Dollar stablecoins on Lightning: how they move and whom they depend on
In brief
Dollar stablecoins handle far more payment volume than Bitcoin's Lightning Network, and several projects now aim to bring those dollars onto Lightning. This article explains how that works, what is actually running today, and what a sender or business must rely on to use it.
The two protocols built for this purpose, Taproot Assets and RGB, share a core idea: the token's records are kept by its holders, and Bitcoin stores only a cryptographic fingerprint of them. This design has a practical cost that is easy to overlook. A wallet's seed phrase is not enough to recover the tokens; the holder must also back up the token data itself.
The two protocols differ in how a payment crosses Lightning. Taproot Assets converts the stablecoin to bitcoin at the edge of the network, so it can pay any Lightning user, but it depends on a node willing to perform that conversion. RGB carries the stablecoin along the whole route, so no conversion is needed, but every node on the route must run RGB software, and RGB's developers have divided into two groups over which version is correct. USDT already runs on Taproot Assets; a Tether-backed company plans to launch it on RGB in late November 2026, with Lightning support to follow in 2027.
The main conclusion is that none of these designs removes the need to trust someone; each places that trust in a different party. Taproot Assets relies on conversion nodes and on a single company's software, and RGB relies on every node along the route. Above all of them sits the stablecoin issuer, which can freeze balances and decides which tokens it will redeem. Finally, none of this activity appears in public Lightning statistics, because the channels involved are hidden by design. As a result, no one can currently say how many dollars move over Lightning. Section 6 compares the options in a single table.
Key terms
Brief definitions for readers without a Bitcoin or Lightning background. Readers familiar with the protocol can skip this section.
- Lightning Network: a payment network built on top of Bitcoin. It settles payments instantly and at very low cost by routing them through a web of two-party payment channels, instead of recording every payment on the Bitcoin blockchain. (JBI translation of River Research)
- Channel: a balance shared by two parties and funded by a single on-chain Bitcoin transaction. Payments inside the channel change how the balance is divided until the channel closes. (Basis of Lightning Technology (BOLT) 2: channel management)
- Commitment transaction: the current, unbroadcast Bitcoin transaction that records who owns what inside a channel. Each payment replaces it with a new one. (BOLT 3)
- Hashed Timelock Contract (HTLC): the mechanism that lets one payment move safely across several channels. Each intermediate node is paid only if it forwards the payment; otherwise the funds return to the sender. (BOLT 3)
- Gossip protocol: the broadcast system Lightning nodes use to tell each other which channels exist, so that payments can be routed through nodes the sender does not know. (BOLT 7)
- Unannounced (private) channel: a channel deliberately left out of gossip. It works normally for its two owners, but no third party or public explorer can see that it exists.
- Client-side validation: a design in which a token's records stay with the people transacting. Bitcoin carries only a small cryptographic commitment to those records, and each party checks the history relevant to them rather than the whole network checking everything. (RGB and client-side validation)
- Universe server: in Taproot Assets, a public repository that stores and serves the proof records token holders need. It has no special authority; anyone can check its data against Bitcoin. (Universes)
- Edge node and Request for Quote (RFQ): an edge node holds both bitcoin and a token and exchanges one for the other in the middle of a payment. The RFQ is the step in which the edge node quotes an exchange rate, valid until a stated expiry, before the payment is sent. (Edge nodes, RFQ)
1. Why put dollars on Lightning
A stablecoin is a token whose issuer promises to redeem it for a fixed amount of a currency, usually one US dollar. Stablecoins have become the main way value moves on blockchains. Forbes estimated stablecoin transaction volume at about 33 trillion US dollars over the year to March 2026. Tether's USDT alone had a circulating supply of about 184 billion dollars, roughly 58 percent of a stablecoin market worth around 316 billion dollars. Almost all of that supply is held on other blockchains, chiefly Tron and Ethereum.
Lightning is small by comparison. The bitcoin committed to its public channels peaked at 5,851 bitcoin (BTC) on 28 December 2025 and stood at 3,754 BTC on 30 August 2026, the most recent figure mempool.space has published. Even at a price of 100,000 dollars per bitcoin, 3,754 BTC would be worth about 375 million dollars, roughly 0.2 percent of USDT's supply.
This gap explains the interest. Lightning offers fast, low-cost settlement, and stablecoins offer the price stability that most users and merchants want. Combining the two is attractive, but Lightning was designed to carry only one asset, bitcoin. This article therefore asks three questions:
- How can a network built for bitcoin carry a dollar token?
