber? Contact us for more information.