When Is the Last Bitcoin Halving? The Final Cut, 2140

When is the last Bitcoin Halving? Block 6,930,000, expected around the year 2140. At that height the block subsidy, already down to a single satoshi, rounds to zero and new Bitcoin issuance stops permanently. Twenty-nine cuts remain, the next at block 1,050,000 in 2028. After the last one, miners are paid by transaction fees alone.

That is the answer, and it is the kind of answer that gets repeated without ever being felt. A date past everyone. A block height nobody in this room will refresh. The schedule does not care that the number is inconvenient to the human attention span, and that indifference is the most interesting thing about it.

So when is the last Bitcoin Halving, exactly?

Block 6,930,000. Not a year, a height. Roughly 2140 is the translation into calendar time, and the translation is the soft part.

The chain does not keep years. It keeps blocks, and it aims for one every ten minutes, and it misses constantly in both directions. Difficulty drags the average back. Over a century of dragging, an error of a few seconds per block compounds into years, which is why every honest source says “around 2140” and no honest source says a month.

On 29 September 2026 the chain tip stood at block 968,988, according to mempool.space. That leaves 5,961,012 blocks to walk before the subsidy reaches zero.

So there are two numbers, and only one of them is real. The height is a fact. The year is a forecast wearing the clothes of a fact.

When was the most recent Bitcoin halving?

Half the people who type this question mean the other “last”. The most recent Bitcoin Halving fired at block 840,000 on 20 April 2024, at about 00:09 UTC, and cut the subsidy from 6.25 BTC to 3.125 BTC. It was the fourth cut. It was not the final one.

Every cut so far, by block height and UTC date, per BitRef’s halving record:

Halving	Block height	Date (UTC)	Subsidy before	Subsidy after
1st	210,000	28 Nov 2012	50 BTC	25 BTC
2nd	420,000	9 Jul 2016	25 BTC	12.5 BTC
3rd	630,000	11 May 2020	12.5 BTC	6.25 BTC
4th	840,000	20 Apr 2024	6.25 BTC	3.125 BTC

The long version of that story, block by block, lives in the Bitcoin Halving history.

The cuts landed in 2012, then 2016, then 2020, then 2024, four rehearsals, each one loud, each one photographed, each one treated at the time as the event. None of them was. They were small-h halvings, drills for a thing that has not arrived. The schedule walks toward the Halfture and does not run.

How many Bitcoin halvings are left?

Twenty-nine, counting the one at block 1,050,000 in 2028, which has not happened yet. Twenty-eight after that. The halving count runs to thirty-three, and four are spent.

The subsidy is set in satoshis, so the subsidy columns are exact and the year column is approximate. The rough years assume the old four-year pace and will drift.

Halving	Block height	Subsidy after (BTC)	Subsidy after (satoshis)	Rough year
5th	1,050,000	1.5625	156,250,000	2028
6th	1,260,000	0.78125	78,125,000	2032
7th	1,470,000	0.390625	39,062,500	2036
10th	2,100,000	0.04882812	4,882,812	2048
20th	4,200,000	0.00004768	4,768	2088
30th	6,300,000	0.00000004	4	2128
32nd	6,720,000	0.00000001	1	2136
33rd	6,930,000	0	0	2140

The very next one has its own page: the Bitcoin Halving in 2028, and what it does and does not change.

Twenty-nine is a strange number to sit with. It is short enough to count on paper and long enough that the counting is an act of faith rather than planning. Nobody reading this will see the last one. Everyone reading this is inside the same schedule as the person who will.

Why does the reward round to zero instead of halving forever?

Because the subsidy is an integer count of satoshis, and integers do not divide forever. They run out.

Fifty Bitcoin is five billion satoshis. Halve it thirty-two times and you are left with one satoshi, the smallest unit the ledger can express. Halve it once more and there is nothing left to halve. The arithmetic does not gracefully approach zero the way a curve does. It hits the floor of the unit and stops.

Look at the tenth row of the table. 4,882,812 satoshis, not 4,882,812.5. The half satoshi was simply dropped, and the same truncation happens at every cut where the number goes odd. That is where the missing coins in the supply cap come from.

This is the part people miss when they picture an infinite tail. There is no infinite tail. There is a last paying block, height 6,929,999, and it pays one satoshi, and then the next block pays nothing and every block after it pays nothing forever. The Halfture is that terminal cut, and the Halfture is the Rapture, the one appointment on the schedule that closes rather than repeats.

When will all Bitcoin be mined, and has anyone checked the arithmetic?

Around 2140, and yes, and the way they checked it is stranger than the answer.

Bitcoin Core carries a consensus unit test called subsidy_limit_test. It walks the chain from height zero upward, adds up the block rewards, and asserts that the total comes to exactly 2,099,999,997,690,000 satoshis. Not twenty-one million. Twenty million, nine hundred and ninety-nine thousand, nine hundred and ninety-nine point nine seven six nine. The famous round number is shorthand, and the real cap is an ugly integer that a machine checks every time the project builds.