- What is actually running today, and who issues the dollars?
- Whom does a sender or business have to trust for a payment to complete?
Taproot Assets and RGB are the two protocols built specifically for this purpose, and most of the article concerns them. They are not the only way dollars reach Bitcoin, however. Liquid, a sidechain operated by a federation of companies, has carried a Tether dollar for several years, and Spark, a newer project, issues its own dollar and converts it at the boundary with Lightning. Both are included in the comparisons below.
2. How Bitcoin can carry a second asset
Bitcoin has no built-in way to issue a second asset. One option would be to add a token ledger that every Bitcoin node stores and checks, but that would place on the whole network the burden of data that most of its users do not need. Client-side validation avoids this. The token's records stay with the people who hold it, and Bitcoin stores only a small cryptographic commitment, which works as a fingerprint of those records.
When a sender pays in such a token, the sender also gives the recipient the token's history, from its issuance up to the current payment. The recipient's wallet checks that history against commitments already confirmed on the Bitcoin blockchain. If everything matches, the tokens are genuine. Bitcoin nodes themselves never see the token.
Bitcoin Research JP explained this mechanism in 2020, when RGB was at an early stage, and the mechanism has not changed since. A 2024 survey of programming on Bitcoin treats client-side validation as a family of designs rather than a single protocol (Programming on Bitcoin).
Two ways to anchor the record
Taproot Assets and RGB share this idea but anchor the fingerprint differently.
RGB uses what it calls a single-use seal. A token's state is tied to a specific Bitcoin output, and spending that output authorizes exactly one change of state. Because Bitcoin already prevents the same output from being spent twice, it also prevents the token from being spent twice (What is RGB).
Taproot Assets places its commitment inside the tree structure that Bitcoin's 2021 Taproot upgrade made available within a transaction output. The protocol was first announced under the name Taro and later renamed, so older coverage uses the earlier name.
A seed phrase is not enough to restore your tokens
The most important practical consequence of client-side validation concerns backups. In an ordinary Bitcoin wallet, the seed phrase (a list of 12 or 24 words) is enough to recover every coin, because the coins are recorded on the public blockchain. Client-side-validated tokens are different: the blockchain holds only the fingerprint, so the holder must keep the records.
Both projects state this clearly. Lightning Labs' safety notes for Taproot Assets say the seed phrase is "NOT enough to restore assets minted or received," and that deleting the software's database without a backup destroys the funds (safety.md). Tether's wallet development kit gives the same warning for its RGB module: users should not assume that the seed phrase restores token balances, and must keep and test backups of the data directory (RGB wallet guide).
Universe servers help by letting holders retrieve proof records from a public source, but they do not remove the need for a backup. For any business holding these tokens, backing up the token data is as important as protecting the keys.
3. How a stablecoin payment crosses Lightning
Both protocols can use Lightning, but they take opposite approaches. A third system, Spark, keeps its dollar off Lightning entirely and converts it at the boundary.
Taproot Assets converts at the edge of the network
In Taproot Assets, a token is held in a Lightning channel alongside bitcoin, in the same funding output, so opening such a channel costs no more on-chain than opening an ordinary one. These token channels connect to edge nodes, which hold both the token and bitcoin and also maintain ordinary channels with the rest of the network.
When a user pays, the edge node exchanges the token for bitcoin in the middle of the payment. The token therefore travels only on the first hop (from the sender to an edge node) or the last hop (from an edge node to the recipient). Every hop in between is a normal bitcoin payment. As a result, a wallet holding dollar tokens can pay an ordinary Lightning invoice, and the recipient does not need to know that a token was involved. A merchant can also do the reverse: issue a bitcoin invoice and receive dollar tokens after an edge node converts the payment (Edge nodes).
The cost of this reach is a conversion spread, together with dependence on an edge node that holds enough of both assets.
The price quote protects the converting node
Before the payment is constructed, the sender requests a quote, and the edge node replies with an exchange rate and an expiry time (RFQ). A node that converts a token into bitcoin carries the price risk until the payment settles. Without an agreed rate and deadline, a sender could start a payment, wait, and complete it only if the exchange rate moved in the sender's favor. The sender would hold a free option at the node's expense. Bitcoin Research JP described this weakness in April 2022, two days after the protocol was announced. A quote with a short expiry limits this risk but does not remove it: the price can still move before the quote expires, and the edge node bears that movement.
