What Are Bitcoin Ordinals? How Inscriptions Put Data on Satoshis

Bitcoin Ordinals are best understood as a way to identify individual satoshis and track them as they move through ordinary Bitcoin transactions. An inscription can attach digital content to one of those sats, creating a Bitcoin-native collectible or other data-bearing asset. Neither idea creates a new coin or a separate blockchain.

That distinction matters. Ordinals use Bitcoin’s existing transaction system, but they add a new way to interpret certain data and outputs. Supporters see digital artifacts and new fee-paying activity; critics worry about scarce block space and the long-term cost of storing and validating more data. Here is how the pieces fit together—and what they do not promise.

Start with the satoshi

Bitcoin is divisible into 100 million satoshis, or sats, per BTC. The Ordinals protocol gives each sat a theoretical serial number based on the order in which it was mined. The numbering is an interpretive scheme: Bitcoin’s consensus rules do not assign sats serial numbers or treat one sat as more special than another.

To follow a numbered sat through transactions, Ordinals applies a simple convention: satoshis flow from transaction inputs to outputs in order. Imagine a transaction with two inputs containing 2 sats and 3 sats, and outputs worth 4 sats and 1 sat. Under this convention, the first four input sats go to the first output and the last one goes to the second.

Wallet software does not ordinarily need to know which sat is which to send bitcoin. But an Ordinals-aware wallet must account for input and output ordering if it wants to preserve a particular sat. A normal-looking transaction can therefore move a collectible sat unintentionally if the sender is not using compatible tools.

What an inscription is

An inscription associates content—such as text, an image, or another file—with a satoshi. The Ordinals reference implementation stores inscription content in a Taproot script-path spend, in data that is placed inside an unexecuted script envelope. A commit transaction creates the relevant Taproot output; a later reveal transaction spends it and exposes the inscription data on-chain.

The content is not stored in a separate cloud link that might disappear while the inscription remains. Its bytes are included in Bitcoin transaction data, so the blockchain carries the content itself. That persistence has a cost: the data consumes block space and is replicated by nodes that download and validate the chain.

The satoshi association is maintained by Ordinals conventions, not by a new Bitcoin consensus field. The Bitcoin network validates the transactions according to Bitcoin’s existing rules; Ordinals indexers interpret those transactions to assign and track inscription numbers and sat locations.

Why Taproot matters

Ordinals inscriptions became practical after Bitcoin’s SegWit and Taproot upgrades. SegWit introduced a separate witness-data structure and a weight-based block limit. Witness data counts toward block weight at a discount compared with equivalent data in the transaction’s non-witness portion. Taproot, activated in 2021, added script-path spending that can reveal a script when spent.

The inscription method uses that script path to carry data. It does not mean that every image is “hidden” in a way that makes it free or invisible. The reveal transaction is public, must fit within block-weight limits, and competes for space with payments and other transactions. The witness discount changes how that data is counted for block weight; it does not eliminate the storage burden.

Are Ordinals Bitcoin NFTs?

People often call inscriptions Bitcoin NFTs because they can represent unique, transferable digital artifacts. The analogy is useful, but the mechanics differ from many NFT systems. The inscription content is recorded on Bitcoin itself, and the associated sat can move through Bitcoin transactions. There is no ERC-721-style token contract required for the basic inscription model.

That does not guarantee that an item is scarce, valuable, authentic, or legally owned in every sense. A protocol can help identify and transfer an inscription, but it cannot establish copyright permission or make buyers agree on its cultural value. Marketplaces, wallets, and indexers also supply interfaces and interpretations that are not themselves Bitcoin consensus rules.

The distinction between the content and the sat matters too. A buyer may care about the image or text, while the protocol tracks a particular sat and inscription identifier. Wallets must handle that relationship carefully, especially when spending inputs that contain both ordinary funds and an inscription-bearing sat.

