On August 1, 2017, one blockchain became two. At block 478,558, a group of miners and developers began producing blocks the rest of the Bitcoin network considered invalid, and Bitcoin Cash was born. Anyone holding bitcoin at that moment suddenly held coins on both chains.
That event was a Bitcoin hard fork: a change to Bitcoin's consensus rules that is not backward compatible, splitting the blockchain into two separate chains that each follow their own rules and carry their own coin. Hard forks are how a network with no CEO and no board resolves its deepest disagreements.
This guide explains what a hard fork is, how it differs from a soft fork, every major fork in Bitcoin's history, what a chain split means for your coins, and why 2026 has produced the most serious fork debate since the blocksize war.
Manage your Bitcoin securely with the self-custody NewsPulse Wallet app.
Key takeaways
- A Bitcoin hard fork is a backward-incompatible rule change that permanently splits the blockchain into two chains, usually creating a new cryptocurrency in the process.
- A soft fork tightens the rules and stays compatible with older software. Bitcoin's major upgrades, including SegWit (2017) and Taproot (2021), were soft forks.
- Bitcoin Cash (BCH), created in August 2017 during the blocksize war, remains the most significant hard fork of Bitcoin.
- If you hold your own private keys at the moment of a chain split, you automatically hold coins on both resulting chains.
- Bitcoin has never carried out a deliberate, consensus-breaking hard fork on its main chain, which is exactly why the 2026 quantum-resistance proposals (BIP-360 and BIP-361) are so contentious.
What is a Bitcoin hard fork?
A hard fork is a change to a blockchain's protocol that breaks compatibility with previous versions of the software. Nodes running the old rules reject blocks produced under the new rules, so the network splits into two chains that share an identical history up to the fork point and then diverge permanently.
The word "consensus" does a lot of work here. Every Bitcoin node independently checks each block against the same set of consensus rules: the 21 million coin supply limit, the block size and weight limits, valid signature formats, and so on. These rules are what make thousands of strangers agree on a single ledger without trusting each other. Bitcoin's own developer documentation describes forks as the natural result of nodes enforcing different rule sets. Change the rules in a way old software cannot accept, and you no longer have one network. You have two.
Because Bitcoin has no central authority to force an update, a hard fork only succeeds if miners, node operators, exchanges, and users voluntarily move to the new rules. When only part of the community moves, both chains survive, each with its own coin, market price, and development team.
One distinction trips people up constantly. A chain fork (like Bitcoin Cash) splits an existing blockchain, so both coins share transaction history up to the split. A codebase fork (like Litecoin) copies Bitcoin's open-source software and launches an entirely new blockchain from its own genesis block, sharing no history with Bitcoin at all. Only the first kind gives existing holders new coins.
Hard fork vs soft fork
The difference comes down to backward compatibility. A soft fork makes the rules stricter, so blocks created under the new rules still look valid to old software. A hard fork makes previously invalid blocks valid (or otherwise breaks the old rules), so old software rejects the new chain.
A useful analogy: a soft fork is like a restaurant trimming its menu. Regulars can still order everything they see. A hard fork is like the restaurant switching to a cuisine the old menu cannot describe. Diners must either accept the new menu or find another restaurant.
| Hard fork | Soft fork | |
|---|---|---|
| Backward compatible | No | Yes |
| What happens to non-upgraded nodes | Rejected from the new chain, remain on the old rules | Continue operating, just cannot use new features |
| Chain split | Yes, if any miners keep the old rules | No, as long as the majority of hash power upgrades |
| Creates a new coin | Usually | No |
| Coordination required | Near-universal agreement | Majority of miners |
| Bitcoin examples | Bitcoin Cash (2017), Bitcoin Gold (2017) | SegWit (2017), Taproot (2021) |
Soft forks are how Bitcoin has actually upgraded. Segregated Witness (SegWit) activated in August 2017 after a grassroots user-activated soft fork (UASF) campaign pressured miners to signal support. Taproot followed in November 2021, improving privacy and enabling more flexible smart contracts. Both changed Bitcoin meaningfully without splitting the network.
How does a hard fork happen?
A hard fork is a process, not a single event. The typical sequence looks like this:
- Proposal. Someone drafts the change, often as a Bitcoin Improvement Proposal (BIP) submitted to the public BIPs repository on GitHub, where it gets a number and community scrutiny.
- Implementation. Developers write the new rules into a client, either as a change to an existing implementation or as an entirely new one, the way Bitcoin ABC implemented Bitcoin Cash's rules in 2017.
- Activation. The software specifies a trigger, usually a block height or a timestamp (a "flag day"). Until that point, both rule sets see the same chain.
- The split. At activation, the first block that is valid under the new rules but invalid under the old ones divides the network. Miners, nodes, and wallets sort themselves onto one side or the other.
- Two networks. From that block forward, each chain has its own transactions, its own difficulty adjustments, and its own market. Exchanges list the new coin (or refuse to), and the market decides what each chain is worth.
