Bitcoin ยท a personal thesis
After eight years of holding through every drawdown and halving, I've stopped trying to guess tops. Instead I anchor to the one part of the model that has actually held, the support line, and to a number the model and the loudest buyers both land on, for the same underlying reason: ~30% a year.
BTC price
$77.0K
Power-law floor
$60.4K
Fair value
$151.8K
Price / floor
1.27ร
Distance above floor
+27.5%
Model status
๐ข INTACT
101days
The floor has not been broken
Since the model froze 6 Jun 2026.
The power-law floor is a statistical long-term support level derived from the model. It is not a guarantee that Bitcoin cannot trade below it. See today's full model snapshot or run your own dates in the calculator.
What is the floor? The power-law floor is this model's long-term statistical support level, derived from the fitted Bitcoin power-law trend and the observed distribution of historical price residuals.
It is not a guarantee that Bitcoin cannot trade below this level. A sustained or significant breach would be evidence against the model, not a reason to move the line (see Methodology).
Both land in the same neighbourhood because they're describing the same fifteen years of history two different ways, not because one confirms the other independently.
The strongest test of a model is not how well it explains the past, but whether it survives attempts to disprove it.
Attack the model โ01 / THE THESIS
For a while I kept hearing the CEOs of Bitcoin treasury companies put Bitcoin's long-run return at roughly 30% a year (call it the next ten to twenty years). It's oddly specific, and oddly consistent across people who don't coordinate. For a while I treated that agreement as if it meant something on its own.
It doesn't, and that's the first thing to be honest about. Giovanni Santostasi's power law, where price tracks a straight line on a logโlog chart against time since the 2009 genesis block, has a support line whose near-term forward growth works out to around 30% a year. The CEOs and the model aren't two independent oracles landing on the same figure. They're two descriptions of one curve. Anyone honestly extrapolating Bitcoin's decelerating log-growth ends up in the same neighbourhood, because they're all reading the same fifteen-year history. So the agreement isn't a second witness, and I won't dress it up as one. The single fitted curve is the signal. The "convergence" is just what that one curve looks like when several people describe it out loud.
And 30% itself isn't a constant or a recurring magic number. It's a snapshot of where the maths lands from here, today. Two things stack to produce it: a line-to-line growth rate that's already decaying (high-20s% over the coming decade, easing toward the teens over twenty years), plus a one-time bonus from Bitcoin currently sitting near its floor and drifting back toward fair value. Change the horizon, or start from a point above the floor, and the number moves. 30% is true for this starting point and this horizon, not a law I can bank on staying put.
So what I'm actually anchoring to isn't the agreement, and it isn't a fixed return. It's the one structural feature of the model that has held through every drawdown, halving, and capitulation in my eight years of holding: the support line. I use it as a reference for where Bitcoin is cheap against its own history, not as a promise about what it will pay.
02 / STRESS TEST
Putting the thesis down honestly means stress-testing it, not just admiring it. Here's where it's solid and where I've had to tighten my thinking.
Where it holds
The arithmetic is real. Run the power-law exponent (about 5.7) forward from today and the line-to-line growth rate comes out near 29% over a ten-year horizon, easing toward the low-20s% over twenty years. Because Bitcoin currently sits close to its floor, projecting from here up to the model's fair-value line two decades out lands right around 30%. You capture a little extra from reverting off the floor toward trend. So "the support line projects ~30% CAGR" isn't a vibe. It's what the maths says for the next decade.
The convergence isn't a coincidence. As section 01 lays out, the CEOs and the power law are two descriptions of one curve, so their agreement confirms the framing above rather than acting as a second, independent vote.
Their described behaviour matches the model even if they never name it. Michael Saylor, the most vocal of them, doesn't forecast a flat 30%. He describes a declining rate (roughly 44% easing to 40, 35, 30, 25 and onward, averaging out near 30) that reaches around $13 million per coin by 2045. That declining shape is the unmistakable fingerprint of a power law (an exponential model could never describe itself that way), and his $13M target sits almost exactly on the power-law fair-value line. So whether or not he's "using" it, what he's describing is it.
Anchoring to the support line is the right choice of line. Above the floor, Bitcoin's price is dominated by sentiment, leverage, and liquidity, which are noisy and not reliably predictable in timing or size. The support line is the most constrained, most empirically reliable part of the whole model. If I'm going to anchor to anything, anchoring to the predictable boundary and ignoring the chaotic region above it is internally consistent.
Where it bends
30% is a declining average, not a constant. This is the easiest trap and the one I'm most careful about. If I bake a flat 30% into a fifteen-year plan, I'll overestimate the back half. The model, and Saylor's own description, has the rate bleeding down toward the high-teens and low-20s over time. I treat 30% as a near-term figure that decays, not a locked annual return.
The CEO number is corroboration, not proof. A Bitcoin-treasury CEO is the single most incentivised person alive to project a high-but-credible-sounding Bitcoin return. It's "talking their book" to attract capital to their own stock. Since the opening already grants that their agreement isn't a second, independent witness, the only honest weight to give it is small: it leans on the same chart and the same hope, so I let it colour the picture without double-counting it as evidence for the model.
The conservatism of "support" lives in the price level, not the growth rate. The floor, fair-value, and ceiling lines are parallel: they share the same exponent, so they imply the same CAGR. Choosing the support line doesn't give me a lower assumed growth rate. It gives me a lower assumed price level. In effect I'm planning as if Bitcoin might only ever trade at its historical floor. That's the genuinely conservative bet, but I shouldn't kid myself that "support" automatically means "lower return assumption."
Model risk is permanent. The fit is descriptive, not a law of nature. It's sensitive to the window you fit it on, and the floor has broken before, briefly, three times across fifteen years, most notably during the March 2020 crash. A fifteen-year regularity is strong, but it is a regularity, not a guarantee. To keep myself honest about that, the model's parameters are frozen and published in advance with a stated break condition, a sustained close below the floor, that would count against the thesis. I don't get to refit the line to erase it.
Saturation is the long-horizon headwind. Compounding at 30% off today's ~$1.2T is a very different proposition from sustaining it once Bitcoin is a $20T-plus asset. The absolute inflows required grow enormous, and real-world adoption could decelerate faster than the curve assumes. The power law already bakes in deceleration, and reality may decelerate harder.
03 / IN PRACTICE
The support line is my valuation and accumulation anchor: a way to judge when Bitcoin is cheap relative to its own history, to size my expectations at a declining ~20โ30%, and to keep me from panic-selling into drawdowns that the model treats as routine visits to the floor. It is not a price target, a guarantee, or a promise that the floor can't break.
Held that way, the thesis is disciplined rather than wishful: I'm anchoring to the most reliable part of the best-fitting model, using a number that stands up on its own maths and happens to match what the most committed institutional buyers are publicly underwriting. The conviction comes from the convergence and the track record, not from believing anyone has a secret.
Want to see the model itself? Read the floor explained simply, or run the numbers in the calculator.