A Bitcoin hard fork launched on September 1, 2026, replacing the SHA-256d proof-of-work algorithm with BLAKE2b. The chain, championed by longtime Bitcoin Core contributor Luke Dashjr, has attracted near-zero hashrate, no major exchange listings, and a single illiquid trading venue where the forke...
"Live by the fork, die by the fork." — Adam Back, CEO, Blockstream
A Bitcoin hard fork launched on September 1, 2026, replacing the SHA-256d proof-of-work algorithm with BLAKE2b. The chain, championed by longtime Bitcoin Core contributor Luke Dashjr, has attracted near-zero hashrate, no major exchange listings, and a single illiquid trading venue where the forked token (BTCB2) trades at $82 — roughly 99.9% below BTC's $77,000 price level.
The fork represents the culmination of a year-long dispute over non-financial data on Bitcoin's blockchain. BIP-110, the soft-fork proposal that preceded the hard fork, drew just 2.53% miner signaling support against a 55% threshold, produced two blocks in eight hours before stalling, and was formally marked Closed in the BIP repository. Dashjr was subsequently removed as a BIP editor within 26 hours of a motion filed by fellow editor Mark Erhardt. The episode has become a case study in how Bitcoin's informal governance mechanisms handle internal dissent.
The conflict originated in Bitcoin Core v30.0, released in October 2025, which raised the default -datacarriersize parameter from 83 bytes to 100,000 bytes. This change formalized the network's accommodation of Ordinals inscriptions, BRC-20 tokens, and Runes — protocols that embed non-financial data into Bitcoin transactions and consume block space.
Proponents of these protocols argued they generated fee revenue for miners and expanded Bitcoin's utility. Opponents, led by Dashjr, characterized the data as "spam" that bloated the blockchain, raised transaction fees for monetary transfers, and imposed storage costs on node operators. The debate was not abstract: Bitcoin blocks carrying heavy inscription payloads could push marginal fee-paying monetary transactions into later blocks or higher fee tiers.
BIP-110, formally titled "Reduced Data Temporary Softfork," proposed seven consensus restrictions with a one-year sunset clause. New output scripts would be capped at 34 bytes. OP_RETURN payloads would revert to the prior 83-byte maximum. Data pushes and witness stack elements would be limited to 256 bytes. Certain Taproot functionality would be restricted. The proposal explicitly grandfathered existing inscriptions.
The BIP-110 draft was completed on October 24, 2025, and formally assigned a BIP number on December 3. An activation client based on Bitcoin Knots was published on January 28, 2026. On March 1, Barefoot Mining produced the first signaling block via the OCEAN mining pool.
The mandatory signaling window opened on August 7, 2026, at block height 961,632. Activation required 55% of blocks within a 2,016-block difficulty adjustment period — approximately 1,109 blocks — to signal support.
The result was decisive. Only 51 blocks signaled in favor: 2.53% of the total. F2Pool explicitly refused to signal. No other major mining pool joined the effort. OCEAN was the primary and nearly sole source of signaling blocks.
At block 961,632, nodes running BIP-110 rules split off onto a minority chain. The Roughnecks mining pool produced two blocks on the fork chain over eight hours. The main Bitcoin chain advanced by 48 blocks in the same period. According to estimates cited by Michael Saylor, 99.85% of hashrate remained on the original chain. The minority chain inherited Bitcoin's full difficulty with roughly 0.15% of its hashpower, meaning the next 2,016-block difficulty adjustment would take an estimated 350 days to complete — effectively freezing the chain.
BIP-110 was marked Closed in the BIP repository.
Rather than accept BIP-110's rejection by miners, its supporters pursued a different strategy. On August 30, 2026, at block height 961,640, the chain hard-forked its proof-of-work algorithm from SHA-256d to BLAKE2b. The September 1 launch date served as the public-facing activation.
The rationale was straightforward: since SHA-256d miners had rejected BIP-110, the fork would use a different algorithm that rendered existing ASIC mining hardware incompatible. BLAKE2b was designed to allow CPU mining, theoretically lowering the barrier to participation. The fork also implemented BIP-110's data restrictions, capping arbitrary block data at approximately 300 kilobytes.
The chain branded itself as "Bitcoin" and its proponents, including Dashjr, asserted that the BTC ticker rightfully belonged to their chain. The official FAQ on btc-blake2b.org argued that Bitcoin's identity follows its rules, not its hashrate. This framing echoed arguments from prior Bitcoin forks, none of which succeeded in claiming the Bitcoin name in practice.
The market's verdict was swift and largely negative.
No major exchange — Coinbase, Binance, Kraken, or others — listed the forked token. The sole venue offering BTCB2 trading was Neoxa Exchange, a small, unverified custodial platform. On Neoxa, BTCB2 bids stood at $82 and asks at $190, producing a spread of $108, or 131.7%. Ask prices extended as high as $389. At $82, BTCB2 traded approximately 99.9% below BTC's concurrent price of roughly $77,000.
