The short answer: around the year 2140, the block reward rounds down to zero and no new bitcoin is ever issued again. Miners keep mining — paid entirely by transaction fees — and the system keeps running exactly as designed. The real story is that this isn't a far-future event: as of block 961,748, 95.56% of all bitcoin already exists (live number, computed from the chain). The era of meaningful new supply is already almost over.
How the 21 million limit actually works
Nothing in bitcoin says "21 million" directly. The cap emerges from two rules: every block mints a reward, and every 210,000 blocks (~4 years) that reward halves. 50 → 25 → 12.5 → 6.25 → 3.125 today, on toward zero. Sum the infinite series and it converges just under 21 million BTC. The supply schedule is a geometric series wearing a monetary policy costume — fixed since 2009, executed by every node on earth, with four halvings already behind it.
Because issuance is front-loaded, the drama is asymmetric: the first four years created half of all bitcoin; the last century will create a rounding error. From here, scarcity isn't coming — it's arrived. What remains is a long, thin tail.
So what actually changes in 2140?
Remarkably little, mechanically. Blocks continue, transactions confirm, keys sign. The one real change is who pays for security. Miners today earn the subsidy (new coins) plus the fees users bid in the mempool auction. The subsidy is the training wheels — a temporary inflation tax that bootstrapped security while adoption grew. Every halving shifts weight from subsidy to fees; 2140 is just the date the training wheels formally come off, a century after they stopped mattering.
The honest version of the security-budget question
Will fees alone pay for enough mining to keep the network safe? Anyone who answers with certainty — in either direction — is selling something. What can be said honestly:
- The optimistic case: block space is permanently scarce, demand for final settlement grows, and fee spikes have already — briefly — out-earned the subsidy. A mature fee market plus a century of adaptation time is a long runway.
- The skeptical case: fees are volatile and cyclical; quiet years could mean thin security budgets, and the equilibrium between fee revenue and attack cost is unproven at final scale.
- The shared ground: the transition is the slowest-moving deadline in finance — halvings arrive on schedule, everyone sees them coming, and the system gets a live rehearsal every four years.
Questions people actually ask
When will the last bitcoin be mined?
Around the year 2140. Halvings every 210,000 blocks keep shrinking the reward until it rounds to zero satoshis. The famous 21 million is an asymptote: over 99% will exist by the 2030s, but the final satoshis trickle out for another century.
Will miners quit when the block reward ends?
Miners are paid twice: new coins plus transaction fees. As the subsidy fades, fees carry more of the load — in high-demand periods, fees have already briefly exceeded the subsidy. Whether fees alone sustain today’s level of security is bitcoin’s most serious open question, and pretending it is settled — either way — is dishonest.
Can the 21 million cap be changed?
Technically, code is just code — but the cap is enforced by every node independently, and a change requires near-universal voluntary adoption. The overwhelming economic majority runs software that rejects any block violating the cap, and holders have every incentive to keep it that way. The cap is best understood as a social contract enforced by mathematics.
Will there really be 21 million usable bitcoin?
No — the protocol will issue slightly under 21M (rounding quirks already destroyed small amounts), and millions more are permanently lost with their keys. The spendable supply is meaningfully below the headline number, and it only shrinks.