RGB carries the asset the whole way
RGB takes the opposite approach. It records the token's change of state inside each channel's commitment transaction, so the token itself moves along the entire route. Routed payments use the same Hashed Timelock Contract mechanism as bitcoin payments: each hop adds an output with a token amount attached, and whoever claims that output receives both the bitcoin and the tokens (RGB Lightning roadmap). Each payment must also carry a small amount of bitcoin, because Bitcoin outputs below a minimum size, known as the dust limit, cannot be spent. The reference node software is built on the rgb-lib library and a modified version of the Lightning Development Kit (LDK).
This design has no conversion spread and no edge node. The trade-off is that every node along the route must run RGB software. A merchant that installs an RGB wallet cannot yet receive RGB payments on that basis alone; every node on the path between sender and recipient must support RGB first.
Spark converts before the payment reaches Lightning
Spark is a separate Bitcoin system with its own dollar token, USDB, issued by Brale. USDB exists only inside Spark and never travels over a Lightning channel. When a Spark user pays a Lightning invoice, a service provider exchanges the USDB for bitcoin and sends an ordinary bitcoin payment; incoming payments are converted the other way (Spark's landscape note). How Spark itself works is outside the scope of this article; Bitcoin Research JP explained it in June 2026.
The comparison is still useful. Two separate groups, issuing different dollars, both chose not to require the existing Lightning network to understand a new asset. RGB is the only design that does require this, which is why its adoption depends on the nodes along the route and not only on senders and recipients.
4. What is running today
The table below summarizes each option as of early November 2026. The status column is based on each project's own statements unless noted otherwise.
| Option | Dollar token (issuer) | Status (November 2026) |
|---|---|---|
| Taproot Assets | USDT (Tether) | USDT integration completed March 2026; software still labeled alpha |
| RGB 0.11.1 (RGB Protocol Association) | Bridged USDT (Utexo, backed by USDT locked on other networks) | Protocol on the main Bitcoin network since July 2025, according to the Association. USDT in testing; public launch planned for 28 November 2026 and Lightning for early 2027, according to Utexo |
| RGB 0.12 (RGB-WG) | None confirmed | Declared ready for production by its developers; not on the main network, according to the Association |
| Spark | USDB (Brale) | Live since January 2026, according to Spark |
| Liquid | L-USDT (Tether) | In circulation for several years |
Taproot Assets has one implementation and is still labeled alpha
Taproot Assets is a single product from a single company. Its software, tapd, is distributed together with the company's Lightning node software, Lightning Network Daemon (LND). The current version, 0.8, was released in June 2026 (v0.8), with maintenance releases through October.
The specification exists as seven draft Bitcoin Improvement Proposals (BIPs), kept on a Lightning Labs engineer's personal branch rather than merged into the official BIP repository. Because there is only one implementation, it has not yet been tested whether independent software can work with it. Lightning Labs' own install guide still describes the software as alpha and states that use on the main network is at the user's own risk (get tapd).
RGB is divided between two incompatible versions
RGB cannot be assessed as a single protocol, because its developers have divided into two groups with incompatible software.
The RGB Protocol Association is a Swiss non-profit formed in July 2025 by Bitfinex, Fulgur Ventures, Plan B Network, Kaleidoswap, Bitcoin Tribe, and LNFI. It states that version 0.11.1 is the version running on the main network, and describes version 0.12 as an unfinished rewrite whose Lightning tests are disabled (v0.11.1 vs v0.12). Its libraries dropped their release-candidate label on 16 September 2026 (core libraries).
The other group, RGB-WG, is associated with the LNP/BP Standards Association. It published version 0.12, told issuers they could use it on the main network, and expected products in summer 2025 (v0.12 consensus). It describes version 0.11.1 as an unauthorized fork whose rule changes affect security (on the fork).
Some of these claims can be checked from outside and some cannot. Which version is running on the main network, and which is receiving updates, can be observed, and on both points the evidence currently favors the Association's version. Whether either version is secure cannot be determined without an independent audit. In practice, a business asking whether a wallet supports RGB must first ask which of the two versions it uses.
Tether backs both protocols
Tether announced USDT on Taproot Assets in January 2025 (Tether). In August 2025 it also announced a plan to issue USDT on RGB "in the near future" (announcement via PR Newswire). The press release was distributed by the RGB Protocol Association and names the Association's version, 0.11.1, as the one ready for the main network.
