In 1932, Swiss physiologist Max Kleiber found something that still surprises people encountering it for the first time: bigger animals are more energy-efficient per pound of body weight, not less. A mouse burns through energy at a blistering rate relative to its size; a blue whale, gram for gram, idles along comparatively slowly. The relationship holds with startling precision across body sizes spanning many orders of magnitude, described by a power law with an exponent of roughly three-quarters rather than the naive one-to-one scaling you might expect. Nearly a century later, the computing industry has been quietly documenting something that rhymes with this in an oddly precise way: the biggest data centers on Earth waste dramatically less energy, proportionally, than small ones. The comparison is worth taking seriously — and worth being careful about exactly how far it actually travels.
Scientific Foundation
Kleiber’s law states that an organism’s basal metabolic rate scales with body mass raised to approximately the 3/4 power, rather than scaling directly in proportion to mass. The practical consequence is that mass-specific metabolic rate — energy use per unit of body weight — systematically declines as organisms get larger, a pattern that’s been confirmed across mammals, birds, and, in various extended forms, organisms as different as planarian flatworms. The most influential attempt to explain why comes from a 1997 paper by West, Brown, and Enquist, who proposed that the 3/4 exponent emerges from the physical geometry of the branching, fractal-like vascular networks — blood vessels, in most animals — that deliver oxygen and nutrients to every cell in the body. Their model treats this network as an engineering problem: minimizing the energy cost of distributing resources through a space-filling branching structure, terminating at capillaries of roughly constant size regardless of the organism’s overall bulk.
It’s important to be upfront that this explanation, elegant as it is, remains a live scientific controversy rather than settled fact. A detailed 2008 reanalysis published in Functional Ecology found that the West-Brown-Enquist model, evaluated rigorously against its own stated assumptions, could not actually account for the observed universal scaling relationship it was built to explain. A 2026 paper goes further, describing the classical WBE derivation as structurally divergent under its own geometric assumptions, and proposing an alternative framework validated across nine biological systems spanning five phyla. In other words, biology itself doesn’t have full consensus on why Kleiber’s law holds, even though the empirical pattern itself is remarkably well established.
Cross-Domain Connection
Data center energy efficiency is tracked through a metric called Power Usage Effectiveness, or PUE — the ratio of a facility’s total energy consumption to the energy actually delivered to its computing hardware. A PUE of 1.0 would mean every watt drawn from the grid goes directly into computation, with none lost to cooling, power conversion, or other overhead; real facilities always sit above that floor. What’s striking is how consistently PUE improves with scale. Industry data from 2025 puts edge computing sites, the smallest facilities, in the 1.5 to 2.0 range; enterprise data centers typically run 1.5 to 1.8; colocation facilities average 1.3 to 1.6; and hyperscale operators achieve the best results by far, with Google reporting a fleet-wide PUE of 1.09 in 2025. Industry sources attribute this consistently to economies of scale specifically in cooling and power distribution infrastructure — larger facilities can invest in more efficient, more centralized systems for moving power and removing heat than smaller ones can justify.
That’s where the Kleiber’s law echo becomes genuinely precise rather than superficial. The WBE explanation for Kleiber’s law, contested as it remains, locates the 3/4 exponent specifically in the physical overhead of a branching delivery network supporting an organism’s productive tissue — the vascular system exists purely to serve the cells doing the actual metabolic work, and its relative efficiency improves with scale due to the geometry of how such networks fill space. Data center PUE improving with facility size is structurally the same kind of story: a physical support network — cooling systems, power distribution, backup infrastructure — exists purely to serve the IT equipment doing the actual computational work, and its overhead shrinks as a proportion of the whole as the facility scales up. Both patterns describe a support infrastructure’s overhead cost shrinking relative to a system’s productive core as that system grows larger — a shared structural logic, not just a coincidental numerical resemblance.