Ordinals and BRC-20 are not the same thing

BRC-20 is an experimental token convention built using text inscriptions and off-chain indexing rules. It uses JSON-like inscription content to describe operations such as deploying a token, minting units, or transferring them. Bitcoin does not execute a BRC-20 smart contract or maintain a native BRC-20 balance in its consensus state.

Instead, indexers interpret the inscription history and decide how those operations affect balances under the convention. That means two services can disagree if their indexing rules differ or if they handle edge cases differently. Users should not confuse a token’s ticker or marketplace listing with a protocol guarantee from Bitcoin itself.

Ordinals can exist without BRC-20: an image or text inscription does not automatically create a fungible token. Conversely, BRC-20 activity is one particular use of inscriptions, not the definition of the whole Ordinals system.

What happens when you transfer one

An inscription is transferred by spending the Bitcoin output that contains its sat and arranging the transaction so the sat flows to the intended output under Ordinals’ ordering convention. That requires an Ordinals-aware wallet or a service that understands the inscription’s location. A plain wallet may still create valid Bitcoin transactions, but it may not protect the specific sat from being sent as change or combined carelessly with other inputs.

Before sending or receiving an inscription, check that the wallet supports the relevant format and that the destination can display and manage it. Confirm the exact inscription identifier and receiving address. Keep separate ordinary BTC available for transaction fees so you are less likely to spend an inscription-bearing output accidentally.

No wallet can reverse a confirmed Bitcoin transaction. A mistaken transfer may be recoverable only if the recipient cooperates or the relevant service can intervene; often it cannot. Treat the sat itself as the asset being moved, not just the picture shown in an app.

The trade-off: block space and node costs

Bitcoin blocks have limited capacity, so inscription transactions compete with payments for inclusion. When demand for block space rises, users generally need to offer higher fees to get timely confirmation. Ordinals can therefore contribute to fee competition, though the size of that effect varies with overall demand and changes over time.

The impact is not captured by one simple “inscriptions equal storage” number. BitMEX Research’s September 2025 analysis estimated about 97.4 million inscriptions by then, including roughly 92.5 million BRC-20 entries and 2.7 million images. It found image and other inscription data occupied about 30.0 GB, while BRC-20-related data occupied about 27.8 GB. The report also argued that BRC-20 outputs can add pressure to the unspent transaction output (UTXO) set, a resource nodes must maintain, while large image payloads have a different validation profile.

Those are dated research estimates, not a live count or a universal verdict on the costs. The report’s broader point is that inscription formats can affect node resources in different ways. Bitcoin users disagree about whether such demand is a legitimate use of block space or an avoidable burden. The protocol does not settle that policy debate; miners choose transactions by fees, and node operators choose what software and policies to run within consensus rules.

Risks collectors should understand

An inscription’s permanence can be a benefit, but it also means mistakes are hard to fix. Content may be difficult or impossible to remove from the chain after confirmation. Uploading someone else’s work can raise copyright or privacy concerns, and buyers should not assume that an inscription’s presence proves the uploader had permission.

There are also market and technical risks. Prices can be thin or volatile; a marketplace’s floor price is not a guaranteed exit price. Indexers and wallet implementations may show different metadata, and BRC-20 balances depend on compatible indexing. Scams can imitate collections, tokens, or support accounts. Verify identifiers and software sources rather than relying on a ticker or thumbnail alone.

Finally, custody requires attention. A wallet recovery phrase controls the underlying Bitcoin keys, but not every wallet will interpret Ordinals the same way. Test small transfers only when appropriate, keep backups secure, and understand exactly which output a transaction will spend before approving it.

The practical takeaway

Ordinals add an identification-and-tracking convention for individual sats, while inscriptions put content into Bitcoin transaction data and associate it with a sat. BRC-20 builds an experimental token system on top of inscriptions, with balances interpreted by indexers rather than enforced by Bitcoin consensus.

If you are simply learning, remember the layers: Bitcoin validates transactions; Ordinals software numbers and tracks sats; inscriptions carry content; marketplaces and indexers provide additional services and interpretations. That separation makes the system easier to understand—and helps you judge what a wallet, token listing, or collectible actually guarantees.