Bitcoin split into two chains at block height 961632 on the night of Aug. 7, after nodes running BIP-110 software began rejecting every block that did not signal support for the proposal.
At that height, AntPool mined the first standard block. Bitcoin’s main network accepted it, while BIP-110 nodes rejected it. A miner working through Ocean later produced an alternative block, and the BIP-110 chain followed that version instead.
Eight hours later, the result was stark. The minority chain had advanced only to block 961633, producing just two blocks in total. During the same stretch, Bitcoin’s main chain had moved on to block 961681, 48 blocks ahead. By technical standards, the fork had already failed in less than a day.
What BIP-110 was trying to do
BIP-110, short for “Reduced Data Temporary Softfork,” was published under the pseudonym Dathon Ohm. Bitcoin Core developer Luke Dashjr is widely seen as the original drafter and main advocate behind it.
The proposal aimed to use consensus rules to temporarily limit non-financial data storage in Bitcoin blocks for roughly one year. Its intended targets included Ordinals inscriptions, BRC-20 tokens, the Runes protocol, and other methods of writing arbitrary data such as images and text onto the blockchain through transaction scripts and witness data.
Supporters argued that this kind of data consumes scarce block space, pushes up transaction fees, raises the cost of running a full node, and pulls Bitcoin away from what they see as its core role as “sound money infrastructure.” Dashjr described such data as “spam” and argued that its long-term accumulation could threaten Bitcoin’s decentralization.
Critics took the opposite view. In their argument, Bitcoin block space belongs to anyone willing to pay to use it. Any valid transaction that pays sufficient fees should be accepted, and consensus rules should not decide which kinds of data are worthy of inclusion.
Michael Saylor published a 110-point essay in July opposing the proposal. His central line was: “Bitcoin does not need guardians of purity; it needs guardians of neutrality.” Adam Back also warned that pushing BIP-110 without broad consensus could split the network.
BIP editor Murch added a note when assigning the proposal its number, calling it “an ill-considered and unusually hasty soft fork proposal,” and said the number reflected procedural compliance rather than endorsement.
Why the chain produced only two blocks
The minority chain’s paralysis was built into the mechanics.
Bitcoin adjusts mining difficulty every 2016 blocks, with a target of one block every 10 minutes on average. At the moment of the split, the minority chain inherited the full difficulty settings of the main chain, but had access to only a small fraction of total hash power. Over the prior two weeks, only 2.53% of blocks had signaled support for BIP-110, far below the 55% needed for activation. Most of those signaling blocks came from Ocean.
That left the minority chain trying to solve puzzles set at 100% of main-chain difficulty with less than 3% of the main network’s hash power. Based on the observed block rate, the minority chain would need about 350 days to reach its next difficulty adjustment. The main chain would need 14 days.
Until that adjustment point, the minority chain was set to remain close to frozen. BIP-110 also required every block in the two-week window before block 963647 to signal support. At the current pace, the minority chain would not even reach the end of that window.
As of 11:40 UTC on Aug. 9, the main chain had advanced to block 961725. The minority chain was still stuck at 961633. None of the 113 newly mined blocks on the main chain had signaled support for BIP-110.
The practical risk for ordinary users
The minority chain’s stagnation does not remove the risk. Both chains accept the same signature format for transactions, creating a replay-attack window.
If a user tries to sell fork coins on the minority chain, assuming such a market even exists, they may sign a transaction sending those coins to a buyer. That same signed transaction would also be valid on the main chain. The buyer could then broadcast it on the main chain and withdraw the same amount of real BTC from the seller.
The minority chain’s extremely slow block production makes the problem worse. Confirmation times for fork-coin transactions would be very long, giving an attacker more time to act on the main chain. Because BIP-110 does not include built-in replay protection, any transaction involving fork coins carries the possibility of unintentionally losing BTC on the main network.
For most Bitcoin users, the safest course is to ignore the minority chain, avoid moving any fork coins, and stay out of any related trading.
The governance dispute did not end with the fork
BIP-110 failed, but the split exposed a governance divide that does not disappear with the chain’s collapse.
Bitcoin’s last major soft fork was SegWit, or BIP-141, in 2017. That upgrade went through two years of governance conflict and required 95% miner signaling to activate, eventually passing only narrowly. Taproot in 2021 gained broad consensus and activated smoothly. BIP-110 lowered the activation threshold to 55% and still managed only 2.53% support.
Before the fork, Dashjr sent a sharper signal. If BIP-110 failed, changing Bitcoin’s proof-of-work algorithm “may be the only option left,” he said, implying that node operators might try to bypass miners and change the rules directly if miners refused to cooperate. That would challenge a basic part of Bitcoin’s security model.
At its core, the dispute is about who gets to decide how Bitcoin block space is used.
Saylor’s position represents one answer: block space is a neutral public resource, and anyone who pays for it should be allowed to use it, regardless of transaction purpose. Dashjr’s position represents another: Bitcoin’s core function is money, and uses that move away from that function amount to misuse of limited resources; protecting the network’s long-term viability matters more than short-term freedom.
In this fork, miners cast their vote with hash power and chose neutrality. That outcome favors supporters of Ordinals and BRC-20, while handing a setback to those who want Bitcoin block space kept “clean.”
That does not mean the argument is over. As long as inscriptions and Runes continue to consume block space, the question of what Bitcoin should be used for is likely to return in new forms.
BIP-110’s two-block showing made one point plain: in Bitcoin, an idea without hash power behind it does not travel far.