Bitcoin's spam war splits again: a 'DOG Mode' client challenges BIP 110 from below

On 17 July 2026 a new Bitcoin implementation began circulating in developer channels under the name DOG Mode. Its pitch is deliberately simple: rather than ask miners to vote on shrinking the block template, it tells node operators they can relax the rules themselves, no consensus change required. Within hours, Runestone co-founder Leonidas had publicly endorsed the approach as a way to "expand Bitcoin transaction access" for the projects that have spent three years arguing that the chain was never meant to be a global filing cabinet (CryptoBriefing, 17 July 2026, 06:40 UTC).

The dispute is the same one that has split the network since Ordinals turned up in late 2022: how much non-financial data should a Bitcoin block be allowed to carry, and who gets to decide. DOG Mode's arrival crystallises a tactical question that the protocol's loudest fights have so far dodged. The bigger fight is still the consensus-level proposal known as BIP 110, which would restrict the kind of data that can ride through Bitcoin's relay layer. That fight is, by every visible measure, losing.

The proposal that almost no miner wants

BIP 110's authors want Bitcoin's consensus rules tightened so that arbitrary data appended to transactions in the witness field is rejected at the protocol level. The mechanism is straightforward in design and politically poisonous in execution: it requires an activated soft fork, which in Bitcoin means overwhelming miner signalling. As of mid-July 2026, that support has not materialised (CoinDesk, 17 July 2026, 05:32 UTC).

The arithmetic is brutal. A soft fork of this magnitude needs sustained hashpower backing measured in the high nineties of percent. BIP 110 has, in the words of one wire summary, "almost no miner support". Pool operators have, with limited exceptions, refused to commit publicly. The reason is not mysterious: the pools that dominate the network earn fees from every kind of transaction, including the ones BIP 110's sponsors consider spam. Restricting that revenue at consensus level is a redistribution, and the entities whose balance sheets would shrink are not in a hurry to vote for it.

That leaves proponents of a cleaner chain with a tactical problem. They can keep arguing. They can fork. Or they can route around the consensus bottleneck altogether.

DOG Mode as a consensus-light workaround

DOG Mode is the routing answer. Instead of asking the network to agree on a new rule, it asks individual node runners to relax a policy choice they already make every day. Bitcoin nodes already filter transactions before relay; that policy is local, not consensus. DOG Mode flips the default, telling operators who want a more permissive mempool to opt in with a flag rather than opt out by hand.

The political geometry is what makes the proposal interesting. It costs nothing to miners. It does not require a signalling campaign. It does not require pool coordination. It appeals directly to the constituency that BIP 110 alienates: the projects that issue inscriptions, run Runes-style protocols, and build wallets and marketplaces on top of Bitcoin's data layer. For them, DOG Mode is a survival mechanism. For chain-pragmatists, it is a quiet concession that the data layer is here to stay.

The involvement of Leonidas, the co-founder of the Runestone project, is the tell. Runestone sits at the centre of the Bitcoin data economy that BIP 110's authors would like to constrain. DOG Mode is, in effect, a counter-offer from inside that economy, pitched not as a competing standard but as a way to keep the chain usable without waiting for the consensus fight to resolve.

What the framing leaves out

The case for restricting non-financial data is not aesthetic. Bitcoin's block size is finite. Every byte used to store an image, a token pointer, or a recursive inscription is a byte unavailable for a payments transaction. During fee spikes the substitution is direct and visible: users willing to pay for block space are crowded out by users whose transactions are denominated in something other than money.

The case for leaving the data layer alone is also not aesthetic. It is the same case that ran underneath the Blocksize War of 2017: the market should decide what block space is for, and node operators who find the data layer harmful can already refuse to relay those transactions. Policy at the node, in other words, has always been the libertarian answer. DOG Mode just systematises the inverse: policy at the node, in the permissive direction.

A counter-narrative worth taking seriously holds that DOG Mode is not neutrality but capture. The projects with the loudest megaphones tend to be the ones whose business models depend on the data layer staying open. Their preferred default is therefore not a neutral default. It is the default that maximises their throughput. From this view, both BIP 110 and DOG Mode are efforts by interested parties to set the dial where they want it, dressed up as network hygiene.

What to watch next

Two things matter over the next quarter. First, whether any of the major mining pools break their silence on BIP 110. A single large pool announcing support would not be decisive, but it would reset the conversation. Without that, BIP 110's proponents are running out of road and will eventually have to choose between a contentious fork attempt, which carries its own existential risks, and a tactical retreat.

Second, whether DOG Mode finds a wallet or node distribution that gives it real-world users. The technical mechanism is trivial. The hard part is getting ordinary node operators to flip the switch, and getting wallets and explorers to treat DOG Mode-flagged transactions as first-class. If that distribution work happens, the proposal becomes a parallel network in everything but name. If it does not, DOG Mode will be remembered as a clever idea that arrived a year too late to settle a fight that has already moved on.

The honest reading is that neither side has won. Bitcoin's data layer is neither going away nor being voted out. It is settling, by attrition, into a permanent feature of the protocol, contested at the margins and accommodated in practice. The interesting question is no longer whether the chain will carry inscriptions and runes. It is who builds the rules that decide what they look like, and whether those rules are written by consensus, by policy, or simply by which client you happen to run.

Desk note: Monexus frames this as a governance fight inside one network rather than as a story about cryptocurrency in general; the wire coverage so far has emphasised the personalities, this desk is tracking the institutional question, who sets defaults when consensus is unreachable.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://t.me/CryptoBriefing
• https://en.wikipedia.org/wiki/Ordinals