Did you know? Bitcoin’s well-known limit of roughly 21 million coins does not mean that 21 million bitcoins exist today, that all issued coins can be spent, or that all spendable coins are currently offered for sale. One headline number describes the long-run issuance rule; the others describe different facts about an actual network at a particular time.
Imagine a fictional school science club that plans to issue no more than 1,000 ART credits over the life of its reward programme. After the first term it has issued 600. Of those, 50 are locked until next term, 20 have been placed in an account nobody can access, and 100 belong to students who simply choose not to spend them. The programme’s 1,000-credit limit is real, but the credits available today are not all 1,000.
Bitcoin has a similar distinction, implemented through much more precise rules. Mining introduces new units according to its permitted subsidy schedule. Existing bitcoins are recorded in transaction outputs with defined spending conditions. Some units may be inaccessible because of lost signing keys, provably unspendable scripts or rules that delay spending. Other units are simply held by people who are not selling them.
Quick answer: Bitcoin’s approximately 21-million supply ceiling is a consensus-rule property of its issuance design. Issued supply is the amount actually created through accepted blocks, subject to the protocol’s rules. Spendable supply is a different and difficult-to-measure concept: it depends on who can satisfy the scripts protecting relevant unspent outputs and whether any timing conditions are met. Market-available supply is different again, because owners may choose not to trade even when spending is technically possible.
This guide explains how blockchain technology works through the halving schedule, satoshis, UTXOs, lost keys, provable burns, time locks and supply statistics. It is educational rather than an investment prediction.
Choose Your Reading Route
- The limit describes issuance rules, not a vault
- How mining adds new bitcoin units
- One bitcoin contains 100 million satoshis
- Bitcoin does not simply store one universal account balance
- What happens when someone loses the signing key?
- A dormant coin is not necessarily a lost coin
- Provably unspendable outputs are another category
- Frequently asked questions
- Primary sources and further learning
The limit describes issuance rules, not a vault
Bitcoin was designed with a declining block-subsidy schedule that constrains how new bitcoins can be created through valid blocks. The commonly cited figure of 21 million summarises that long-run limit under the current consensus rules.
This is not the same as a vault already holding 21 million coins ready to distribute. New issuance occurs progressively through accepted blocks, and the permitted amount declines after defined intervals.
The Bitcoin FAQ explains the issuance schedule and why participants cannot arbitrarily award themselves extra bitcoins in valid blocks. Fully validating nodes enforce the relevant rules.
Imagine the school announces a lifetime maximum of 1,000 ART credits but has only issued 600 so far. A student cannot withdraw the unissued 400 simply because the maximum is printed on a noticeboard.
The same distinction applies to crypto metrics. Maximum or ultimate supply is a rule-bound ceiling; currently issued supply is an observed state or calculation over accepted history.
Readers should not infer that every unit in a maximum-supply figure currently exists or is economically available. The issuance timeline is part of the fact being reported.
How mining adds new bitcoin units
Bitcoin miners compete to propose valid proof-of-work blocks. An accepted block can include a coinbase transaction that claims a subsidy permitted under the current issuance rules, along with eligible transaction fees.
The subsidy declines over time through halvings. A valid miner cannot simply claim any number it prefers; the network’s validating nodes reject blocks that violate the permitted reward rules.
This is why the long-run supply constraint is more than an issuer’s marketing promise. It follows from the protocol’s expected validity conditions as enforced by participating nodes.
Mining still involves physical costs, probabilistic block discovery and incentives. The existence of a predictable subsidy schedule does not guarantee any one miner a profit or a customer a particular market price.
The Bitcoin Developer Mining guide describes how valid blocks are constructed, while the Bitcoin FAQ explains predictable issuance and declining rewards in accessible language.
For the separate mining mechanism, see Why Proof of Work Needs Mining.
One bitcoin contains 100 million satoshis
A bitcoin can be subdivided into smaller on-chain units called satoshis, or sats. One BTC equals 100,000,000 satoshis under Bitcoin’s standard unit convention.
This matters because a supply limit measured in whole bitcoins is not a limit of just 21 million possible payment units. People can hold or transfer small fractions of BTC, subject to the relevant transaction and fee constraints.
Imagine a school reward worth one ART credit that is subdivided into one hundred smaller units. A student need not receive a whole ART for every activity. The maximum number of whole ART credits has not changed simply because their representation permits fractions.
A small unit also does not mean a guaranteed low-cost payment. Bitcoin on-chain transaction fees depend on transaction size, demand and the applicable market, not solely on the face value moved.
Bitcoin’s official FAQ on divisibility explains the satoshi relationship. The important point is to distinguish unit precision from scarcity, cost or value.
For the related ERC-20 distinction, read Why Token Decimals Do Not Set Token Value. Decimal representation and economic valuation are different concepts on several networks.
