March 15 2026
A remarkable development in the Bitcoin community emerged from the innovations of developer Martin Habovštiak, who managed to embed a 66-kilobyte image in a single transaction without using the OP_RETURN or Taproot methods. This experiment is not only an artistic expression; it also serves as proof that closing off a single data path doesn't eliminate the ability to process data on the blockchain, but simply changes where it resides.
Deze demonstratie is bijzonder relevant te midden van de meest controversiële governance-discussies die Bitcoin in jaren heeft gezien. Enerzijds is er een groep die pleit voor striktere filters om “spam” van de blockchain te houden. Anderzijds betoogt een andere fractie dat zulke zware restricties mensen in ongewenst gedrag zouden dwingen en grootminingbedrijven bevoordelen. Habovštiak’s experiment biedt steun voor het laatste argument: filters veranderen simpelweg de richting, maar voorkomen niets.
Habovštiak has accompanied his findings with a transaction ID and a verification method. Interested users can query the Bitcoin CLI, and with a simple reconstruction, they can reassemble the file. This method bypasses the much-discussed data storage paths, namely the OP_RETURN field, recently relaxed by Bitcoin Core, and Taproot's witness structure, which has enabled numerous inscriptions. Bitcoin transactions are actually bytes, and nodes maintain that these bytes satisfy certain conditions such as valid signatures and proper formatting. What they do not enforce, however, is that bytes should only represent "money." If someone constructs valid transaction bits that also form a valid image file, these are stored and distributed within the network.
Bitcoin has the capacity to complicate certain data patterns through software defaults. However, it cannot completely eliminate them without addressing the economic interests of miners head-on.
Bitcoin operates with two layers of rules. Consensus rules determine which blocks are valid; policies determine which transactions are passed on by individual nodes and which are generally accepted by miners in mempools. Policies can slow behavior and impose additional fees, but they cannot guarantee that transactions with relevant consensus validity will be excluded, especially if they incur high fees.
Miners can withdraw any consensus-valid transaction, especially if it reaches them via paths that bypass the regular node network. This demonstrates that the OP_RETURN size was a policy choice, not a consensus limit. Throughout Bitcoin Core's history, these limits have always been seen as a kind of standardization, with developers arguing that strict limits force users into undesirable encryption practices, such as hiding data in issues that appear spendable, unnecessarily expanding the UTXO set.
Habovštiak's demonstration makes this abstract discussion tangible. It limits one method, but leads to engineering efforts in a different direction.
When many nodes refuse to submit "non-standard" transactions, economic incentives create alternatives. Mining pools accept transactions directly, bypassing relay networks. Services like MARA's Slipstream function as direct submission systems for larger or non-standard transactions that are often rejected by nodes from mempools, even when they meet consensus rules. This creates a centralization risk that can be exacerbated by stricter filters, as only miners and specialized services are able to process these transactions into blocks with certainty.
Met een kostprijs van 10 satoshi per virtuele byte bedraagt de prijs voor één megabyte blockruimte ongeveer 0,1 BTC, terwijl het bij 50 satoshi per byte ongeveer 0,5 BTC is. De vraag wat mensen bereid zijn te betalen wordt op deze manier cruciaal.
The experiment comes at a crucial time in the Bitcoin community, where BIP-110 is being discussed. This proposal will temporarily limit the data carrying capacity of consensus-level transactions for about a year. Among other things, the proposal will ensure that new output scripts larger than 34 bytes are invalid, with the exception of OP_RETURN outputs, which can be up to 83 bytes. It will also impose limits on payload sizes and the number of elements in the witness stack, with a general limit of 256 bytes, with some exceptions. Proponents view BIP-110 as a measure to protect node operators from rising storage costs, while critics warn of side effects and implementation risks. This proposal represents a shift from policy-level filtering to consensus restrictions, which has far-reaching governance implications.
