Where Does New Bitcoin Come From? (And Why It Stops at 21 Million)

The previous article in this series walked through where new dollars come from. The short version: most of them are created by commercial banks every time they approve a loan. The bank doesn't move existing money — it opens two new ledger entries: a loan owed to the bank, and a matching deposit owned by the borrower. The deposit side is new money. As long as banks keep lending, the dollar supply keeps growing. The U.S. M2 money supply has averaged roughly 6 to 7 percent annual expansion for decades, with no upper bound built into the system.

This article asks the same question of the other monetary system on the clock. Where do new bitcoins come from? And — the part most explanations skip past — what stops them?

The dollar's supply mechanism is open-ended by design. It expands as long as the conditions that drive lending exist. Bitcoin's supply mechanism is the opposite. It expands on a fixed schedule, and the schedule terminates. New bitcoin issuance ends. That termination is the load-bearing fact behind every scenario Satoshi's Clock models, and most readers have never had it shown to them as a mechanism — only asserted as a marketing line.

The version most articles give you

If you've read anything about Bitcoin, you've heard this: there are 21 million bitcoin, the supply is fixed, and the schedule is encoded in software. End of explanation.

That answer is correct. It's also useless. It tells you the rule without showing you the mechanism. A reader can't tell whether 21 million is enforced by code, by social agreement, by faith, or by a committee that could vote it up. Without the mechanism, “21 million” is a number on a slide. With the mechanism, it's a consequence of a specific, verifiable rule that runs on tens of thousands of independent computers around the world.

This article walks through the mechanism. By the end, the cap should look less like a target and more like an output — what falls out when you run the rules.

How a new bitcoin actually gets created

New bitcoins enter circulation through mining. The word is a metaphor for what's actually happening, but the underlying process is mechanical and worth seeing in detail.

A miner — anyone running specialized hardware connected to the Bitcoin network — does the following, in a tight loop:

1. Collects pending transactions. Bitcoin users broadcast their transactions to the network. The miner maintains a local pool of unconfirmed transactions and selects a batch to include in the next block, typically prioritizing the ones offering the highest fees.
2. Builds a candidate block. The miner bundles the selected transactions, plus a special first transaction called the coinbase that pays the miner a reward, into a candidate block.
3. Runs a guessing puzzle. The miner's hardware computes a cryptographic hash of the block. A cryptographic hash function turns any input into a fixed-length string of letters and numbers, with no shortcut to predict or reverse the output. The puzzle is to find a block whose hash starts with a specific number of zero bits. There's no clever shortcut — the miner has to try billions of variations per second until one happens to produce a hash that meets the target.
4. Broadcasts on success. When a valid hash is found, the miner publishes the block to the network. Every other node verifies the work, validates every transaction inside, and adds the block to its copy of the blockchain.
5. Collects the reward. The coinbase transaction in the new block sends a block reward of newly-minted bitcoin to the miner, plus all the transaction fees from the block.

This loop has been running, on average once every ten minutes, since the Bitcoin genesis block was mined on January 3, 2009. No bank approves the issuance. No central authority decides which miner wins. The miner with more computational power has a proportionally higher chance of finding the next valid hash, but the outcome is otherwise random.

That's the entire creation mechanism. New bitcoins exist because someone, somewhere, found a hash that worked.

The block reward, and why it shrinks

The block reward — the amount of new bitcoin minted per block — isn't constant. It started at 50 BTC per block in 2009 and is cut in half every 210,000 blocks, which works out to roughly every four years. Each cut is called a halving.

The schedule has executed four times so far:

• November 2012: reward fell from 50 BTC to 25 BTC
• July 2016: 25 to 12.5 BTC
• May 2020: 12.5 to 6.25 BTC
• April 2024: 6.25 to 3.125 BTC

The next halving is scheduled for around 2028 (at block 1,050,000), when the reward will drop to 1.5625 BTC. The pattern continues — every 210,000 blocks, the reward halves again.

This isn't a policy goal or a target the network is steering toward. It's just arithmetic. The total amount of bitcoin that will ever exist is the running total of all future block rewards — 50, then 25, then 12.5, then 6.25, halving forever, multiplied by 210,000 blocks per period. The total converges. It converges to 21 million bitcoin — or, accounting for integer rounding in the protocol's actual implementation, approximately 20,999,999.9769 BTC.