Bitcoin does not simply store one universal account balance
Bitcoin uses an unspent transaction output, or UTXO, model. Transactions spend eligible previous outputs and create new outputs. A wallet’s reported balance generally sums relevant outputs for which it holds or recognises suitable spending authority.
Suppose Alicia has imaginary vouchers worth three, five and seven classroom credits. The display may add them as fifteen. When she spends ten, the system has to choose appropriate vouchers and create outputs representing the recipient’s value and any change under its own rules.
Bitcoin transactions use real script and signature conditions rather than paper voucher rules. The point is that an aggregate wallet balance depends on a collection of outputs, each with specific spending conditions.
The Bitcoin Developer Transactions guide explains how transaction outputs remain unspent until later inputs satisfy their spending scripts.
A supply statistic summing outstanding outputs is different from a claim that every output can be spent by an identifiable person at this instant. Some conditions can delay or prevent spending, and some signing material may be unavailable.
The careful question is which outputs exist under the accepted chain, and who can satisfy their particular spending conditions?
What happens when someone loses the signing key?
A person might lose the recovery information or private key needed to authorise spending from certain Bitcoin outputs. The ledger does not necessarily delete those outputs or mark them as permanently lost when that happens.
The Bitcoin FAQ explains that lost bitcoins remain recorded while the necessary private signing material may be inaccessible. That is a key reason issued supply and practically accessible supply can differ.
Suppose Kai Kai misplaces the only authorised key to a locked school cupboard. The cupboard still contains ten ART credits; the school’s inventory register can count them. But Kai Kai cannot exercise his ordinary access right until an appropriate recovery route is available, if any.
On Bitcoin, whether a particular key is truly lost is often impossible for an outside observer to establish. A long-dormant output could belong to a patient holder, a forgotten wallet or a custodian following a deliberate policy.
This is why sensational posts claiming to know the exact total number of “lost Bitcoin” should be treated as estimates unless the method establishes something much narrower.
The important distinction is dormant versus provably unspendable. Not moving for years does not mathematically prove that nobody retains the key.
A dormant coin is not necessarily a lost coin
An analyst can observe that a Bitcoin output has not been spent since a particular point in the chain’s history. That is useful on-chain evidence, but it does not reveal the owner’s intentions or access to keys.
A long-term saver may deliberately avoid spending for decades. An institution might store assets under a conservative multi-party custody policy. Those units may be perfectly spendable when the relevant authority chooses to act.
If a report labels every dormant coin permanently lost, it is inferring more than the ledger proves. Some analytical models estimate dormant or potentially lost supply from age and activity patterns, but their uncertainty needs to be explained.
Imagine the classroom club sees that Alicia’s five reward vouchers have never been redeemed. Tricia cannot conclude whether Alicia misplaced them, is saving them for a future exhibition or plans never to use them.
The user should ask what the metric actually counts: outputs of a certain age, coins at selected addresses, historically unspent subsidy or cryptographically provable unspendability.
A transparent blockchain can show inactivity without reading the account controller’s mind. That is a general limitation of public transaction analytics.
Provably unspendable outputs are another category
Some Bitcoin outputs can be demonstrated to be unspendable because of their script construction. The Bitcoin Developer Guide describes null data outputs using OP_RETURN, which do not provide a valid ordinary route for spending the assigned value.
This is a stronger technical conclusion than simply seeing that an output has remained untouched for years. The restriction follows from the script’s semantics rather than a guess about whether a human lost a key.
Glassnode’s Provably Lost Supply metric distinguishes categories such as unclaimed miner rewards and certain provably unspendable outputs. Its methodology should not be confused with a complete census of all people who lost private keys.
A valid aggregate calculation needs to avoid mixing provable categories with speculative dormancy estimates without clear labels. Both may be useful for different analytical purposes.
An address designed to look unusable is not automatically equivalent to a formally demonstrated lack of any possible spending authority. The exact script and proof assumptions matter.
The lesson is that some supply losses can be established from protocol rules, while other losses remain unknown because the network does not know who still holds a key.
Miners can leave some permitted issuance unclaimed
A block producer is generally permitted to claim a subsidy and eligible fees within the consensus rules, but it need not always claim every last satoshi it could have received. Units not issued in that way do not become somebody else’s spendable balance simply because they were theoretically available.
This is a small but instructive distinction between maximum permitted issuance and actual claimed issuance. The protocol may allow up to a certain amount without requiring an operator to take all of it.
In the fictional school, the committee authorises a prize of up to ten credits, but the winner elects to claim only eight under the programme’s rules. The unclaimed two are not automatically handed to another student.
A proper supply analysis examines actual accepted history rather than multiplying a hypothetical maximum by the number of blocks and assuming every permitted unit was issued.