Liquidity was negligible. Even small orders could produce significant price swings in either direction. The bitcoinbip110.org website itself warned users that Neoxa was a "high-risk, unverified custodial exchange" and urged caution about promoting its price as a reference point.
The situation deteriorated further when a screenshot surfaced from the Bitcoin Knots Discord server. An individual identified as Andy, associated with Neoxa Exchange, posted: "Hey guys, after all the hate we got from you guys, I will rather delist the coin. There be lots of time to withdraw your coins." As of September 3, the delisting had not been confirmed, but the statement underscored the fragility of the fork's single trading venue.
A critical technical risk compounded the market problem. The BLAKE2b fork lacked replay protection. Transactions broadcast on one chain could potentially be replayed on the other, creating the possibility of unintended fund transfers. Bitcoin developer Kevin Loaec and hardware wallet manufacturer Ledger both issued warnings about this vulnerability.
The BIP-110 episode triggered Bitcoin's most significant governance action in years. On August 9, 2026, at 18:53:50 UTC — two days after the soft fork stalled — BIP editor Mark Erhardt (known as Murch) filed a formal motion to remove Dashjr as a BIP editor.
Erhardt cited four grounds: heavy involvement in BIP-110's creation combined with inconsistent editorial standards; attempting to assign a BIP number via Twitter before mailing list discussion and rapidly merging a pull request; minimal participation, with fewer than 1% of BIP editor comments since additional editors were appointed in April 2024; and championing a contentious fork that undermined community trust.
Dashjr responded 49 minutes later, calling the accusations "false" and stating he had followed the BIP process consistently for years.
The removal was confirmed on August 10 at 20:52 UTC by fellow editor Bryan Bishop on the Bitcoin Development Mailing List — 25 hours and 58 minutes after the motion was filed. Pull request #2248, a one-line deletion of Dashjr's name from the BIP 3 governing document, was merged by editor Jon Atack. The vote was recorded via GitHub reactions: more than a dozen thumbs-up reactions and three explicit "ACK" comments.
Two minutes after the removal confirmation, Dashjr announced a sabbatical from OCEAN, the mining pool he co-founded. No successor was named and no duration was stated.
The remaining five BIP editors are Bryan Bishop, Jon Atack, Mark Erhardt, Olaoluwa Osuntokun, and Ruben Somsen.
A notable procedural gap emerged: BIP 3, the document governing the BIP editorial process, contains no formal removal procedure and no appointment guidelines. The removal was executed through informal consensus among editors — a process that, while effective, exposed the absence of codified governance rules for Bitcoin's standards body.
The BLAKE2b fork joins a long list of Bitcoin chain splits. According to available data, there have been approximately 105 Bitcoin forks to date. None has approached BTC's market position.
Bitcoin Cash (BCH), the most prominent fork, split at block 478,558 on August 1, 2017, with backing from major miners, exchanges, and a prominent advocacy campaign. As of mid-2026, BCH trades at less than 1% of BTC's value. Bitcoin SV (BSV), which forked from BCH in November 2018 with support from Craig Wright and Calvin Ayre, has declined further.
The pattern is consistent: forks that lack majority hashrate, exchange support, developer continuity, and user adoption fail to sustain economic value regardless of their technical arguments. Michael Saylor characterized BIP-110 as Bitcoin's "biggest self-inflicted risk," arguing that rejecting valid, fee-paying transactions "could be used to censor any class of Bitcoin activity." Adam Back identified "fatal flaws" in the activation mechanism.
David Schwartz, Ripple's former CTO, dismissed the claim that the BTC ticker could follow a minority chain as "nonsense."
The BLAKE2b fork starts from a substantially weaker position than BCH did in 2017. BCH had exchange listings, mining pool support, and institutional backing at launch. BTCB2 has a single unverified exchange, near-zero hashrate, and no wallet or Lightning Network implementation support.
The BLAKE2b fork is, by measurable standards, the least-supported Bitcoin chain split in recent history. Its hashrate is negligible, its sole exchange venue is flagged as high-risk, and its replay protection is absent. The governance repercussions — Dashjr's removal and the exposed procedural gaps in BIP 3 — may prove more consequential than the fork itself.
The underlying dispute over non-financial data on Bitcoin's blockchain remains unresolved. Ordinals and BRC-20 activity continues on the main chain. Miners have demonstrated through BIP-110's rejection that they will not voluntarily restrict fee-paying transactions, regardless of data content. Whether future proposals find more nuanced approaches to block space allocation — or whether the market resolves the question through fee dynamics alone — remains an open question.
What the data shows is that Bitcoin's consensus mechanism works as designed: minority positions cannot impose rule changes, and the market prices forked chains accordingly.