Everyone knows Bitcoin has a hard cap of 21 million coins. What most people don’t know is that the effective supply is almost certainly much lower – and some of the best on-chain researchers in the space put the real circulating supply somewhere between 14 and 16 million BTC.
I’ve been tracking this data for years, and the honest answer to “how many Bitcoin are lost forever?” is: we don’t know for certain, but the estimates are large enough to materially change how you think about Bitcoin’s scarcity.
TLDR
- Chainalysis estimates ~3.7 million BTC are permanently lost; James Lopp’s on-chain analysis puts the number closer to 5 million since 2017 alone.
- The effective Bitcoin supply may be only 14-16 million coins out of the theoretical 21 million cap – making each surviving coin scarcer than the headline number suggests.
- Lost coins are a feature for long-term holders: every permanently lost coin increases the ownership percentage of everyone who still holds liquid Bitcoin.
Buy Bitcoin on Coinbase – easy, reliable custody.
The 21 Million Cap Is the Ceiling, Not the Floor
Bitcoin’s supply schedule is one of the most studied properties in finance. The protocol will never produce more than 21 million BTC – that’s hardcoded, and it’s the foundation of every scarcity argument Bitcoin bulls make.
But here’s what gets less attention: the 21 million figure is a ceiling, not a floor. It’s the maximum that can ever exist. How many are actually in play – meaning held in wallets that can be accessed, traded, or spent – is a completely different question.
I’ve written before about the supply math behind Bitcoin’s 20 million milestone and what it means for investors. The lost-coin analysis takes that a layer deeper.
How Chainalysis Estimates Lost Bitcoin
Chainalysis, the leading blockchain analytics firm, published analysis estimating that approximately 3.7 million BTC are permanently lost. Their methodology focuses on what they call “zombie UTXOs” – unspent transaction outputs that haven’t moved in years and show behavioral patterns consistent with abandonment rather than intentional long-term storage.
Their key criteria for flagging coins as likely lost:
- No movement for 5+ years combined with creation date and address type patterns from early Bitcoin eras
- Wallet software that is no longer supported and had known backup failures
- Address types associated with era-specific wallets that commonly had key loss issues (early Bitcoin-QT wallets, paper wallets from 2010-2012)
- Absence of any consolidation or dusting transactions that would suggest the holder is aware of and managing the wallet
Chainalysis is careful to frame this as a probability estimate, not a certainty. The core problem with on-chain analysis is that there’s no reliable way to distinguish between a coin that’s lost forever and a coin held by someone who simply hasn’t moved it in a decade.
That said, 3.7 million is their conservative estimate. Some analyses push higher.
James Lopp’s On-Chain Methodology: Up to 5 Million
James Lopp, CTO of Casa and one of the most respected technical Bitcoin analysts in the space, has conducted independent on-chain research suggesting the figure may be closer to 5 million BTC lost since 2017 alone.
Lopp’s approach uses UTXO age analysis – examining the age distribution of all unspent transaction outputs and applying models to estimate what fraction of aged UTXOs represent genuinely lost coins versus intentional long-term holding. His work accounts for: If you want to understand the supply-scarcity implications, see our breakdown of how much BTC you need to be wealthy by 2032.
- Dormant pre-2013 addresses – coins that were mined or received in Bitcoin’s first four years and have never moved. Many of these are from a period when Bitcoin had essentially no monetary value and wallet software was experimental.
- Known wallet failure events – cases where specific wallet software had documented backup issues, data corruption problems, or was discontinued
- Statistical mortality modeling – accounting for the probability that Bitcoin holders from 2010-2014 have passed away without proper key transfer
The honest caveat here is that any model like this involves assumptions. The 5 million figure from Lopp’s analysis should be read as a plausible upper estimate, not a verified fact. Chainalysis’s 3.7 million is more conservative and uses stricter methodology. The true number almost certainly sits somewhere in this range.
The Early Mining Era: Where Most of the Losses Happened
If you want to understand where the losses are concentrated, the answer is clear: 2009 to 2013 is the primary loss era for Bitcoin.
Here’s why this period was so different:
Bitcoin had almost no monetary value. When Satoshi Nakamoto mined the genesis block in January 2009, Bitcoin wasn’t worth anything. Early miners were experimenting with cryptographic protocol, not building a financial system. Coins were treated like tokens in a test environment. Many early miners deleted wallets, reformatted hard drives, or simply never backed up their private keys because there was nothing to protect.