The RGB path runs through Utexo, a company whose funding round Tether led. In October 2026 Utexo opened a test version of USDT on RGB to exchanges and wallets on Bitcoin's test network. The token is a bridged version of USDT. A user locks Tether's USDT on Arbitrum, an Ethereum-based network; a single signing server, running in an isolated Amazon Web Services enclave, then creates the same amount on RGB. To redeem, the user destroys the RGB tokens and a group of separate signing servers must agree before the locked USDT is released (Utexo Mint). Utexo's node software builds on the RGB Protocol Association's line of code. Utexo plans a public launch on 28 November 2026 and Lightning support in the first quarter of 2027 (Utexo). These are company targets; earlier targets of July and October 2026 were not met. Tether's wallet development kit also ships RGB modules maintained by Utexo, with the Lightning module still marked beta (RGB Lightning module).
The issuer has therefore not chosen between the two protocols. It has put USDT on one and is backing a second, bridged version on the other.
5. Why public data cannot see this activity
Asset counts measure issuance, not use
The figures that are commonly cited do not measure how much these tokens are used. Lightning Labs' default universe server lists 187,493 asset mints, but most are single-issue collectibles and .btc domain names, and the list includes only assets synchronized to that server. The RGBMap registry lists two assets on the main network, but it counts only assets whose issuers chose to register. Neither figure indicates payment volume.
No published figure exists for USDT on Lightning
No source has published the supply or payment volume of USDT on Lightning. For comparison, Forbes reported 83.9 billion dollars of USDT on Tron alone in March 2026 (Forbes). The amount on Lightning is very likely small by comparison, and it cannot have been moving for long. Tether announced the integration on 30 January 2025, and Forbes dates its completion to 21 March 2026, fourteen months later. Descriptions of "USDT on Lightning since early 2025" therefore date it from the announcement rather than the completion.
What wallets actually offer
Which wallets offer a dollar balance, and how they handle it, can be observed. The table below lists widely used Lightning wallets that offer one, based on each company's own documentation as of early November 2026.
| Wallet | Dollar balance | What moves over Lightning |
|---|---|---|
| Speed Wallet | USDT-L, Speed's own wrapped USDT on Taproot Assets, backed by USDT on Ethereum that Speed holds (Speed) | The token itself, over Taproot Assets channels |
| Wallet of Satoshi | USDB on Spark, issued by Brale (disclosure) | Bitcoin; a swap provider, Flashnet, converts between USDB and bitcoin |
| Strike | A USDT balance held by Strike for the user (Strike FAQ) | Bitcoin; the dollar balance stays in Strike's accounts |
| Blink | A synthetic dollar balance, not a token: bitcoin hedged with derivatives on an exchange (Blink) | Bitcoin |
| AQUA | USDT on Liquid (AQUA v0.2.0) | Bitcoin, through a swap service that converts to and from Liquid |
| Tether Wallet | USDT on other blockchains | Bitcoin only, as of version 1.4 in June 2026 (Crypto Economy) |
Only one of these wallets moves a dollar token over Lightning, and that token is Speed's own wrapped USDT, not the USDT that Tether issues on Taproot Assets. In every other case the dollar stays in a company's accounts, on another network, or in a hedge, and Lightning carries bitcoin. This is a further reason Lightning data cannot measure dollar use: in most wallets, no dollar ever enters a channel.
Token channels are hidden from public statistics by design
Better dashboards cannot close this gap. Taproot Assets channels are unannounced, so they never appear in the gossip data that public explorers collect. RGB's token records are kept inside commitment transactions, which are not broadcast while a channel is open. Every public Lightning capacity chart therefore counts announced bitcoin channels only.
The previous article in this series argued that public capacity is not a reliable measure of Lightning's liquidity. Token channels add a further problem: if users moved funds from announced bitcoin channels into unannounced token channels, public statistics would show a decline even if total activity grew. Two other changes in 2026 have the same effect. Splicing, now enabled by default in the Core Lightning software, resizes an existing channel instead of opening a new one. LND's newer simple taproot channels cannot be announced at all until the taproot gossip specification is finalized.
Public capacity fell about 36 percent between the December 2025 peak and 30 August 2026, a period that includes the reported completion of the USDT integration. Whether these two facts are connected is a hypothesis that current public data cannot test.