Well-designed hard forks add replay protection at step 2: a tweak to the transaction format so that a transaction broadcast on one chain cannot be copied, or "replayed," on the other. Bitcoin Cash included it. SegWit2x's decision to make replay protection opt-in rather than mandatory was one of the reasons that fork collapsed before launch.
Why do hard forks happen?
Every Bitcoin hard fork traces back to a disagreement the community could not resolve within the existing rules. The recurring causes:
- Scalability. How many transactions should fit in a block? The block size limit dispute drove the biggest forks in Bitcoin's history.
- New features. Some changes, like altering the proof-of-work mining algorithm, simply cannot be done as a soft fork.
- Security and bug fixes. A critical vulnerability can force incompatible changes. Bitcoin's 2010 overflow bug (covered below) is the canonical example.
- Ideology. Should Bitcoin be digital gold or electronic cash? Should lost coins ever be touched? Technical questions are often proxies for philosophical ones.
- Reversing transactions. Not a Bitcoin example, but Ethereum's 2016 hard fork to undo the DAO hack (which produced Ethereum Classic) shows a chain split used to rewrite recent history.
The Bitcoin hard fork list: a complete history
Understanding Bitcoin's forks means understanding the blocksize war, the community's defining conflict from roughly 2015 to 2017. Bitcoin's blocks were capped at 1 MB, and as adoption grew, blocks filled up and fees rose. One camp wanted bigger blocks so Bitcoin could scale as peer-to-peer cash on the base layer. The other camp argued big blocks would make running a node too expensive, centralizing the network, and preferred keeping blocks small while scaling through soft forks like SegWit and second layers like the Lightning Network.
Early attempts to raise the limit came as alternative node software: Bitcoin XT (2015), Bitcoin Classic (2016), and Bitcoin Unlimited (2016). None attracted enough support to activate, but they set the stage for what followed.
| Fork | Date | Fork block | Key change | Status as of July 2026 |
|---|---|---|---|---|
| Bitcoin Cash (BCH) | Aug 1, 2017 | 478,558 | Raised block size to 8 MB, later 32 MB; rejected SegWit | Active; the most widely traded and supported Bitcoin hard fork |
| Bitcoin Gold (BTG) | Oct 24, 2017 | 491,407 | Switched mining algorithm to Equihash for GPU mining | Active but marginal; suffered 51% attacks in 2018 and 2020 |
| SegWit2x (B2X) | Planned Nov 2017 | 494,784 (planned) | Would have doubled the block size to 2 MB | Cancelled on Nov 8, 2017, before launch |
| Bitcoin Diamond (BCD) | Nov 24, 2017 | 495,866 | Increased supply 10x, faster blocks | Largely inactive; minimal trading and development |
| Bitcoin SV (BSV) | Nov 15, 2018 | 556,766 (split from BCH) | Pushed block sizes to 128 MB and beyond, "Satoshi Vision" | Active; delisted from several major exchanges since 2019 |
| eCash (XEC) | Nov 15, 2020 | Split from BCH | Bitcoin ABC's chain with a development fund built into mining rewards | Active but small |
A few of these deserve more than a table row.
Bitcoin Cash emerged directly from the blocksize war's endgame. When SegWit locked in without a block size increase, the big-block camp forked off with 8 MB blocks. It is the only Bitcoin hard fork that built a lasting economy, with ongoing development, merchant use, and its own subsequent upgrade cycle.
SegWit2x is the fork that never happened, and its cancellation may be the most important event on this list. The "New York Agreement" of May 2017 saw dozens of companies and a majority of hash power commit to activating SegWit and then hard forking to 2 MB blocks six months later. Node operators and users revolted, seeing a backroom deal between corporations as an attack on Bitcoin's governance. Facing collapsing support and the lack of mandatory replay protection, the organizers called it off three weeks before activation. The lesson the community drew: miners and companies do not decide Bitcoin's rules, users running nodes do.
Bitcoin Gold aimed to democratize mining by switching to a GPU-friendly algorithm. Its later history became a cautionary tale instead: with only a small fraction of hash power protecting it, the chain suffered 51% attacks in May 2018 and January 2020, with attackers double-spending millions of dollars. A fork inherits Bitcoin's code, but not its security.
Bitcoin SV shows that forks can fork. In November 2018, the Bitcoin Cash community split over protocol direction, and two years later it split again, producing eCash. Each division fragmented the community and hash power further.
By most counts, well over 100 projects have forked Bitcoin's chain or code since 2009, the majority during the 2017-2018 "fork season" when new coins airdropped to bitcoin holders briefly seemed like free money. Nearly all of them are now ghost chains with no meaningful mining, development, or liquidity.
Has Bitcoin itself ever hard forked?