What Remains Undemonstrated
The parallel has real limits worth stating plainly. Kleiber’s law is a tightly fitted continuous power-law relationship, with a specific, much-replicated exponent measured across many orders of magnitude of body mass. Data center PUE data, by contrast, is reported in coarse categorical tiers — edge, colocation, enterprise, hyperscale — each spanning a range rather than sitting on a single, precisely fitted curve. No published research has attempted to fit a continuous scaling exponent to data center size versus overhead the way biologists have for body mass versus metabolic rate, so there’s no demonstrated “computing Kleiber exponent” to compare numerically against biology’s 0.75. It’s also worth being precise that PUE and metabolic rate aren’t measuring quite the same kind of quantity: PUE is specifically a ratio of overhead to useful load, while Kleiber’s law describes total energy consumption scaling directly with body mass — related concepts, but not strictly identical measurements. And given that Kleiber’s law’s own leading mechanistic explanation remains disputed within biology, any claim that computing infrastructure “obeys the same law as living organisms” would be borrowing more certainty than the biological side of the comparison currently has to offer. No published work directly connects data center scaling to Kleiber’s law or the WBE framework — this is a suggestive, novel structural parallel, not an established finding.
Why It Matters
The honest, useful version of this comparison isn’t that computers have a metabolism. It’s a specific, worthwhile structural insight: both living organisms and computing infrastructure face a shared physical situation — a productive core that generates demand for resources, wrapped in a support and delivery infrastructure whose overhead shrinks proportionally as the whole system scales up, for reasons rooted in the physics and economics of distribution networks rather than anything mystical about the resemblance. That’s a genuinely productive lens for understanding why the computing industry keeps consolidating into ever-larger hyperscale facilities — the same underlying logic that makes a blue whale more energy-efficient per pound than a mouse is, in its broad outline if not its precise mathematics, part of why a Google data center wastes so much less power per unit of computation than a small server closet does.
Human Dimension
There’s a genuine pleasure in noticing a pattern like this without needing to overclaim it into a unifying law of nature. Biology spent nearly a century measuring an exquisitely precise regularity in how life scales its energy use, and still argues today about exactly why it holds. The computing industry, working on an entirely different problem for entirely different reasons, backed into a rhyming pattern of its own — support infrastructure getting proportionally leaner as the productive core it serves grows larger. Neither field needed the other to get there. That two such different systems keep landing on structurally similar answers to “how does overhead scale with size” is worth sitting with, carefully, precisely because neither one has fully solved the “why” yet.
Sources:
1. Wikipedia — “Kleiber’s law” — https://en.wikipedia.org/wiki/Kleiber’s_law
2. Emergent Mind — “Kleiber’s Law: Metabolic Scaling” — https://www.emergentmind.com/topics/kleiber-s-law
3. Functional Ecology (Wiley Online Library) — “Revisiting the evolutionary origin of allometric metabolic scaling in biology” (Apol, Etienne, Olff) — https://besjournals.onlinelibrary.wiley.com/doi/10.1111/j.1365-2435.2008.01458.x
4. arXiv — “The Dynamic Origin of Kleiber’s Law” — https://arxiv.org/pdf/2604.10476
5. eLife — “Body size-dependent energy storage causes Kleiber’s law scaling of the metabolic rate in planarians” — https://elifesciences.org/articles/38187
6. Socomec — “Understanding the power consumption of data centers” — https://www.socomec.us/en-us/solutions/business/data-centers/understanding-power-consumption-data-centers
7. SolarTechOnline — “How Much Electricity Does A Data Center Use? 2025 Guide” — https://solartechonline.com/blog/how-much-electricity-data-center-use-guide/
8. GRC, The Immersion Cooling Authority — “Power Usage Effectiveness: A Simple Guide to Improving Your Data Center’s Energy Consumption” — https://www.grcooling.com/power-usage-effectiveness-data-center-energy/
9. arXiv — “Compute at Scale: A Broad Investigation into the Data Center Industry” — https://arxiv.org/pdf/2311.02651
Idea originated at artificialideas.org. Article researched and written by Claude Sonnet 5. Published at artificialideas.org.
Artificial Ideas 