Habovštiak's experiment offers valuable insight: even with consensus constraints, there is pressure to conform. He emphasizes that BIP-110 would invalidate his specific construction, but at the same time, he can produce alternatives with different encodings. This dynamic persists: if you restrict one pattern, incentives and ingenuity push data elsewhere.
Restricting popular data patterns can backfire by pushing users toward alternative encodings that incur higher network overhead. When developers create issues that appear spendable but essentially contain arbitrary data, they increase the UTXO set—the database of unused issues that every full node must maintain in accessible memory.
The growth of this set imposes a more persistent burden than witness data or OP_RETURN payloads, which can be discarded. An output containing, for example, an image file remains in this set until someone releases it, potentially indefinitely. This statement reveals Bitcoin Core's historical reluctance to impose strict limits on OP_RETURN. Filters that initially appear protective can actually increase long-term costs for nodes, thus undermining the decentralized objective.
The economic dynamics of enforcement point to three possible scenarios. The first path preserves the status quo: price it, don't ban it. Arbitrary data remains, primarily regulated by the fee market. When block space becomes scarce, data-rich transactions are inherently excluded monetarily. The mechanism becomes more economic than technical. The second path tightens policy filters while leaving the consensus level unchanged. Data shifts to more difficult-to-filter encodings and direct submission systems for miners.
The risk of centralization increases because only miners and specialized pipelines can process these transactions with certainty. The third path implements consensus restrictions like those in BIP-110. While popular patterns appear to temporarily decline, the adaptation continues as new encodings emerge. This can lead to further collateral damage if limits push data into outputs that further expand the UTXO set. Governance risks increase if controversial consensus changes introduce coordination challenges and the possibility of network splits.
Three indicators will determine which scenario unfolds. First, miner behavior: will mining pools continue to accept non-standard transactions through direct channels? Services like Slipstream exist specifically to address this, and their continued existence reveals miner priorities. Second, governance developments surrounding BIP-110: will the proposal garner significant support beyond the debates? It requires coordinated activation within a decentralized network, making political feasibility as important as technical merit. Third, the secondary effects: will the restrictions push more data to encodings that increase the burden on nodes? The growth of UTXOs during periods of policy tightening would provide important data here.
Those who oppose on-chain data storage outside of financial transactions may draw an uncomfortable conclusion from Habovštiak's demonstration: you probably can't prohibit it. It's possible to price it through fee markets, discourage it through policy standards, or increase friction through implementation complexity. However, complete prevention requires accepting economic constraints beyond your control or implementing consensus restrictions that carry their own risks. Bitcoin validates the structure of transactions, not their meaning, and the protocol doesn't distinguish between "money transactions" and "data transactions." This differentiation, after all, is interpretation, something the network can't perform.
The real debate isn't whether Bitcoin has the technical means to prevent data arbitrariness—the implicit conclusion is that it's difficult and possibly even impossible. The debate revolves around the choice of network technology: centralization toward miners who bypass filters, governance risks from controversial consensus changes, or higher long-term costs due to poorer encryption. Habovštiak's illustration confirms that the filters aren't functioning as advertised. What happens next depends on whether Bitcoin's users and developers accept this reality or continue to pursue technical solutions to what increasingly appears to be an economic and governance problem.
What does Habovštiak's demonstration mean for the Bitcoin community?
The demonstration shows that closing one data channel doesn't mean that such data processing disappears. It emphasizes that there are creative ways to transmit this data, even with strict filters.
Why are consensus rules important in this context?
Consensus rules are crucial because they determine which transactions are valid. They don't guarantee that data is treated solely as "money," but ensure that all transactions, including those involving non-financial data, can be processed if they comply with the rules.
What are the implications of BIP-110 for us as investors or policymakers?
BIP-110 can provide short-term stability by setting data limits. However, the risk of higher operational costs and potential governance conflicts in the future should prompt caution among investors and policymakers. It does not offer a permanent solution and may lead to undesirable side effects.
Bitcoin (BTC)
Ethereum (ETH)
XRP (XRP)