Two details in that test are worth the trouble. It does not walk every block; it samples every thousandth one and multiplies by a thousand, because the subsidy only changes every two hundred and ten thousand blocks and walking all of them would be slow. And it walks to fourteen million, more than twice as far as the last cut, well past a guard clause at the sixty-fourth halving that has no business ever firing. The test harness is built to cross the end of the money and keep going, just to prove that nothing happens out there.

Somebody wrote that loop bound. Somebody chose to look past the end.

Who argued about the end of the money, and how did that go?

Briefly, and the transcript is public.

On 11 March 2014 a contributor opened a pull request to force the block reward to zero once the halving count reached sixty-four, closing a defect in how the original code handled the bit shift after block 13,440,000. Peter Todd NACKed it, saying the project should use the roughly 250 years it had to study the issue more carefully. Gregory Maxwell said he was confident it would be fixed sometime in the hundreds of years before it mattered. Another commenter observed that the longer the delay, the more likely a miner alive today would still be around when it did.

Gavin Andresen merged it on 3 April 2014.

That thread is the only famous argument about the end of Bitcoin’s money, and it is not an argument about the end. It is an argument about whether there is time to argue. Which is the same argument every generation has about every deadline it will not personally meet, conducted here in a comment box, in public, by people who then fixed it anyway. Distrust the institutions that would rather manage your horizon than show you the code. This one showed you the code.

What happens to miners after the last Bitcoin halving?

They get paid in fees or they stop. There is no third option written anywhere.

Each cut before the last one already halves what a unit of hashrate earns from the subsidy, which is why miners watch hashprice the way farmers watch rain. The final cut takes the subsidy part of that number to zero. Fee revenue already spikes and collapses on a schedule nobody controls, and the long argument about Bitcoin’s security budget lives in the gap between “fees will be enough” and “fees will be enough reliably”. Nobody knows. The people who say they know are describing a market that does not exist yet, in a century they will not visit, with a straight face.

What can be said is narrower and more useful. The security of the chain has always rested on the honest majority doing the boring thing, and after 2140 the boring thing has to be paid for by the people using the chain rather than by new issuance. That is a change of who pays, not a change of whether. It is also the most examined open question in the whole design, which is the opposite of a hidden flaw.

What skeptics say

In several places the argument breaks, and they are the interesting places.

The first crack: a date in 2140 is not a plan, it is a horizon, and horizons are what people reach for when they want their position to feel inevitable instead of chosen. A man holding an asset has an obvious reason to like a story in which the asset’s scarcity is a law of nature. I am that man. Weigh it accordingly.

The second: the cap held because humans noticed a defect and wrote a guard by hand. It was edited once, in public, by people who had never met the author. Anything edited once can be edited again, and “the code says so” is a weaker sentence after you have read the pull request than before.

The third: nobody alive has any idea what a fee-only chain looks like under stress. The optimistic case and the pessimistic case are both extrapolations across a century, which is another way of saying neither is evidence. And a schedule is not a promise about value. The schedule promises supply. That is all it has ever promised.

And the last: maybe none of this saves anyone. Maybe it is a clock in an empty room, ticking correctly, watched by nobody, and the correctness was never the point.

What do you do with a date you will not live to see?

You hold, and you hand it down, and you let the arithmetic be longer than your life.

This is where the doctrine sits, and it sits quietly. You have to hold Bitcoin to be saved. That is theology, not a prospectus, and this site has no idea what the price does next week or next decade; it is a claim about custody and posture, not a claim about returns, and nothing here is financial advice. Holding a thing you did not sell in a drawdown is a different act from predicting a thing. One of them is available to you. The other is not available to anyone.

So the work is small and it is the same every year. Keep the keys. Know what happens to them when you are gone, because the keys you hand down are the only part of this that crosses a generation. Twenty-nine cuts remain and the schedule is indifferent to whether you are ready for any of them.

Look at the height.

FAQ

Q: When is the last Bitcoin halving?
A: The final halving is the thirty-third, at block 6,930,000, expected around the year 2140. It sets the block subsidy to zero. If you mean the most recent halving, that was the fourth, at block 840,000 on 20 April 2024.

Q: How many Bitcoin halvings are left?
A: Twenty-nine as of 2026: the fifth at block 1,050,000 in 2028, then one every 210,000 blocks until block 6,930,000. Four of the thirty-three have already happened.

Q: When will all Bitcoin be mined?
A: The last satoshi of new subsidy is paid at block 6,929,999, roughly in 2140. The total ever issued will be 20,999,999.9769 BTC, slightly under 21 million, because the subsidy is truncated to whole satoshis at each cut.

Q: Can the 2140 date change?
A: The block height cannot change without consensus. The calendar date can and almost certainly will drift, because blocks arrive faster or slower than the ten minute target and difficulty only corrects the average. Treat 2140 as an estimate and block 6,930,000 as the fact.

Q: What secures the network once the block subsidy is gone?
A: Transaction fees alone. Whether fee revenue is large and steady enough to fund the hashrate the chain needs is the central open question in Bitcoin’s long-run security model, and it is argued in public rather than settled.