Halving

Bitcoin pays miners for securing the network, and the size of that payment is written down in advance rather than decided by anyone. Four halvings have happened so far, the most recent on April 20, 2024 at block 840,000, which took the subsidy from 6.25 coins to 3.125. For a buyer it is a supply event with a known height, an uncertain date, and no guaranteed price effect.

How it works

The halving is not an event the network schedules. It is a consequence of counting blocks.

Every node computes the allowed subsidy from the height of the block in front of it, and the answer drops by half each time the height crosses a multiple of 210,000. No vote happens, no announcement is made, and no software update is required. A node running code from years ago arrives at the same number as a node compiled this morning, because the arithmetic is fixed and the input is a count that cannot be faked without doing the work.

At the ten minute target that difficulty aims for, 210,000 blocks take 2,100,000 minutes, or 1,458 days, three days short of four calendar years. Reality has run faster. The gap between the first two halvings was 1,319 days, then 1,402, then 1,440. Blocks have tended to arrive slightly ahead of schedule because hash rate has grown within each retarget window, so every halving so far has landed earlier than the naive four year estimate, and the gaps have been converging on it.

The four so far: November 28, 2012 at block 210,000, cutting 50 coins to 25. July 9, 2016 at block 420,000, cutting 25 to 12.5. May 11, 2020 at block 630,000, cutting 12.5 to 6.25. April 20, 2024 at block 840,000, cutting 6.25 to 3.125. Next comes block 1,050,000 and a subsidy of 1.5625, projected for 2028 with the usual caveat that the date is a guess and the height is not.

The sequence ends. After 33 halvings the shifted value rounds away to nothing, which happens at block 6,930,000, somewhere around the year 2140. From that point miners are paid by fees alone.

Why this matters when you buy bitcoin

A halving changes what miners earn. It does not change anything about the coins you already hold, the addresses you use, or the software you run.

The measurable effect is on new supply. Before April 2024 the network issued roughly 328,500 coins a year; after it, roughly 164,250. That is a genuine tightening, and it is worth being precise about what it tightens. Newly mined coins are the only bitcoin that has never had an owner. Everything else on an exchange order book already belongs to someone, and a halving does nothing to change how much of it they want to sell, which is why the honest description is a slow constraint rather than a switch someone flips.

The marketing around halvings is where buyers actually get hurt. Every cycle brings countdown clocks, "pre halving" bundles, leveraged products timed to the date, and altcoins advertising a halving schedule of their own as though the schedule were the point. Bitcoin has both risen and fallen in the months following previous halvings. If an offer needs the halving to make sense, it is a price forecast wearing a protocol rule as a costume.

The practical response is boring and works: decide on an amount and a rhythm and stop timing the event. Apps built for that, such as Relai in Europe or Bitpanda for SEPA savers, let you set a recurring purchase and forget the countdown. Whatever you use, the halving should not change your plan, because it was public knowledge before you made the plan.

One more thing to watch is the scam that arrives with every halving: messages telling you to move your coins to a "halving compatible" address or to claim a bonus. Nothing about a halving requires you to touch your wallet. Any instruction to do so is theft.

Block 840,000, the halving that paid mostly fees

The fourth halving produced the most valuable block bitcoin has ever mined, and almost none of that value was the subsidy.

Block 840,000 was found by ViaBTC on April 20, 2024. The subsidy had just fallen to 3.125 coins. The block collected around 37.6 coins in transaction fees, roughly twelve times the subsidy, for a total of about 40.7 coins. The reason was timing rather than luck: the Runes protocol was designed to launch at exactly that height, so a crowd of people bid against each other for a place in that one block.

It is the clearest illustration of what a halving does and does not do. Bitcoin's issuance was cut in half at that moment, and the miner's revenue for that block went up. Fees are the other half of what a block pays, they are set by an auction rather than by code, and over the next century they are the half that has to grow.

Halving vs block reward

The halving cuts the subsidy; the block reward is the subsidy plus the fees of every transaction included. Confusing the two makes headlines like "the halving cut miner revenue in half" wrong twice over: revenue also depends on fees, and on the price. Block 840,000 is the counterexample sitting right at the boundary, a block where the subsidy halved and the reward set a record.

Halving vs stock to flow

Stock to flow is a price model, not a part of bitcoin. The analyst known as PlanB published it in March 2019, using the halving schedule as the input to a formula projecting bitcoin's value from its scarcity, and that paper forecast a one trillion dollar bitcoin market value after the May 2020 halving, roughly 55,000 dollars a coin. The model has been revised more than once since, which is the tell: rules do not get rewritten when they turn out to be inconvenient. The halving is a rule every node enforces. Stock to flow is one person's regression on top of that rule, and treating the second as a consequence of the first is how people end up buying a forecast.