Permanent bytes: what to store on Bitcoin, and how

Status on blockspacelabs.com: Research.

What it asks #

The card opens: "Every byte on Bitcoin is permanent and costly. What deserves that, and how is each byte best spent?" It splits the question into seven parts:

1. What's been stored: every inscribed byte by kind, over time, and what it paid.
2. Format efficiency: bytes per pixel by image format; how often brotli compression is used and what it saves.
3. Duplication: identical files inscribed more than once, and what the copies cost.
4. Reuse, not repeat: how many bytes recursion and delegation save.
5. Cost per permanent byte: fee per content byte, against the witness discount.
6. Link rot: inscriptions that point off-chain, and how many of those links still answer.
7. What lasts: which formats stay readable, and what belongs on-chain whole, as a hash, or not at all, "since nothing can be deleted".

The data it will use #

"The byte census, ord's inscriptions, the largest-inscriptions ranking, BTI books and fees." Each is already kept by Blockspace Holdings' node and indexers: the byte census, the ord index behind blockspace.to, the ranking at https://block.space/largest-inscriptions, and the three books.

The figure on the card #

The card shows one current figure: 128,169,266 inscriptions in ord, as of 2026-10-20 09:07:08 UTC. The wiki's own reading is 128,169,401 (as of block 972,304, 2026-10-20 10:17:29 UTC). No result of the study is claimed. Inscriptions explains how files are inscribed.

Related #

• blockspacelabs.com
• Inscriptions
• The byte census (T1)
• The three books
• The witness discount
• recursive inscription
• delegate inscription

Sources #

1. The blockspacelabs.com page (index.html) · blockspacelabs.com
2. Blockspace Labs facts (data/facts.json) · blockspacelabs.com
3. ord /status (the blockspace.to ord server) · blockspace.to/status
4. Blockspace Holdings records