A Bitcoin developer embedded a 66-kilobyte picture inside a single transaction with out utilizing OP_RETURN or Taproot.
The transaction adopted consensus guidelines. Anybody can confirm the bytes utilizing normal node software program. Martin Habovštiak did not do that to make artwork, however to show that closing one information doorway does not take away the aptitude, it simply adjustments the place bytes conceal.
Replace (March 5, 2026, 15:30 UTC): Builders and node operators have continued debating Martin Habovštiak’s on-chain picture demonstration, with critics disputing whether or not the file is really “contiguous” and early node telemetry exhibiting restricted adoption of the proposed BIP-110 restrictions up to now. The underlying governance query, whether or not Bitcoin can technically stop arbitrary information storage, stays unresolved.
A Bitcoin developer embedded a 66-kilobyte picture inside a single transaction with out utilizing OP_RETURN or Taproot.
The transaction adopted consensus guidelines. Anybody can confirm the bytes utilizing normal node software program. Martin Habovštiak did not do that to make artwork, however to show that closing one information doorway does not take away the aptitude, it simply adjustments the place bytes conceal.
The demonstration lands amid Bitcoin’s most contentious governance battle in years. One faction needs stricter filters to maintain “spam” off the blockchain.
One other argues that harsh restrictions push folks into worse behaviors and benefit giant miners. Habovštiak’s experiment offers proof for the second place: filtering redirects quite than stopping them.
What really occurred
Habovštiak’s write-up features a transaction ID and verification technique.
Customers can run bitcoin-cli getrawtransaction, then xxd -r -p to reconstruct the file. The development avoids the 2 pathways most cited in information storage debates: the OP_RETURN discipline that Bitcoin Core lately relaxed, and Taproot’s witness construction that enabled many inscriptions.
Bitcoin transactions are bytes. Nodes implement that bytes observe structural guidelines, equivalent to legitimate signatures, correct formatting, and bonafide spending circumstances.
They do not implement that bytes “imply cash solely.” If somebody constructs legitimate transaction bytes that additionally kind a legitimate picture file, the community shops and relays them.
Bitcoin can discourage sure information patterns via software program defaults. It can not stop them with out straight confronting miners’ financial incentives.
The excellence no person explains
Bitcoin operates with two layers of guidelines. Consensus guidelines decide what blocks are legitimate. Coverage guidelines decide what transactions particular person nodes relay and what miners sometimes settle for into mempools by default.
| Rule layer | What it controls (plain English) | What it could’t assure | Why it issues right here |
|---|---|---|---|
| Consensus guidelines | What makes blocks/tx legitimate | Can’t implement “money-only that means” | If it’s legitimate, it may be mined |
| Coverage / standardness | What nodes relay / mempools settle for by default | Will be bypassed | Filters add friction, not certainty |
| Miners’ inclusion | What will get into blocks | Incentives override preferences | Charges can “purchase” inclusion |
| Direct submission pipelines | Bypasses relay community | Concentrates entry | “Pay-to-play” danger (Slipstream-type routes) |
Coverage can gradual habits, increase friction, and impose prices. It can not assure prevention if a transaction stays consensus-valid and pays ample charges.
Miners can embody any consensus-valid transaction, particularly when it reaches them via paths that bypass common node relay.
OP_RETURN measurement limits have at all times been coverage decisions, not consensus partitions. Bitcoin Core has traditionally handled these as standardness nudges, with builders arguing that harsh limits push folks into worse encodings, equivalent to stuffing information into outputs that seem spendable, bloating the UTXO set that each node should keep.
Habovštiak’s demonstration makes this summary argument concrete. Cap one technique, and engineering effort flows towards one other.
The pay-to-play drawback
Even when many nodes refuse to relay “non-standard” transactions, financial incentives create workarounds. Mining swimming pools settle for transactions straight, bypassing the relay community. Providers explicitly launched for this exist already.
MARA’s Slipstream operates as a direct submission pipeline for “giant or non-standard” transactions that nodes typically exclude from mempools even once they observe consensus guidelines. The service routes round defaults quite than breaking guidelines.
This creates a centralization vector that stricter filters could amplify. When common nodes will not relay sure transaction sorts, solely miners and specialised companies can reliably land them in blocks.
At 10 satoshis per digital byte, one megabyte of blockspace prices roughly 0.1 BTC. At 50 satoshis per byte, roughly 0.5 BTC. The “ban” query turns into “what is going to folks pay?”
BIP-110 and the governance battlefield
The demonstration arrives as Bitcoin debates BIP-110, a proposal to quickly prohibit data-carrying transaction fields on the consensus stage for roughly one yr.
| Area / space | What BIP-110 proposes (plain English) | What it’s making an attempt to stop | Fundamental tradeoff / danger |
|---|---|---|---|
| New output scripts | New scriptPubKeys > 34 bytes invalid (besides OP_RETURN allowance) | Knowledge stuffed into outputs | Danger of pushing information elsewhere |
| OP_RETURN exception | OP_RETURN allowed as much as 83 bytes | Small provable notes | Critics: nonetheless doesn’t “ban information” |
| Payload limits | Caps sure pushed information components (common 256-byte ceiling with exceptions) | Massive embedded blobs | Workarounds could emerge |
| Witness stack components | Limits witness factor sizes (common 256 bytes) | Inscription-style payloads | Would possibly redirect to worse encodings |
| Length framing | Non permanent (~1 yr) | Tactical slowdown | Implies “no clear everlasting repair” |
| Second-order impact | If information shifts into UTXO-like outputs | Keep away from long-term node burden | Backfire danger: UTXO bloat will increase |
The draft would make new output scripts exceeding 34 bytes invalid, apart from OP_RETURN outputs, which could be as much as 83 bytes. It additionally proposes limits on payload sizes and witness stack components, typically capping them at 256 bytes with slim exceptions.