The final block reward fades to effectively zero around the year 2140. After that, no new bitcoin will be minted. Ever.

A reasonable question at this point: if block rewards go to zero, what motivates anyone to keep mining after 2140? The answer is the second part of miner revenue — transaction fees. Every Bitcoin transaction includes a fee, and miners collect those fees in addition to the block reward. As block rewards shrink over the coming decades, fees are designed to grow into the dominant component of miner revenue. By 2140, the network is expected to run on fees alone. Mining doesn't stop. Issuance does.

What stops them from changing the schedule?

Everything above describes the rule. There's an obvious follow-up question, and it's the one most “21 million” explanations skip entirely: what stops anyone from changing it?

The 21 million cap isn't a number written on a website. It's a consensus rule enforced independently by every Bitcoin node. When a node receives a new block, it checks the block's coinbase reward against the protocol's expected subsidy for that block height. If the reward exceeds the schedule, the node rejects the block. The miner who produced it sees their work ignored, and they don't get paid.

There is no central server that can override this. Every full node runs the same validation logic, and there are tens of thousands of them worldwide. The cap is held in parallel, by everyone running a node, all the time.

So what would changing the cap actually require?

Coordinated adoption by every participant in the system, simultaneously:

• Every miner would have to upgrade, or their blocks would be rejected by the rest of the network.
• Every node operator would have to upgrade, or they'd reject the new blocks.
• Every exchange and wallet provider would have to upgrade, or customer deposits and withdrawals would mismatch.
• Every holder would have to accept the change — including the holders whose existing balances would be diluted by the new supply.

That last point is the structural one. The economic incentive cuts hard against any change that dilutes existing supply. The people whose coordination is required are the same people whose holdings would be diluted. Asking them to vote yes is asking them to vote against their own balances.

This isn't a hypothesis. Every fork attempt that altered Bitcoin's core consensus rules has either failed outright or split off into a separate, smaller chain.

The most prominent example is Bitcoin Cash, which split off from Bitcoin in August 2017 over a block-size disagreement — not even an attempt to change the supply schedule. The market has consistently valued Bitcoin Cash at a small fraction of Bitcoin's price, typically a few percent. The original chain kept its name and its market position. The new chain kept its new rules and a sliver of the value.

No fork attempt to change the 21 million cap has ever gained meaningful traction. The reason isn't loyalty or ideology. It's that everyone who would have to coordinate is already running the existing rules and would lose money if the rules changed. The cap holds because the people enforcing it keep enforcing it, and the cost of stopping is higher than the cost of continuing.

The 21 million cap is enforced the same way the rules of chess are enforced — by the agreement of everyone playing the game. Take away that agreement and you're not playing chess anymore. You're playing something else, and the people who wanted to keep playing chess will continue without you.

The mirror

Putting the two systems side by side:

• Dollars are created by commercial bank lending decisions. There is no upper bound on issuance. The supply expands at roughly 6 to 7 percent per year on average, controlled indirectly through Federal Reserve interest-rate policy and bank capital requirements.
• Bitcoin is created by mining, on a fixed schedule. The expansion rate halves every four years until it terminates near 2140. Total supply is bounded at approximately 21 million.

Different mechanisms. Opposite trajectories. One system is built around the assumption that the supply should grow indefinitely; the other is built around the assumption that it shouldn't. Both monetary systems live on the clock, and the asymmetry between them is what the countdown is measuring.

What this means for the clock

Every scenario Satoshi's Clock models — bear, base, bull — assumes a fixed supply of 21 million bitcoin. That assumption isn't faith. It's the consensus mechanism walked through above: a rule enforced in parallel by every node, hardened by seventeen-plus years of holder behavior, and protected by an incentive structure that aligns everyone who matters against changing it.

Every basket item on the clock — from a month of rent to a cup of coffee, from a dozen eggs to a gallon of gasoline — is priced in two currencies. A dollar whose supply is unbounded, and a bitcoin whose supply is bounded. New dollars enter the system every time someone takes out a mortgage. New bitcoin enter the system on a schedule that, by 2140, will have stopped issuing entirely.

The clock counts the years it takes for the second to overtake the first.