Here is the nuance most explainers miss: the chain we call Bitcoin (BTC) has never executed a deliberate, rule-loosening hard fork. Every intentional upgrade, from SegWit to Taproot, was engineered as a backward-compatible soft fork. The forks in the table above created new coins; Bitcoin itself kept its rules.
Two accidental events came close.
- The value overflow incident (August 15, 2010). A bug let someone create a transaction at block 74,638 containing 184 billion BTC, nearly 9,000 times the total supply cap. Satoshi Nakamoto and early developers shipped a patched client within hours, and the network abandoned the poisoned chain. The fix tightened the rules, making it technically a soft fork, but it remains the only time Bitcoin's ledger was deliberately rolled back.
- The March 2013 database fork. Bitcoin version 0.8 switched database software, and at block 225,430 it accepted a large block that version 0.7 nodes choked on. The network briefly ran as two chains, an unintentional hard fork. Developers and mining pools coordinated within hours to downgrade, deliberately abandoning the longer chain to restore consensus.
Both incidents shaped Bitcoin's deep conservatism about consensus changes. The community treats a contentious hard fork on the main chain as a failure mode, not a governance tool. Keep that in mind when reading the next section.
What a hard fork means for your coins
If a chain split happens while you hold bitcoin, the practical effects depend on where your coins sit.
- Self-custody. If you control your private keys at the fork block, those keys are valid on both chains. You automatically own the same balance of the new coin, no action required. This is why chain-split coins are often compared to an airdrop, although the mechanism differs: nothing is sent to you, your existing keys simply work in two places.
- Exchange custody. If your coins are on an exchange, the exchange holds the keys and decides whether to credit you the forked coin. Major exchanges credited Bitcoin Cash in 2017; many smaller forks were never supported. Policies vary, so never assume.
- Replay attacks. If a fork lacks replay protection, a transaction you broadcast on one chain can be copied to the other, spending coins you meant to keep. Waiting for wallet guidance and using wallets that split coins safely matters in the days after a fork.
- Volatility and scams. Fork events attract fake wallets and "claim your coins" phishing sites that harvest private keys. Never type a seed phrase into a website to claim a forked coin.
Taxes deserve a mention because they surprise people. In the United States, the IRS ruled in Revenue Ruling 2019-24 that new coins received from a hard fork are ordinary income at their fair market value when you gain control of them, and your cost basis is set at that value. Other jurisdictions treat forked coins differently, with some taxing them only when sold. Rules change, so check current guidance where you live. This is general information, not tax advice.
The quantum question: Bitcoin's next fork debate
For years, "will Bitcoin ever hard fork again?" was a theoretical question. In 2026 it has a live answer being argued in real time, and the trigger is quantum computing.
Bitcoin's signatures rely on elliptic curve cryptography, which a sufficiently powerful quantum computer could break, deriving private keys from exposed public keys. No such machine exists today, but the risk horizon is real enough that NIST finalized its first post-quantum cryptography standards in August 2024. The exposure is not evenly spread: according to a March 2026 Ark Invest analysis, more than a third of the bitcoin supply, roughly 8 million BTC, sits in outputs with publicly exposed keys, including early pay-to-public-key coins attributed to Satoshi Nakamoto.
Two proposals now frame the debate:
- BIP-360, merged into the BIPs repository in February 2026, defines a quantum-resistant address type that users could migrate to voluntarily. It is designed as a soft fork and does not force anyone to move.
- BIP-361, published in April 2026 by Jameson Lopp and five co-authors, goes much further. It would phase out spending from quantum-vulnerable address types on a fixed timeline of roughly five years, after which unmigrated coins are frozen. That includes an estimated 1 million BTC attributed to Satoshi and millions more that have not moved in over a decade. The full text is public in the BIPs repository.
The disagreement is sharp. Supporters argue that letting a future quantum attacker silently drain dormant coins and dump them on the market is the worst outcome, so a scheduled sunset is the responsible path. Critics, including Blockstream CEO Adam Back, favor optional upgrades and a migration window closer to a decade, arguing that quantum machines remain lab experiments. Others contend that invalidating existing signature types is a hard fork in everything but name, colliding with the anti-hard-fork culture described throughout this article, and that freezing coins crosses a line Bitcoin has never crossed: confiscation by consensus.
However it resolves, this is the most consequential fork debate since the blocksize war, and it tests the same question in a new form. Who decides what Bitcoin is: developers, miners, institutions, or the users running nodes? History suggests the answer is the same as in 2017, and that any change without overwhelming agreement will simply produce another line in the fork table above.
Conclusion
A Bitcoin hard fork is a backward-incompatible rule change that splits the blockchain into two chains, each with its own coin. Forks are Bitcoin's pressure valve: when a disagreement cannot be settled, the minority can exit with a copy of the ledger, and the market judges the result. As of July 2026, that judgment has been consistent. The original chain has kept the name, the security, and the overwhelming majority of value, while upgrading itself only through soft forks. The quantum debate will test whether that pattern holds through Bitcoin's third decade.