6. What each option asks you to trust
The table below brings the earlier sections together. For each option, it shows whom a payment depends on, what happens if that party stops cooperating, and what the recipient needs.
| Option | A payment depends on | If that party stops cooperating | The recipient needs |
|---|---|---|---|
| Taproot Assets | An edge node holding both dollars and bitcoin, and Lightning Labs' software (the only implementation) | The holder can no longer pay ordinary Lightning invoices, and must restore proof records from a backup or another universe server | Any Lightning wallet, if an edge node converts the payment on the last hop |
| RGB | Every node on the route, all running the same RGB version; for USDT, also Utexo's bridge | The payment cannot be routed, and there is no conversion to fall back on | A wallet running the same RGB version as the sender |
| Spark | Spark's operators and the providers that convert USDB (according to Spark) | Depends on Spark's own exit design, which this article did not assess independently | A Spark wallet, or any Lightning wallet through edge conversion (according to Spark) |
| Liquid | The federation that holds the backing bitcoin and signs blocks | The sidechain stops processing transactions | A Liquid wallet |
None of these designs removes trusted parties; each places them somewhere different. Taproot Assets reaches all of Lightning through edge nodes and one company's software. RGB removes the edge node but needs every node on the route to run the same version, and its developers disagree about which version that is. Spark and Liquid rely more heavily on a named operator or federation.
The issuer sits above every option
One dependency applies to every option and is independent of the protocols. A stablecoin is a claim on its issuer. If a protocol splits into two versions, the issuer decides which version's tokens it will continue to redeem; a bitcoin holder would keep coins on both sides, but a stablecoin holder has only what the issuer recognizes. The issuer can also freeze a balance at any time. A bridged token adds one more party: the bridge that holds the locked collateral. Utexo's design also lets the asset name a list of rejected coins; tokens descended from a listed coin can still move between RGB wallets, but the bridge refuses to redeem them. The control moves from the wallet to the exit. Shoma Moriya made this argument for Bitcoin Research JP in June 2026: stablecoins do not give Japanese exporters payment sovereignty, because Tether and Circle can freeze funds. None of the protocols described here changes that. They determine how the dollar moves, not who controls it.
For a business considering stablecoin payments over Lightning, the useful question is therefore not which protocol is best, but whose cooperation each payment needs, what users lose if that party stops cooperating, and whether the business can keep the token records that the blockchain will not keep for it. The answers currently differ for each option, and the public data that would show which option is gaining real use does not yet exist.
Sources: Forbes, 31 March 2026 (stablecoin volumes, USDT scale, USDT on Tron, 21 March 2026 completion date); Tether, 30 January 2025 (the Taproot Assets announcement); Tether RGB plan, 28 August 2025 (USDT on RGB, distributed by the RGB Protocol Association); Utexo, 22 October 2026 (bridged USDT testing, 28 November launch, Lightning in early 2027); Utexo Mint documentation (lock-and-mint design, signer enclaves, reject list); Lightning Labs, Taproot Assets v0.8 (current release); Taproot Assets install guide (alpha warning); Edge nodes and RFQ (conversion at the edge); Universes (proof storage); taproot-assets safety.md (seed phrase is not sufficient); RGB and client-side validation (2020 Bitcoin Research Institute explainer); How TARO differs from Counterparty and Omni (the free-option weakness, April 2022); Programming on Bitcoin (client-side validation as a family of designs); What is RGB (single-use seals); RGB Lightning roadmap (commitment transactions, dust limit); RGB v0.11.1 vs v0.12 and core libraries update (the Association's account); RGB v0.12 consensus and on the fork, 18 July 2025 (RGB-WG's account); Tether WDK RGB wallet and RGB Lightning module (Utexo-maintained modules, beta Lightning wallet, seed does not restore token data); Spark, stablecoins on Bitcoin (Spark's account of USDB and edge conversion); Oshikawa, Spark protocol guide (how Spark works, June 2026); Speed, USDT on Lightning (USDT-L design and backing); Wallet of Satoshi disclosure v1.23 (USDB, Brale, Flashnet); Strike FAQ (USDT balance); Blink Stablesats (synthetic dollars); AQUA v0.2.0 (USDT on Liquid); Crypto Economy, Tether Wallet 1.4 (Lightning for bitcoin only); Lightning Terminal universe (mint count); RGBMap (registry count); mempool.space Lightning statistics (capacity peak 28 December 2025, 3,754 BTC on 30 August 2026, the latest published point as of 5 November 2026); taproot gossip (why taproot channels stay unannounced); Lightning capacity is not liquidity (previous article); Moriya, payment sovereignty (issuer freeze argument, June 2026).