What Happens to the Power Law After the Last Halving?

In short: The last halving changes nothing about the Power Law itself. The model only uses the time since bitcoin started, not the number of new coins. Issuance of new bitcoin ends around 2140, but in practice it is over by around 2035, when 99 % have been mined. What really changes concerns the miners: they will then live on fees alone.

Many holders hear about the Power Law and the halving at the same time, often in one sentence: "Every four years the supply halves, that's why the price rises." That leads to an obvious question: when there are no more halvings, does the model break too? This article calculates when issuance really ends, why the Power Law does not depend on it and what actually changes after the last halving. How the model works in general is explained in The Bitcoin Power Law Explained.

When is the last halving?

Every 210’000 blocks the reward for a new block halves. Bitcoin started in 2009 with 50 bitcoin per block; since April 2024 it has been 3.125. How this works in detail is in The Bitcoin Halving Explained Simply.

The software counts in sats, the smallest unit. At every halving the number is halved and everything after the decimal point is cut off. After 32 halvings only 1 sat per block is left. At the 33rd halving it becomes 0. From then on no new bitcoin are created.

The years assume one block every ten minutes from today. In the past, blocks came slightly faster on average because computing power usually grows. The late halvings may therefore come a few years earlier. Only the block numbers are fixed.

99 % are mined by around 2035

The year 2140 sounds like a long issuance period. In reality almost everything is distributed much earlier. Each epoch brings half as many bitcoin as the one before. The first four years brought 10.5 million, the next four 5.25 million, and so on.

On 13 Nov 2026 there are about 20’111’000 bitcoin, 95.8 % of the final supply. In February 2035, around block 1’411’200, it is 99 %. In February 2048 it is 99.9 %. The entire last 92 years of issuance therefore bring only about 21’000 bitcoin.

The chart shows it: after 2035 the curve is practically flat. The final supply, by the way, is not exactly 21 million but 20’999’999.9769 bitcoin, because fractions of sats are cut off at every halving. Why the limit is 21 million at all is explained in Why There Are Only 21 Million Bitcoin.

For the price this means: the step from "almost no new bitcoin" to "no new bitcoin at all" around 2140 is tiny. The big changes in supply lie in the next ten years, not in 114 years.

Why the Power Law needs no halvings

The Power Law describes the price as a function of a single quantity: the days since the genesis block on 3 January 2009. Neither the number of bitcoin nor the number of new coins per block appears in the formula.

Time stands for something else here: adoption. More users, more capital, more infrastructure, more computing power. Such networks often grow along power laws, fast at first and ever slower afterwards. The Power Law says: bitcoin grows the way a network grows that gradually reaches the world. Whether there are four, eight or zero halvings in between makes no difference to the formula.

That also shows where the model comes from. The line has no steps at the halvings. It is a smooth curve through 16 years of price history.

The contrast: Stock-to-Flow

The Stock-to-Flow model is completely different. It derives the price directly from the ratio of existing supply to new supply. Every halving doubles that ratio, and the modelled price jumps. Around 2140 the new supply becomes zero, the ratio infinite, and the model has no answer left. The comparison of the two models is in Power Law vs. Stock-to-Flow.

And the cycles?

To be fair: the big highs and lows of the past were close in time to the halvings. Many explain the four-year cycles that way. But these cycles are deviations around the line, not the line itself. If the cycles weaken because the halvings barely take any bitcoin away anymore, that changes the swings around the trend, not the trend.

The data already show exactly that. The peaks above the Power Law got smaller with every cycle: in the epoch up to 2012, 5 % of days were above 5.8 × the line; in the epoch since 2024, only above 1.2 ×. The bear-market lows, on the other hand, stayed at 0.4 to 0.56 × the line. The criticism that the cycles may have shaped the model is covered in Criticism of the Bitcoin Power Law.

What changes for miners

This is where the real change lies. For each block, miners receive two things: the subsidy, meaning new bitcoin, and the fees of all payments in the block. As of November 2026 by far the largest part of their income comes from the subsidy. After the last halving there are only fees.

Satoshi planned it this way from the start. In the white paper he writes that once a predetermined number of coins have entered circulation, the incentive can transition entirely to transaction fees and be completely inflation free.

It is interesting how this interacts with the Power Law. The subsidy halves every four years. So the price would have to more than double in four years for miners' income in dollars to still rise.

If the price follows the model, the fee question does not become important only in 2140 but from the 2040s, when the subsidy shrinks in dollars. Whether fees will then be enough to pay for the network's security is an open question. It depends on how many people want space in the blocks and what it is worth to them. How fees arise is explained in Understanding Bitcoin Transaction Fees and the Mempool.

The difficulty adjustment provides stability: if income drops and miners switch off, mining becomes easier, and blocks keep coming about every ten minutes. So the rules work without a subsidy too. What is open is only how much computing power the fees will pay for.

What the model would say far in the future

If you keep applying the formula stubbornly, numbers soon come out that have nothing to do with the world anymore.

For comparison: the world's entire economic output in 2024 was about 110 trillion dollars a year, according to the International Monetary Fund. Already around 2050, all bitcoin along the line would be worth more than three years of world output. In 2140 it would be about 2’500 years. A single sat would then be worth $134.

Why a century of extrapolation doesn't hold

No asset can grow faster than the economy it trades in forever. As long as bitcoin is small, it can grow much faster than the world through new users. Once it is a large share of global wealth, it can only grow as fast as that wealth. Somewhere before that point every growth curve has to flatten, a power curve included.

On top of that, the model is measured on 16 years of data. That allows a trend to be carried forward a few years, with large uncertainty. A century is more than six times as long as the entire history the model comes from. For such periods it is no longer a forecast but an arithmetic exercise.

That is why the site's charts only calculate as far as 2040. How much the parameters themselves already wobble is shown in Criticism of the Bitcoin Power Law.

The model's growth slows every year

One point is often overlooked: the Power Law is not an exponential curve. On an exponential curve the price grows by the same percentage every year. On the Power Law this percentage falls every year, because each new year is a smaller part of bitcoin's age so far.

So the slowdown is already built into the model. The site's trend currently grows even a little slower than the line, because it approaches the floor from above. For your planning this means: if you calculate with the Power Law today, don't assume constant returns but returns that get smaller year by year. And for the time after 2140 the model says nothing reliable, with or without a halving.