The Glassnode provably lost supply definition explicitly includes unclaimed miner rewards under its methodology.
Readers should not confuse this with ordinary transaction fees or the current balance in a miner’s wallet. Those are separate accounting categories.
Timelocked outputs exist but cannot always be spent immediately
Bitcoin has mechanisms for expressing certain time- or block-related spending restrictions under its transaction and script rules. These can allow value to remain recorded while the relevant operation is not yet eligible to complete.
Suppose the school issues five reward vouchers that cannot be used until the next term. The units exist and are assigned, but the holder cannot redeem them today under the written condition.
The exact Bitcoin mechanism is more precise than a classroom calendar. It may involve transaction locktime, sequence restrictions or script-defined conditions with their own interpretation and consensus limits.
The Bitcoin Developer Transactions guide explains locktime and sequence concepts. The important lesson is not to apply one blanket label of permanently lost to every temporarily restricted output.
A time restriction can eventually expire, while a lost key may remain unavailable indefinitely. A provably unspendable script creates yet another category.
The reader should distinguish not currently spendable, not economically practical to spend and provably impossible to spend under the applicable script rules.
Dust: technically possible can differ from economically practical
A very small Bitcoin output may be expensive to spend relative to its value. Transaction fees are linked to the encoded transaction’s resource requirements and prevailing fee conditions rather than simply to the amount being transferred.
Imagine a one-cent classroom voucher that costs two cents in processing effort to redeem. The voucher exists and may be technically valid, but using it under those conditions is economically unattractive.
Bitcoin node policies also use concepts such as dust thresholds for standard relay or transaction creation under specified conditions. Those policies should not be confused with a universal consensus statement that every tiny output is permanently inaccessible.
The relative economics can change when fee conditions or available transaction structures change. A reader should not equate “not sensible to spend today” with “cryptographically impossible to spend forever.”
The Bitcoin Developer Transactions guide discusses transaction outputs, fees and policy considerations relevant to small amounts.
This is why practical spendable supply can depend not only on keys and scripts but also on processing costs and the user’s actual intended transaction.
Bitcoin market supply is another separate question
Even a bitcoin that is currently spendable by its owner is not necessarily offered for sale. People can hold their assets for many reasons, and a liquid market’s available orders represent only what participants are willing to trade under the quoted conditions.
Suppose the school has 500 accessible ART credits, but students choose to offer only thirty at the exchange desk today. The available market supply is thirty under that simplified observation, not all 500 accessible credits and not the lifetime programme ceiling.
The same distinction applies to Bitcoin exchanges and liquidity pools. A market quote can reflect a particular venue, time and trade size without guaranteeing that every holder will sell at that price.
This matters when people infer future price increases from a limited issuance schedule. Supply constraints may matter, but demand, market depth, competing assets and user expectations remain independent variables.
No blockchain consensus rule guarantees a particular market price simply because the issuance schedule is predictable.
Read Why Token Market Cap Is Not Cash in the Treasury for the broader difference between an observed price, aggregate valuation and executable liquidity.
Wrapped Bitcoin tokens are not new native BTC issuance
Some other blockchains support tokens that represent Bitcoin under bridge, custodian or issuance arrangements. These may use familiar names or symbols suggesting a relationship to BTC.
Such wrapped representations are not automatically native bitcoins newly created under Bitcoin’s consensus issuance rules. Their backing, redemption and transfer conditions belong to the particular issuer or cross-chain mechanism.
Imagine the school deposits ten actual ART credits into a trusted cabinet and issues ten paper receipts elsewhere. The receipts are useful under the agreed backing arrangement, but counting both the locked original credits and the receipts as twenty newly issued ART would double-count the underlying economic claim.
The same concern affects aggregate reports spanning networks. A balance labelled BTC-like on another chain can be technically valid without being an additional Bitcoin mainnet coin.
If a bridge fails or an issuer cannot redeem, the value of the representation may differ from the asset it claims to track.
For more, read What a Blockchain Bridge Does and Does Not Guarantee.
What the 21-million headline cannot tell you
The long-run supply ceiling is an important protocol property. But it cannot tell you precisely how many private keys are lost, who intends to sell tomorrow, how much liquidity exists on a particular exchange or what future market price people will offer.
It also cannot tell you whether an unrelated token with BTC in its symbol has valid Bitcoin backing. Network identity and custody evidence remain essential.
The maximum does not mean all permitted units have already been issued, that every output is currently transferable or that there is one global list of owners with real-world names.
A useful report therefore identifies the chosen supply measure, the observation block, the method and any assumptions about loss, accessibility or market availability. An analyst’s estimate can be useful when its uncertainty is honestly described.