Supporters body BIP-110 as a measure that protects node operators from runaway storage prices.
Critics warn about unwanted side effects and implementation dangers. The proposal represents an escalation from policy-level filtering to consensus-level restriction, a shift carrying governance implications past the instant technical query.
Habovštiak’s experiment feeds straight into this debate. It demonstrates that even consensus restrictions face strain to adapt. He notes BIP-110 may invalidate his particular development, but in addition that he may produce alternate options utilizing totally different encodings.
The underlying dynamic persists: squeeze one sample, and incentives plus ingenuity push information elsewhere.
The short-term framing, one yr quite than everlasting, acknowledges this actuality implicitly. A everlasting change would require confronting tougher questions in regards to the sustainability of enforcement.
A brief measure admits the issue could lack a clear technical answer, solely tactical administration with a restricted shelf life.
The worst-behavior drawback
Proscribing standard information pathways can backfire by pushing utilization towards encodings that impose increased community prices.
When builders create outputs that look spendable to hold arbitrary information, they enhance the UTXO set, which is the database of unspent outputs each full node should keep in accessible storage.
UTXO progress represents a extra persistent burden than witness information or OP_RETURN payloads, which could be pruned. An output that encodes a picture file stays within the UTXO set till somebody spends it, doubtlessly indefinitely.
The node price accumulates quite than growing older away.
This explains Bitcoin Core’s historic reluctance to impose harsh limits on OP_RETURN. The choice is not essentially higher. Filters that appear protecting can enhance long-term working prices for nodes, undermining the decentralization purpose they purpose to protect.
Three paths ahead
The enforcement economics recommend three situations.
The primary path maintains the established order: value it, do not ban it. Arbitrary information persists, ruled primarily by payment markets. When blockspace turns into scarce, data-heavy transactions are naturally priced out. The lever turns into financial quite than technical.
The second path tightens coverage filters whereas leaving consensus unchanged. Knowledge shifts towards harder-to-filter encodings and direct-to-miner submission. Centralization danger rises as a result of solely miners and specialised pipelines can reliably verify these transactions.
The third path implements consensus restrictions, equivalent to these outlined in BIP-110. Standard patterns could quickly decline, however adaptation continues as new encodings emerge. Collateral harm will increase if limits push information into outputs that bloat the UTXO set.
Governance danger escalates as contentious consensus adjustments increase coordination challenges and the potential for community splits.
What decides the result
Three indicators sign which situation materializes.
First, miner habits. Do mining swimming pools proceed accepting non-standard transactions via direct channels? Providers like Slipstream exist particularly for this, as their sustained operation reveals miner priorities.
Second, governance trajectory. Does BIP-110 collect significant adoption past debate? The proposal requires coordinated activation throughout a decentralized community, making political viability as vital as technical advantage.
Third, second-order results. Do restrictions push extra information into encodings that enhance node burden? UTXO progress charges throughout coverage tightening durations would offer empirical proof.
The uncomfortable actuality
For those who oppose on-chain information storage past monetary transactions, Habovštiak’s demonstration delivers an uncomfortable message: you most likely cannot ban it.
You possibly can value it via payment markets. You possibly can discourage it via coverage defaults. You possibly can increase friction via implementation complexity.
However full prevention requires both accepting financial constraints you can not management or implementing consensus restrictions that carry their very own dangers.
Bitcoin validates transaction construction, not that means. The protocol does not distinguish between “cash transactions” and “information transactions” as a result of that distinction requires interpretation that the community can not carry out.
The actual debate is not whether or not Bitcoin can technically stop arbitrary information, because the demonstrated reply is “not simply, and maybe under no circumstances.”
The talk is which tradeoffs the community accepts: centralization towards miners who bypass filters, governance danger from contentious consensus adjustments, or increased long-term prices from worse encoding decisions.
Habovštiak’s picture proves the filters do not work as marketed. What comes subsequent relies on whether or not Bitcoin’s customers and builders settle for that actuality or proceed pursuing technical options to what more and more seems to be an financial and governance drawback.
Latest developments (March 5, 2026, 15:30 UTC)
- March 2–4: Debate intensified on developer boards and social media after Bitcoin Knots maintainer Luke Dashjr argued the transaction didn’t really comprise a contiguous picture file on the protocol stage, difficult the framing of the demonstration whereas reaffirming his help for stricter limits on arbitrary information. Dashjr’s response targeted on how totally different definitions of “contiguous” apply to transaction parsing.
- Early March 2026: Node telemetry cited by a number of trade trackers signifies roughly 8.8% of reachable nodes at present sign help for the BIP-110 rule set applied in Bitcoin Knots, highlighting that the proposal stays removed from network-wide consensus.
- Early March 2026: Habovštiak stated he additionally constructed a model of the picture transaction that conforms to BIP-110 constraints throughout testing in a Knots regtest setting, however claimed the compliant model required extra whole information than the unique development, doubtlessly undermining the proposal’s purpose of lowering on-chain storage. Particulars of the check have been shared alongside verification directions in his public clarification of the experiment.
- Early March 2026: The proof-of-concept renewed scrutiny of Bitcoin’s governance break up between Bitcoin Core and Bitcoin Knots communities over whether or not coverage filters, payment markets, or consensus guidelines ought to finally regulate non-payment information on the blockchain. A number of builders emphasised that the demonstration itself doesn’t change consensus guidelines however illustrates how different encodings can emerge when particular pathways are restricted.
The follow-up dialogue suggests the demonstration has shifted consideration from whether or not the particular transaction works to a broader query: how far protocol adjustments can realistically go in limiting information patterns with out introducing new technical workarounds or financial incentives that push the habits elsewhere.
