🔖

Ordinals and the Inscription Era

5 min read

culture governance
Share: X Bluesky Mastodon Reddit HN

On December 14, 2022, a developer named Casey Rodarmor minted the first Bitcoin inscription, an image of a skull, embedded permanently in the witness data of a Taproot spend. Five weeks later, on January 20, 2023, he released ord 0.4.0 and declared inscriptions ready for mainnet. Within months, inscriptions had become one of the most controversial uses of Bitcoin block space in years, and one of the largest sources of fee revenue miners had ever seen.

The protocol leans on two features its designers never intended for this: SegWit's witness-data discount (2017) and Tapscript's removal of the 10,000-byte script-size limit (2021), which together make large payloads cheap and valid. Whether that makes inscriptions clever or abusive was the loudest debate in Bitcoin culture in 2023. By October 2025 the same fight over data on chain had moved to Bitcoin Core's OP_RETURN relay policy. The chain itself is neutral. Let's look.

Two ideas in one

Ordinal theory has two distinct components:

The two pieces compose: an inscription on sat N is, by ordinal-theory convention, "owned" by whoever currently controls the UTXO containing sat N. Trade the UTXO, you trade the inscription. NFTs, but built on a deterministic numbering scheme rather than a separate token contract.

The first inscription

Casey Rodarmor minted Inscription #0 on December 14, 2022, in block 767,430. The inscription is a small image of a skull, embedded in the witness of a P2TR script-path spend.

Open in playground → the first inscription's block
Open in playground → block stats around the genesis-of-inscriptions moment

To see the inscription itself, fetch the block at verbose=2 and look for the transaction with a large witness containing the OP_FALSE OP_IF "ord" ... envelope. The data inside is the raw image bytes (around 800 bytes for inscription zero).

Open in playground → full block contents, find the OP_IF envelope

How it works mechanically

A Tapscript can contain a branch that is parsed but never executed. Inscriptions exploit this: the inscription content sits inside a never-executed script branch of a Taproot output, structured as: 

<pubkey> OP_CHECKSIG   // the part that actually runs
OP_FALSE
OP_IF
  "ord"      // the protocol identifier
  OP_PUSH 0x01 // content-type tag (a one-byte data push, not OP_1)
  "image/png"  // content-type
  OP_0
  <raw payload bytes>
OP_ENDIF

Because OP_FALSE OP_IF ... OP_ENDIF never executes, the bytes in the middle don't have to be valid script; they can be anything. And because witness data gets a 4x discount in weight-unit accounting, large inscriptions cost roughly 1/4 what the same bytes would cost as non-witness data.

The fee revenue impact

Inscriptions filled blocks. By mid-2023, Bitcoin's mempool had hundreds of thousands of pending transactions, fee rates spiked dramatically, and miners earned more in fees than they had since January 2018. Some blocks had fee revenue exceeding their subsidy.

Open in playground → May 7, 2023: fees (6.70 BTC) beat the 6.25 BTC subsidy

The totalfee reported is in satoshis. Compare to subsidy (also in satoshis) for the fee-vs-subsidy ratio. During peak inscription activity, this ratio crossed 1.0 multiple times, a glimpse of what the post-subsidy economy might look like.

The debate

Inscriptions are controversial because:

There's no resolution. Bitcoin's design accepts both views: anyone willing to pay can use the bytes. Filtering inscription transactions at the mempool layer (which some node operators do) is a per-node policy decision, not a consensus rule. The chain you're querying right now contains well over a hundred million inscriptions (128 million as of October 2026). They're not going anywhere.

What it means

Inscriptions are an empirical answer to a long-running question: can Bitcoin sustain itself on fees alone? For a year and a half, from early 2023 to mid-2024, the answer occasionally became "yes, today it can" (the April 2024 post-halving spike was Runes, not inscriptions). That's the most important data point Bitcoin has ever produced about its post-subsidy future.

Whether the inscriptions themselves are art, gambling instruments, or just permanent jpegs, for a stretch in 2023 and 2024 they paid miners in a way the protocol's authors didn't anticipate; by October 2026 fees were back under 1% of the subsidy. The chain rewards block space; inscriptions taught the market what block space is worth when somebody's willing to pay for it.

Sources

Read more: The Halvings.