A broad investment slogan such as “only 21 million therefore guaranteed profits” has no mathematical foundation in the issuance constraint alone. Scarcity is not a promise about demand.
Students should learn the positive point without the exaggeration: a predictable issuance rule is verifiable under consensus, while accessibility, market participation and economic value are different questions.
A school worksheet with four supply columns
Draw four columns called Lifetime Ceiling, Actually Issued, Currently Accessible and Offered for Exchange. The school reward programme promises a maximum of 1,000 ART credits over its life.
At the current date, 600 have been issued. Of those, 50 are time-locked, 20 are in a demonstrably inaccessible account under our fictional rules and another 100 are controlled by students who choose not to spend them.
The accessible total depends on which categories and time constraints the worksheet includes. Market availability will be lower still if only a fraction of accessible holders actually offer their units.
Now ask whether 1,000 means there are 1,000 active tokens today. No. Ask whether the 100 deliberately saved credits are permanently lost. No. Ask whether a person finding the lost cupboard key might change the accessible number. Under the fictional model, yes.
Finally let the students convert whole credits to smaller units and record a small fee required for spending. The precision and economic cost do not alter the lifetime issuance ceiling.
The exercise uses paper and imaginary numbers. It trains learners to separate protocol supply, state, authority and market activity before drawing conclusions.
Seven questions for a Bitcoin supply claim
Which number is being quoted? Lifetime maximum, issued supply, UTXO amount, estimated accessible supply and exchange-available supply are different quantities.
What block or date applies? Actual issuance grows under the schedule, and accepted chain state must be measured at a defined point.
Is the inability to spend provable? A restrictive script or unclaimed reward is different from an old address that has not moved.
Could a key still exist? Dormant outputs are not necessarily permanently lost. Avoid presenting guesses about key access as certain facts.
Are there time or script conditions? Some outputs exist but cannot be spent until defined requirements are met.
What asset and network are involved? Wrapped representations on other chains should not be counted as new native Bitcoin issuance.
Does the statistic imply a price? No. Market liquidity, bids, demand and obligations require separate evidence.
A careful statement says which supply category was measured and what assumptions were needed. That is more useful than reciting only the largest or most dramatic number.
Frequently asked questions
Is Bitcoin limited to 21 million coins?
Bitcoin has a long-run issuance ceiling conventionally described as approximately 21 million BTC under the current protocol rules.
Do all 21 million bitcoins already exist?
No. Bitcoin’s subsidy schedule creates units progressively through accepted mining rewards, with issuance declining over time.
Why are some bitcoins lost?
Owners may permanently lose access to required private signing information. Those units can remain recorded even when nobody can currently authorise spending.
Can the blockchain tell exactly how many private keys are lost?
No. An old unspent output does not prove its key is gone. Some unspendability can be demonstrated technically, while human key loss is often uncertain.
What is a UTXO?
It is an unspent transaction output with a defined satoshi amount and spending conditions. Bitcoin wallets commonly aggregate eligible UTXOs into displayed balances.
What is a satoshi?
It is one hundred-millionth of a bitcoin, the usual smallest on-chain unit.
Does dormant mean unspendable?
No. A long-term holder might simply choose not to transact. Dormancy and proven loss must be reported separately.
What is a provably unspendable output?
It is an output whose script rules do not permit ordinary spending under the protocol, such as relevant OP_RETURN null-data outputs.
Do timelocked coins count as existing?
Yes, they can remain recorded while their spending conditions prevent current use. Temporary restriction is different from permanent loss.
Do wrapped BTC tokens add to Bitcoin’s 21-million limit?
No. They represent separate tokens or claims on other networks under their own backing arrangements, not new native Bitcoin issuance.
Does a limited supply guarantee Bitcoin’s price will rise?
No. Demand, competition, market liquidity and broader conditions can change independently of the supply rule.
What should students remember?
Bitcoin’s maximum issuance, actual existing outputs, spendable holdings and coins available in the market are four different questions.
Primary sources and further learning
The Bitcoin FAQ explains the long-run supply ceiling, mining issuance, divisibility and the effect of lost keys. The Bitcoin Developer Transactions guide explains UTXOs, script conditions, locktimes and transaction-fee considerations. The Glassnode Provably Lost Supply definition gives examples of narrower demonstrable categories such as unclaimed rewards and specified provably unspendable outputs; those are not a complete census of lost keys.
For further eduKateSG reading, see Proof of Work and Cumulative Work, UTXO Transactions and Change Outputs, Why Proof of Work Needs Mining, Why a Wallet Password Is Not a Recovery Phrase and What a Blockchain Bridge Does and Does Not Guarantee.
The lasting lesson is that a supply rule can be precise while the number of coins people can actually access or will sell remains uncertain. That distinction is central to understanding Bitcoin without exaggeration.