2026/04/19 by Sebastian Gurovich · 1 voice
Physics and Astronomy · #astro-ph.IM #astro-ph.GA
arxiv published 2026/04/19 · arxiv updated 2026/07/01
We test the standard Kardashev conjecture -- that a technologically advanced civilisation's energy production grows at a fixed one percent per year -- against six decades of global energy-production data (1965-2024), with consequences for the search for extraterrestrial intelligence. Markov Chain Monte Carlo inference falsifies the one-percent rate, returning a posterior r=2.01%±0.03%~\mathrmyr-1; a linear model and a free-rate exponential fit the data equally well and are statistically indistinguishable. Extrapolated to the Type~II threshold (the solar luminosity), the linear form implies a transition time of order 105 Hubble times, while the exponential branch is excluded as a sustained trajectory by waste-heat and carrying-capacity limits. Neither adequate description therefore reaches energy-scaling Type~II coherently -- Kardashev's Conundrum. The observational consequence is that Type~II technosignatures, Dyson spheres in particular, are unlikely on physically realistic trajectories, offering a candidate explanation for the null results of Dyson-sphere surveys and a data-driven entry in Cirkovic's ``logistic'' category of Fermi-paradox solutions. The root cause is that the Kardashev state variable, power, is dimensionally incomplete: it captures energy throughput but not the information richness of its use. We therefore propose the Kardashev-Sagan-Nakamoto renormalisation, dividing energy production by the global proof-of-work hashrate to define B=P/H in \mathrmJ Hash-1 -- the KarNak unit -- a dimensionally complete information--energy variable motivated by the Landauer limit, on which the Type~II concept is preserved.