The 21 Million Cap & Supply Schedule

The single most repeated fact about Bitcoin is “there will only ever be 21 million.” This page asks the why behind it: where the number comes from, how the rule is actually enforced, and what “digital scarcity” really means once you stop taking the phrase on faith.

The schedule, not the number

The 21 million cap is not written anywhere as a target. Nobody coded MAX_COINS = 21000000. (Bitcoin Core does contain a MAX_MONEY = 21000000 * COIN constant, but it’s a sanity check on transaction amounts — the source comment itself notes it is not what limits the supply.) Instead, the cap is an emergent consequence of two simpler rules:

1. New coins are created only in the block subsidy — the reward paid to whoever mines a block.
2. That subsidy halves every 210,000 blocks (roughly every four years).

Run those two rules forward and the total supply converges. The subsidy began at 50 BTC per block and steps down geometrically:

Era   Blocks (start)   Subsidy/block   New BTC this era
 0          0          50              10,500,000
 1     210,000         25               5,250,000
 2     420,000         12.5             2,625,000
 3     630,000          6.25            1,312,500
 4     840,000          3.125             656,250
 ...        ...          ...                 ...
                                       ───────────
                            total →    ≈ 21,000,000

Each era mints exactly half of the one before. That is a geometric series: 210,000 × 50 × (1 + ½ + ¼ + ⅛ + …). The sum in parentheses converges to 2, so the whole thing converges to 210,000 × 50 × 2 = 21,000,000. The cap is just arithmetic.

Why “approximately” 21 million

The true total is slightly under 21M, for a humble reason: the subsidy is stored as an integer number of satoshis and is computed with integer division. Each halving rounds down, shaving off fractional satoshis. After 33 halvings the subsidy rounds to zero and issuance stops. The final figure is about 20,999,999.97 BTC — close enough that everyone says “21 million,” but a nice reminder that the rule, not a round number, is what’s real.

For the block-by-block mechanics of the subsidy and the halving cadence, see halving & issuance.

”Digital scarcity” — enforced by consensus, not decree

Here is the part that matters and is most often glossed over. Scarcity in the physical world is enforced by physics: there is only so much gold in the crust, and digging more is hard. Scarcity in fiat is enforced by decree — a central bank promises to be disciplined, and you trust the promise.

Bitcoin’s scarcity is enforced by neither. It is enforced by every full node independently checking the rules. When a block arrives, each node verifies that the coinbase transaction pays no more than the subsidy the schedule currently allows, plus the fees in that block. A miner who tries to pay themselves 60 BTC when the rules say 6.25 produces a block that honest nodes simply reject as invalid. It never becomes part of the ledger anyone agrees on.

This is the recurring thread of the whole textbook in one sentence: the cap is how untrusting strangers agree on one ledger’s monetary policy. No one has to trust a mint, because anyone can run the arithmetic and refuse coins that break it. Scarcity here is a social fact maintained by software, replicated across thousands of machines that have no reason to trust each other.

Could the cap ever change?

In principle, yes — if a supermajority of users, miners, and businesses all chose to run software that allowed more coins. In practice this is wildly unlikely: holders of a scarce asset have every incentive to reject a change that would dilute them. The cap is protected less by code than by the aligned self-interest of everyone holding it. We pull this thread further in game theory.

Divisibility: scarcity without clumsiness

A fixed supply only works as money if you can split it finely enough. Bitcoin’s base unit is the satoshi: 1 BTC = 100,000,000 sats. So the real ledger isn’t denominated in 21 million units — it’s denominated in about 2.1 × 10¹⁵ satoshis. As purchasing power per coin rises, ordinary payments simply move down into smaller sat amounts. You never run out of room to make change. This is why a hard cap doesn’t doom Bitcoin to deflationary paralysis at the unit level — the unit can always shrink. (Recall the divisibility row in what money actually is; Bitcoin scores excellently precisely because a digital ledger can divide without limit.)

Lost coins: the supply is even tighter than 21M

Coins are controlled by private keys. Lose the key — a discarded hard drive, a forgotten password, a death with no heir who knows the seed — and those coins are provably unspendable but still counted. They sit on the ledger forever, visible, frozen, gone.

Nobody knows the exact figure, but credible estimates put millions of BTC as likely lost, including a large early stash widely believed to belong to Satoshi and never moved. The economic effect is subtle but real: lost coins make the effective circulating supply smaller than the issued supply, and they do so unpredictably. In a fiat system, a central bank can replace lost notes. In Bitcoin, loss is permanent — a quiet, ongoing tax that makes the remaining coins slightly scarcer.

Contrast: discretionary fiat issuance

The cleanest way to see what Bitcoin is for is to put the two policies side by side.