The Chainalysis blockchain forensics data shows a clear concentration of dormant UTXOs from this era. UTXO age distribution analysis reveals that coins created between 2009 and 2013 are dramatically more likely to be permanently inactive than coins from any subsequent period.
Wallet software was primitive. Bitcoin-QT, the original Bitcoin client, had no seed phrase backup system. You backed up a wallet.dat file, period. If your hard drive failed, if you reformatted your computer, if you lost the file – the coins were gone. There was no recovery path. The contrast with modern hardware wallets and 24-word seed phrases couldn’t be more stark.
The hardware failure rate compounds over time. Hard drives fail. If a wallet.dat file from 2010 was on a hard drive that failed in 2013, those coins are permanently inaccessible. No recovery service, no brute force attack, no blockchain forensics can help – the private key simply no longer exists anywhere.
James Howells, the Welsh IT worker who famously threw away a hard drive containing 8,000 BTC, is just the most publicized example of a problem that happened thousands of times at smaller scales throughout Bitcoin’s early years.
Satoshi’s Coins: The Elephant in the Room
No analysis of lost Bitcoin is complete without addressing Satoshi Nakamoto’s estimated 1 million BTC.
The consensus estimate – based on analysis of the early mining patterns by Sergio Lerner and others – is that a single early miner (almost certainly Satoshi) mined approximately 1 million BTC in 2009 and early 2010 using a distinctive pattern recognizable in the blockchain data. These coins have never moved. Not a single satoshi.
There are two possible explanations: either Satoshi is still alive and holding (a possibility that can’t be ruled out), or those coins are effectively gone – either through death, deliberate destruction of keys, or hardware failure.
From a market perspective, the practical distinction matters less than it seems. Whether those 1 million coins are being intentionally held by a living Satoshi or are genuinely inaccessible, they represent approximately 4.7% of the total supply that has never entered circulation and shows no signs of doing so.
I’ve covered how this fits into Bitcoin’s broader scarcity case in my piece on Bitcoin vs. gold scarcity comparisons – the Satoshi coins are one of the most unique scarcity features in monetary history. Gold doesn’t have an equivalent: there’s no founding creator whose known holdings are permanently sequestered by mathematical certainty.
Deaths Without Key Transfer: The Ongoing Loss Rate
Here’s something that doesn’t get enough attention: Bitcoin is still being lost right now, every year, and will continue to be lost indefinitely.
People die. When they die, if they haven’t explicitly transferred their private keys or seed phrases to heirs, the coins become inaccessible. This is a fundamentally different dynamic than traditional assets. If someone dies holding stocks or real estate, those assets pass through probate and eventually land with heirs. The assets still exist in the economy.
Bitcoin held in self-custody with no inheritance planning simply disappears from the available supply permanently when the holder dies.
Mortality statistics suggest this is not a trivial number. If we assume rough estimates of 50-100 million Bitcoin holders globally, and apply a standard mortality rate, somewhere in the range of several hundred thousand to a million BTC may have passed out of accessible hands through death in the years since Bitcoin became meaningfully valuable (roughly 2017 onward). Most of these cases will never make headlines because the amounts involved are modest and the deaths unremarkable.
The proper planning antidote – hardware wallets with documented seed phrase inheritance, or multi-sig setups where heirs hold a key – is well understood but far from universally practiced. The lost-through-death category will grow steadily as the early adopter cohort ages.
Effective Supply: What 14-16 Million Actually Means
Pull all these estimates together and the picture that emerges is striking.
| Loss Category | Estimated BTC | Confidence |
|---|---|---|
| Chainalysis “permanently lost” estimate | ~3.7 million | Medium-high |
| James Lopp upper bound estimate | ~5 million | Medium |
| Satoshi’s coins (never moved) | ~1 million | High |
| Additional deaths/hardware failures (ongoing) | ~500K-1M | Low-medium |
| Conservative effective loss total | ~4-5 million | – |
| Circulating supply (out of ~19.7M mined) | ~14-16 million | – |
The practical implication: if you hold Bitcoin, you own a larger percentage of the actual liquid supply than your nominal BTC balance suggests. If 5 million coins are genuinely inaccessible, then each coin in circulation represents a larger claim on the total real-world supply than the raw math of 21 million would indicate.
