Physicists have long enjoyed applying statistical mechanics to social phenomena. A recent preprint takes this hobby seriously: it uses bosonic statistics to precisely describe Bitcoin's wealth distribution.
Quantum mechanics divides particles into two classes: fermions and bosons. Fermions obey the Pauli exclusion principle—no two particles can occupy the same state. Electrons are fermions, which is why they fill atomic shells one by one. Bosons, by contrast, love to crowd together: unlimited bosons can occupy a single state. Photons are bosons—this is why lasers exist, with countless photons collapsing into one quantum state.
Now ask a seemingly unrelated question: what does Bitcoin's UTXO (unspent transaction output) ownership distribution look like?
From Coins to Bits
Traditional money—gold coins, banknotes—has a key property: each physical unit is distinguishable. One 100-dollar bill is not another; they have different serial numbers. In physics this is "classical distinguishability," and ownership distribution then becomes a problem of classical statistical mechanics governed by exchange, inheritance, and trade.
Bitcoin is different. UTXOs are pure information—digital records. Two 1 BTC UTXOs have no intrinsic physical difference. At the protocol level they are completely indistinguishable, like two photons: swap them, and the state of the universe is unchanged.
Physics says that distinguishability fundamentally determines which statistics a system follows. If UTXOs are indistinguishable information units, they should follow bosonic statistics—not classical exponential, Pareto, or traditional economic distributions.
The Experiment: Measuring Bitcoin's "Temperature"
Between 2018 and 2023, the researchers took 72 monthly snapshots across 63 UTXO denominations (from dust to thousands of BTC). For each denomination, they counted addresses owning k UTXOs of that denomination.
They then fitted the data with a single-parameter geometric model, the parameter being an "inverse temperature" β. Key findings:
1. Remarkable fit quality. In 99.74% of samples, the Jensen-Shannon divergence (a measure of distribution similarity) was below 0.08—values under 0.1 are usually considered "nearly indistinguishable." Bitcoin's UTXO ownership distribution follows a geometric distribution almost perfectly.
2. Self-consistency test passed. The model predicts a definite mathematical relation between the inverse temperature β and the mean holdings. If the model were nonsense, this relation would fail. Instead, the deviation was below 0.1%—in every sample.
3. Striking stability. The inferred inverse temperature stayed within a narrow range across denominations spanning 8 orders of magnitude (from 0.00000001 BTC to thousands of BTC) and 6 years of history. One stable parameter describes everything.
This means Bitcoin's economy has something like a thermodynamic temperature—a global economic parameter—that remained essentially constant for six years.
What It Means
The paper is not just curve-fitting. It proposes a structural mechanism for digital economic inequality.
Core claim: when wealth shifts from distinguishable physical objects to indistinguishable information units, the underlying ownership statistics change—from classical to quantum (bosonic) statistics. Bosonic statistics naturally yields more unequal distributions (geometric rather than normal). Not because of greed, policy errors, or market failure—purely a consequence of the nature of information. Indistinguishable digital money acquires bosons' "clustering tendency," and wealth automatically concentrates into fewer addresses.
A helpful analogy is the laser. Pump energy into atoms and photons do not spread evenly across quantum states—they spontaneously bunch into the same state. Economic incentives are the pump; UTXOs are the photons. Under the right conditions, wealth "lasers" into a few addresses.
Feynman's Review
Statistical-mechanics-for-society papers are usually entertainment rather than science. This one does several unusual things:
- It doesn't just say "sort of resembles." It proposes a specific functional form (single-parameter geometric distribution), derives a self-consistency check (temperature–mean relation), and validates it against large-scale data with less than 0.1% deviation—a falsifiable hypothesis that passed.
- It finds an "invariant": a single parameter stable across 6 years and 8 orders of magnitude. If this is coincidence, it is an extremely stubborn one.
- It offers a structural account of digital inequality—not about greed or human nature, but about the nature of information. Bosonic clustering may be an unavoidable, mathematically driven consequence when wealth becomes fully indistinguishable units.
Reference paper
Chanhee Park, Claudio J. Tessone, Yu Zhang, Jeong-Hyuck Park. "Empirical confirmation of bosonic wealth statistics in Bitcoin UTXOs." arXiv:2605.12853, 2026.