Stablecoin Chains Compared: Arc vs Tempo vs Plasma vs Stable

Four chains, one job

For most of a decade the answer to "where should dollars live onchain" was a general-purpose blockchain that happened to host a stablecoin contract. Four networks now exist to do only the payments part, and as of 16 September 2026 all four are live.

The tables comparing them rank throughput and finality. Those are the numbers vendors publish about themselves, measured on their own terms, and they are the least useful column in the set. The question that actually separates these chains is simpler: what pays for gas, and who is allowed to validate?

How we compared

Every figure below was read on 29 September 2026 from each network's own documentation, pricing pages or launch announcement, linked once from each chain's heading. Where a detail appears only in third-party coverage we label it as such rather than presenting it as documented. We have not run transactions on these chains and we publish no independent latency or throughput measurements.

The comparison

Arc: the institutional one

Circle's announcement of the founding validator cohort names BlackRock, the Depository Trust and Clearing Corporation, Galaxy, Global Payments, ICE, Mastercard, MoneyGram, SBI Group, Standard Chartered, Sumitomo Corporation, Visa and Circle itself. The same document states that Arc "is an open L1 blockchain launched by Arc Network Services LLC and operated by a permissioned validator set", and that transacting "depends on the ability to obtain and use USDC to pay gas fees".

Read those two sentences together and the product is clear. This is a settlement network whose validators are the institutions that already clear and custody money, with a dollar as the fee unit. BlackRock is deploying BUIDL onto it, and the DTCC's tokenisation integration is described as planned for the second half of 2027.

For a bank or asset manager, a named and vetted validator list is the feature. For anyone whose threat model includes "the operators collectively decide not to process my transaction", it is the whole problem. Neither reaction is wrong; they are different requirements.

Tempo: no native token at all

Tempo, incubated by Stripe and Paradigm, went live on 18 March 2026. Its documentation is unusually direct about the fee model: users "pay transaction fees in any USD stablecoin on Tempo", with "no native token required", and fees are "paid directly in TIP-20 stablecoins with automatic conversion". Consensus is "Simplex BFT via Commonware for deterministic sub-second finality", and the site states blocks "finalize in ~0.6 seconds with no re-orgs".

Removing the native token removes the most common failure of onchain payments: the user who holds dollars, wants to send dollars, and cannot, because they do not hold the unrelated asset required to pay the fee. Tempo also shipped a protocol for machine payments alongside its mainnet, which is the thread our explainer on how AI agents pay picks up.

Tempo's documentation does not state whether its validator set is permissioned, which is a gap worth asking about directly if you are evaluating it.

Plasma: the subsidised transfer

Plasma launched its mainnet beta on 25 September 2025 out of the Tether and Bitfinex orbit, and it is the only chain here that kept a conventional native token while making the user experience look tokenless.

Its documentation describes a protocol paymaster, funded by the Plasma Foundation, that sponsors gas for eligible peer-to-peer USDT transfers so a sender never needs XPL. Two conditions travel with that: the paymaster covers eligible transfers rather than arbitrary contract interactions, and a lightweight identity check exists to stop the subsidy being farmed. Plasma also supports whitelisted assets such as USDT and its wrapped Bitcoin as gas tokens, converting at execution time using oracle prices. Consensus is PlasmaBFT, documented as "a pipelined implementation of Fast HotStuff".

A subsidy is a business decision rather than a property of the technology. It works until the foundation decides it does not, which is worth writing into any model that assumes zero transfer cost indefinitely.

Stable: gas that is already a dollar

Stable went live on 8 December 2025 and takes the shortest route to the same destination: the gas asset is USDT0, so the unit you spend on fees is the unit you are sending. Its documentation describes "a payments Layer 1 where USDT0 is the native gas token" with "sub-second finality" and full EVM compatibility, and the launch introduced a STABLE token and a foundation alongside the network.

The consensus mechanism is described in third-party comparisons as a CometBFT-derived design. We could not confirm that from Stable's own documentation in this pass, so we are not stating it as documented.

Why we have not ranked these on speed

Arc advertises sub-second finality. Tempo says about 0.6 seconds with no re-orgs. Stable says sub-second. Plasma's consensus is built for rapid successive finalisation.

Every one of those is a vendor statement about its own network, and none of them is measured against a shared workload, the same transaction mix, or the same geographic spread of nodes. Cold numbers like these are extremely sensitive to all three. A chain finalising an empty block in a data centre and a chain finalising a full block across continents can both honestly report sub-second finality and mean entirely different things.

If settlement latency is a hard requirement for your product, measure it yourself on your own transaction pattern. Until someone publishes a harness that runs the same test across all four, a ranking would be decoration.

Which should you choose?

If you are a regulated institution that needs to name your counterparties: Arc. A validator list containing the DTCC, Visa, Mastercard and Standard Chartered is a compliance artefact as much as an infrastructure choice, and USDC as gas keeps the accounting in one currency.

If you want the simplest possible user experience: Tempo, where fees are paid in any USD stablecoin and no native token exists to explain to anyone.

If your volume is USDT peer-to-peer transfers: Plasma, where those transfers are sponsored today. Model what happens if they stop being sponsored, because that is a foundation policy rather than a protocol guarantee.

If you are already building on USDT rails: Stable, where gas is USDT0 and there is no second asset to manage.

If you need composability with existing DeFi: none of these, yet. The liquidity, tooling and contract ecosystems still sit on general-purpose chains, and our comparisons of cross-chain bridges and stablecoin payment processors cover how value moves between the two worlds today.

Conclusion

The stablecoin chain category has resolved faster than most people expected, and it has resolved along an axis nobody put in a comparison table. Three of these networks are competing on user experience: remove the native token, subsidise the transfer, make the fee unit a dollar. Arc is competing on something else entirely, which is the identity of the institutions running the machines.

If you are choosing between them, start with the governance question and work backwards to the technical ones. And if you want to understand what is being settled on top of these rails, our explainer on tokenised deposits and stablecoins covers the bank-issued side of the same market.

This comparison is an editorial synthesis of network documentation and launch announcements read on 29 September 2026 and linked inline. We did not transact on these chains, run latency tests, or receive briefings from these teams. Where a detail could be confirmed only through third-party coverage we have labelled it. Finality and throughput figures are vendor claims; verify against your own requirements before building on any of them.