What Is the Kardashev Scale, and Does It Still Make Sense?
The most famous yardstick for alien civilizations measures how much they burn. What if that is the wrong thing to measure?
In 1964, a Soviet astronomer asked a strange question. How would we recognize an advanced civilization from across the galaxy? His answer: check its power bill. That idea still shapes how we search for alien intelligence. It also hides an assumption almost nobody questions.
The Kardashev Scale, explained
Nikolai Kardashev was a radio astronomer. He wanted to know what signals a distant civilization might send. To guess, he first needed a way to compare civilizations. He chose energy.
The Three Types
Type I, planetary. Uses the energy available on its home planet. Kardashev put this near 4 × 1012 watts, roughly humanity’s level in 1964.
Type II, stellar. Captures most of its star’s output, about 4 × 1026 watts. The classic picture is a Dyson sphere: a shell or swarm of collectors wrapped around a sun.
Type III, galactic. Commands the energy of a whole galaxy, about 4 × 1037 watts.
Popular retellings use rounder numbers. In 1973, Carl Sagan turned the three steps into a smooth curve and set Type I near 1016 watts. On his version, humanity uses roughly 19 trillion watts, which puts us at about 0.73. We are not yet a Type I civilization.
The scale was never just a ranking. It was a search strategy. Physicist Freeman Dyson had argued in 1960 that a civilization harvesting a star would have to shed waste heat, and that heat would glow in the infrared. So astronomers could look for stars and galaxies that are too bright in the wrong color.
A team led by Roger Griffith at Penn State did exactly that for Type III. They scanned nearly 100 million galaxies in NASA’s WISE data and studied about 100,000 of the most promising in detail. They found no sign of a galaxy-wide civilization. About 50 galaxies glowed unusually bright in the infrared, but nature can produce that glow too.
The assumption hiding in the scale
The scale rests on one idea: more advanced means more energy. It fits our own history. Fire, farming, coal, oil, electricity. Each leap in capability came with a leap in power use. Extend that line far enough and you reach stars and galaxies.
But a trend is not a law. Look at what actually produces intelligence.
Your brain runs on about 20 watts. That is less than a bright light bulb, and it is the most capable general intelligence we know of.
Computation has a physical floor, but the floor is low. In 1961, Rolf Landauer showed that erasing a single bit of information costs a minimum amount of energy. At room temperature that is about 4 × 10–21 joules. Today’s chips sit orders of magnitude above it.
One 2023 analysis estimates that even realistic transistor designs could be roughly 200 times more efficient than a current top AI chip. The gap is room, not a barrier.
Get Future Human Hypothesis’s stories in your inbox
Join Medium for free to get updates from this writer.
Efficiency also compounds. Our machines do more per watt every decade. A civilization a million years ahead would not be stuck with our engineering.
There is a fair objection. Right now, AI is pushing electricity demand up, and data centers are being built at enormous scale. That is real. But early versions of a technology are usually wasteful. Steam engines were. The mature version often does far more with far less.
The scale measures how much a civilization burns. It does not measure how much it understands.
Physicist John Barrow made a related point in 1998. He proposed a scale that runs the other way, from Type I-minus down to Type Omega-minus. Its levels track how finely a civilization can manipulate the world: machines, then genes, molecules, atoms, nuclei, particles, and finally spacetime itself. On his view, advanced means smaller and more precise, not bigger.
Could the most advanced civilizations be the least visible?
Now turn the scale upside down. If advanced intelligence becomes more efficient over time, the most advanced civilizations would use less energy per thought, shed less waste heat, and leave fewer traces. They would not look like glowing spheres around stars. They might look like nothing at all.
This is not a new thought. In 2012, futurist John Smart published the transcension hypothesis. It says advanced civilizations do not expand outward. They turn inward, toward ever smaller and denser forms of computation. From outside, they fade from view.
That would explain the silence. We have searched for loud civilizations and found none. Maybe loudness is a sign of youth. We have made a version of this argument before: the quiet may be what maturity looks like.
Here the Future Human Hypothesis comes in. Suppose some far-future branch of humanity is studying its own past. What would the best observer look like? Not a spacecraft. Not a signal. Not a heat signature. The best observer would be one that leaves no mark on what it watches, because any trace it left could change the very thing it came to study.
That fits a pattern we keep returning to: observers who watch but never intervene. Restraint becomes a feature of capability, not a gap in it.
None of this is proof. The scale may still describe some civilizations, and energy-hungry expansion remains possible. But our searches were built around one expectation, and so far the sky has not matched it.
So here is the question to sit with. If maturity means doing more with less, what exactly are we looking for, and would we ever notice it?
Further Reading
Primary sources and reference material cited or relevant to this essay.
Transmission of Information by Extraterrestrial CivilizationsN. S. Kardashev · Soviet Astronomy · 1964
Kardashev Scale (original values, Sagan’s 1973 interpolation, Barrow’s inward scale)Wikipedia · accessed October 2026
Report on the WISE search for Type III civilizations (Griffith, Wright, Povich et al.)Sci.News · April 2015
Limits to the Energy Efficiency of CMOS MicroprocessorsHo, Erdil & Besiroglu · arXiv · 2023
Beyond Fermi’s Paradox XI: The Transcension HypothesisUniverse Today · on J. Smart, Acta Astronautica 78, 2012
