How does Bitcoin's supply schedule actually work?

Bitcoin's monetary policy is unusual in one specific way: it is written in software rather than decided by anyone, and it was fixed at launch. Whatever you think of that design, it is worth being able to state it exactly, because it is frequently described badly.

New coins enter only one way

There is exactly one mechanism by which bitcoin comes into existence: the block subsidy, paid to whoever produces a valid block. Nothing else creates coins. There is no reserve, no premine, no discretionary issuance.

The protocol targets one block every ten minutes, and enforces that target by adjusting difficulty (next lesson).

The halving

The subsidy halves every 210,000 blocks. At ten minutes a block:

210,000 × 10 minutes = 2,100,000 minutes ≈ 3.99 years

which is where "roughly every four years" comes from. It is a block count, not a date — the calendar dates drift with actual block times.

Where 21 million comes from

This is one line of arithmetic, and doing it yourself is more useful than memorising the number.

Each era issues 210,000 × subsidy. Total issuance is:

210,000 × (50 + 25 + 12.5 + 6.25 + …)

The bracket is a geometric series with first term 50 and ratio ½, which sums to 100. So:

210,000 × 100 = 21,000,000, near enough — the real ceiling is 20,999,999.9769 BTC, because the subsidy is an integer number of satoshis and the halvings round it down until it reaches zero

Issuance stops when the subsidy rounds to zero in the integer arithmetic the software uses — bitcoin is counted internally in satoshis, at 100,000,000 to the coin — which happens around the year 2140.

The important, under-noticed consequence

By 2026 roughly 20 of the 21 million already exist. Current issuance is 144 expected blocks a day × 3.125 = 450 BTC per day, against 900 before April 2024. The overwhelming majority of all bitcoin that will ever exist has already been issued, and the remaining schedule tails off asymptotically for another century.

What a capped supply does not mean

This is a statement about supply only. It says nothing about demand, and price is set by both.

Bitcoin has fallen more than 50% from a high on several occasions and more than 70% in past cycles, with a fixed schedule running unchanged throughout. A capped supply has never prevented a drawdown and does not imply anything about future price. There have been four halvings — 2012, 2016, 2020 and 2024 — and four observations of anything is not a dataset. This course makes no claim, in either direction, about what any of it implies.

The open question the design creates

As the subsidy shrinks toward zero, block producers must be paid increasingly by transaction fees instead. Whether fee revenue alone will fund enough security in the long run is a genuine, unresolved engineering debate among people who work on the protocol — not a settled matter with a comfortable answer.

Try it now

1. Do the series yourself: 50 + 25 + 12.5 + 6.25 + 3.125 + … → 100, then × 210,000 = 21,000,000. Deriving the cap beats remembering it.
2. On a Bitcoin block explorer, find the current block height and divide it by 210,000. The integer part tells you which subsidy era you are in; check it against the table above.
3. The full price history is below. Navigate to 11 May 2020 and to 20 April 2024 — the last two halvings — and Measure six months either side of each. Write the four percentages down, describe what you see in neutral language, and then resist writing a rule from two events. Two observations are an anecdote, and the sample cannot grow faster than one every four years.