	Bitcoin	Discretionary fiat
Who sets supply	A fixed algorithm everyone verifies	A committee using judgment
How much exists	Knowable decades in advance	Decided meeting to meeting
Can it be expanded?	Only by near-universal agreement	Yes, by policy
Enforcement	Every node rejects rule-breakers	Trust in the institution
Failure mode	Rigidity (can’t respond to crises)	Debasement (can over-issue)

Neither column is strictly “better” — this is a genuine tradeoff, and a textbook should say so. Fiat’s discretion lets central banks fight recessions and act as lenders of last resort; Bitcoin’s rigidity removes that flexibility entirely. What Bitcoin offers instead is credible predictability: a monetary policy that cannot be quietly changed by anyone, because changing it would require persuading a global, adversarial, self-interested crowd to vote against their own holdings. That credibility is the product.

By the numbers — Bitcoin’s “inflation rate” vs gold’s

Scarcity is easiest to feel as a yearly issuance rate — new supply divided by existing stock. Bitcoin’s is now low and falling by schedule:

blocks per day      ≈ 144          (10-minute target)
subsidy per block   = 3.125 BTC    (after the 2024 halving)
new BTC per year    ≈ 144 × 3.125 × 365  ≈ 164,000 BTC
circulating supply  ≈ 19.7 million BTC   (mid-2020s)

issuance rate ≈ 164,000 / 19,700,000  ≈ 0.8% per year — and halving toward 0

For comparison, gold’s above-ground stock grows by roughly 1.5% per year as miners dig more — and that rate tends to rise when prices rise. Bitcoin’s already sits below gold’s and is contractually cut in half about every four years, on a schedule every node can verify. (Figures approximate, as of the mid-2020s.)

The architect’s lens

The 21M cap is a designed monetary policy — interrogate it rather than reciting it:

• Why does it exist? To give money credible, unchangeable predictability — a fixed ~21M supply no mint or committee can quietly expand, enforced by every node rather than by decree.
• What problem does it solve? Debasement and trust in an issuer. The cap is emergent from two rules (new coins only in the coinbase subsidy; halving every 210,000 blocks → a geometric series summing to ~21M), and any node rejects a block paying more than the schedule allows — scarcity as a social fact maintained by software.
• What are the trade-offs? The failure mode is rigidity: unlike discretionary fiat there’s no lever to fight recessions or act as lender of last resort — you trade flexibility for predictability. (Divisibility to 100M sats per BTC keeps the fixed cap usable as purchasing power rises.)
• When should I avoid it? A hard cap is the wrong policy if you want an elastic supply that can respond to crises — that’s exactly what central-bank discretion buys, at the price of trusting the institution.
• What breaks if I remove it? Lift the cap and the core value proposition evaporates: holders can be diluted, “digital scarcity” becomes a promise rather than a verifiable fact, and Bitcoin is just another discretionary currency.

Check your understanding

1. The 21M cap isn’t a hardcoded target. What two simpler rules produce it, and why does the total converge to ~21 million?
2. Why is the real cap slightly under 21 million BTC?
3. What does it actually mean to say Bitcoin’s scarcity is “enforced by consensus, not decree”? Who does the enforcing?
4. How does divisibility to satoshis keep a fixed supply from becoming unusable as prices rise?
5. Give one genuine advantage of discretionary fiat issuance that a fixed cap gives up.

Show answers

1. (1) New coins are created only in the block subsidy, and (2) that subsidy halves every 210,000 blocks (started at 50 BTC). Run forward, that’s a geometric series — 210,000 × 50 × (1 + ½ + ¼ + …) — whose bracket converges to 2, giving 210,000 × 50 × 2 = 21,000,000. The cap is just arithmetic, not a hardcoded target.
2. Because the subsidy is stored as an integer number of satoshis and computed with integer division, so each halving rounds down, shaving off fractional satoshis. After 33 halvings the subsidy rounds to zero, leaving about 20,999,999.97 BTC.
3. It means no mint, central bank, or decree sets the supply — every full node independently checks the rules. When a block arrives, each node verifies the coinbase pays no more than the schedule’s current subsidy plus fees; a miner who tries to pay themselves 60 BTC when the rule says 6.25 produces a block honest nodes simply reject as invalid. Scarcity is a social fact maintained by software across thousands of mutually-distrustful machines — exactly how untrusting strangers agree on one ledger’s monetary policy.
4. Because the base unit is the satoshi (1 BTC = 100,000,000 sats, ~2.1 × 10¹⁵ total), so as purchasing power per coin rises, ordinary payments just move down into smaller sat amounts. The unit can always shrink, so a hard cap never runs out of room to make change.
5. Discretion/flexibility: a central bank can expand supply to fight recessions and act as lender of last resort. Bitcoin’s fixed cap gives that up entirely — its failure mode is rigidity (can’t respond to crises), the price of credible, unchangeable predictability.