Share
Former astrophysicist Giovanni Santostasi posted a detailed defence of his Bitcoin Power Law model on X this week, responding to critics who argue no mathematical formula can account for human behaviour. His model, which was published in 2018 and updated regularly since, predicts Bitcoin will reach $2 million by 2036.
The model expresses Bitcoin’s price as a function of days elapsed since the Genesis Block. It has tracked Bitcoin’s price across nine orders of magnitude with an R² above 0.95, an extraordinarily high goodness-of-fit for any financial model.
For such a bold prediction, it’s important to look at the reasoning behind it. Could it be more interesting than the numbers?
The power law doesn’t predict what any individual will do. It describes what remains when you integrate over all possible human actions at large scale.
Santostasi’s framework uses a theory from physics called the renormalization group theory. The core argument of the theory is: at large scales, the smallest details wash out. Who is buying, why they are buying, what the news cycle says, these become irrelevant operators, argues Santostasi.
Dear Mises — the Misesian objection is well taken as a philosophical point, but I think it misidentifies what the power law is actually modelling.
You're right that no model can predict individual human action. But the power law doesn't predict individual actions — it describes…
— Giovanni's BTC_POWER_LAW (@Giovann35084111) March 2, 2026
So what survives if these don’t? Santostasi lists what he describes as “the relevant operators”, including fixed supply, programmatic issuance, and decentralised consensus. According to him, they’re all architectural features that aren’t dependent on human motivation.
The power law model borrows an analogy from the field of thermodynamics. It states that we can’t predict the trajectory of any single molecule in a gas, yet we can predict pressure and temperature with extraordinary accuracy. Macro-regularity is therefore a product of the unpredictability of micro-actions, not despite it.
The model’s most compelling evidence is its resilience. Mt. Gox, the 2018 crash, China’s mining ban, the FTX collapse; each of these events caused massive deviation from the power law trajectory. Each time, the price of Bitcoin returned to the curve.
Not that these were small perturbations. In Bitcoin terms, they were civilisation-scale events. The model absorbed them because, in Santostasi’s framework, they aren’t important to the universality class – basically, noise at the micro level that washes out at scale.
The Austrian economics objection, recently highlighted by critics on X, is direct: no model can predict the aggregate outcome of individual human choices because human action is purposive, not mechanical.
However, Santostasi’s response is that this misidentifies what the model is doing. It isn’t predicting choices, it’s identifying what is invariant across all possible choices.
For example, a US government adoption of Bitcoin as a reserve asset, would likely accelerate movement along the existing curve, rather than destroy the pattern. The law breaks only if the relevant operators change. If Bitcoin’s supply ceiling is overridden, if issuance becomes discretionary, if consensus becomes centralised. Those changes require destroying Bitcoin itself.
That gives the model a specific, testable breaking point, one which is scientifically responsible. What it also means is that the $2 million target isn’t a price prediction in the traditional sense – it’s a conditional statement.
For Santostasi, this is where the physics points. If Bitcoin remains Bitcoin, the ten-year test is now running.
Share