This connects directly to the sound money properties that Bitcoin’s most committed advocates emphasize. Sound money requires not just a hard supply cap but predictable and non-manipulable scarcity. Lost coins, counterintuitively, reinforce the scarcity case rather than undermining it.
The Counterargument: Are They Actually Lost?
I want to be honest about the limitations here, because the “Bitcoin is even scarcer than you think” narrative can get oversimplified.
Dormancy is not proof of loss. A UTXO that hasn’t moved in 12 years might be in a perfectly accessible cold wallet held by a very patient long-term holder. Institutional cold storage often looks identical on-chain to abandoned wallets. If the Winklevoss twins or some early Bitcoin whale simply hasn’t touched their coins in a decade, their coins will show up in dormancy analyses – but those coins are very much not lost.
The models have wide confidence intervals. Chainalysis’s 3.7 million estimate is not a precise measurement – it’s a probability-weighted model. The actual figure could be 2 million or 5 million depending on the assumptions you feed in.
Dormant doesn’t mean gone forever. People sometimes rediscover old wallets. Hard drives get recovered. Seed phrases written on paper surfaces resurface. It’s not common, but it happens. Coins that appear in the “likely lost” analysis occasionally come back to life.
The honest framing is: a significant number of Bitcoin are almost certainly permanently inaccessible, the best estimates cluster around 3-5 million, and this increases effective scarcity meaningfully – but the exact number is unknowable with current technology.
How Lost Coins Affect the Supply and Demand Picture
The supply side of Bitcoin has always been its defining feature. I’ve analyzed the treasury demand dynamics post-halving and the math of new issuance – but lost coins add a dimension to that analysis that even many serious Bitcoin investors overlook.
New issuance from mining is approximately 450 BTC per day at current block reward levels (post-April 2024 halving). Annual new supply is roughly 164,000 BTC. If even 100,000-200,000 BTC per year are leaving circulation through loss and death, the net new liquid supply entering the market is considerably smaller than the raw issuance figure suggests.
This is a deflationary force operating quietly in the background of every Bitcoin supply analysis. It doesn’t move markets quarter-to-quarter, but over multi-year time horizons it compounds significantly.
How to Secure Your Bitcoin So It Doesn’t Become a Statistic
The implication most worth acting on: don’t let your Bitcoin become part of the next analysis’s “lost” estimate.
Modern security practices have largely solved the key management problem that destroyed so many early coins. A hardware wallet with a properly stored 24-word seed phrase is recoverable even if the hardware device itself fails. Multi-signature setups allow you to distribute risk so that losing one key doesn’t mean losing your coins.
More importantly, inheritance planning is the gap most Bitcoin holders don’t address. If you hold meaningful Bitcoin and die without leaving accessible instructions for heirs, those coins are gone. Estate attorneys who work with crypto clients, Ledger’s inheritance guidance, and services like Casa (Lopp’s company) exist specifically to address this problem.
The early mining losses were largely unavoidable given the technology available at the time. Losses happening today are almost entirely preventable with current tools.
Kraken: operating since 2011, no major hacks.
How Many Bitcoin Are Lost Forever? The Bottom Line
The 21 million cap is real and inviolable. But the effective supply of Bitcoin that can actually be accessed, traded, and moved is almost certainly 14-16 million at most, and possibly less. Chainalysis puts permanently lost coins at 3.7 million. James Lopp’s independent analysis suggests up to 5 million. Add Satoshi’s 1 million coins that have never moved, and the scarcity argument gets considerably stronger than the headline 21M number implies.
For income investors and long-term holders, the lost-coin picture reinforces rather than complicates the Bitcoin thesis. If you’re holding Bitcoin expecting it to function as a scarce, sound monetary asset over decades, the ongoing reduction in effective supply is working in your favor.
The one action item worth taking from this analysis isn’t philosophical – it’s practical. Make sure your coins aren’t next in the lost column. Proper key management, hardware wallet storage, and inheritance planning are the only things standing between your Bitcoin and the lost pile.
For a deeper look at how these supply dynamics stack up against gold, I’d recommend my piece on Bitcoin vs. gold in 2026 – the scarcity comparison goes deeper than most people realize, and the lost-coin data is one of the most underappreciated variables in that analysis.
