The Case Against Payment Blockchains
Payments don't just need faster rails, they need programmable money.
Consider what happens every time you pull out a credit card. @Slashapp, a fintech company doing $4 billion in annual volume on Visa rails, recently detailed the process: an elaborate flow tracking authorization, incremental auth, partial capture, reversal, refund, aging, and force capture.
Four accounts move money between them across the lifecycle. Because reliability is table stakes, they also rely on a dedicated health-check service just to monitor an unfathomable number of authorization states.
That's what most payments today actually look like. Every hold, every conditional release, every dispute window exists because buyers and sellers don't trust each other unconditionally, and shouldn't have to.
All of this logic is bundled into the card network and completely abstracted away from the average consumer. Some argue this complexity is inherited from legacy intermediaries: chargebacks, settlement delays, correspondent banking. In reality most of it is structural. Authorization, holds, and conditional capture exist because commerce comes with conditions.
As crypto payments go mainstream with stablecoins leading the charge, this is the stack they compete against. Not just the speed of settlement, but the full programmatic layer: authorization, fraud, disputes, compliance, conditional capture. A high-stakes opportunity.
@Tempo raised $500M to build a payments-specific blockchain in partnership with @Stripe. @Circle launched @Arc. Stablecoin giant @Tether continues to expand distribution across every major network. With the GENIUS act, and CLARITY looming, regulators and the legacy banking system are also jumping onboard.
Stablecoins do need programmable infrastructure.
Modern blockchains were not optimized for payments. Yet every one of these projects assumes the answer is a new blockchain. New consensus mechanisms, new validator sets, new execution environments.
Bitcoin already exists as a neutral settlement network with the deepest liquidity in digital assets. It just wasn't expressive enough for commercial payments, until now.
Programmable money, not platforms
A dollar bill is a bearer instrument. Whoever holds it can spend it, no ledger update, no intermediary approval. The limitation is that a dollar bill can't encode conditions. It's bearer, but dumb.
Bitcoin's outputs (UTXOs) are unique in that they remove that limitation. Each one is a discrete unit of value with spending conditions attached: this value can be spent IF these specific conditions are met.
You order something online. The funds sit in a UTXO with two spending paths: the merchant can claim them with a co-signature from the escrow service, or the timelock expires and the funds return automatically. Authorization, hold, conditional capture, and automatic release, in a single output.
You give your AI assistant a budget for travel. A UTXO with spending conditions: up to $500 per day, expiring at end of trip, revocable at any time. The agent operates within those bounds, and the output enforces them. No risky wallet access, no smart contract gating permissions.
The constraints live in the money.
The account model, used by Ethereum, Solana, Tempo, and most modern chains, stores balances in a global ledger with smart contracts mediating state transitions. Every transaction on the network shares the same state. Your escrow competes for processing with someone else's token swap, someone else's NFT mint, every other operation on the chain.
As payment volume grows, this gets worse, not better. Solana already deals with transactions failing because they hit overlapping accounts. Tempo, a chain purpose-built for payments, has carved out dedicated lanes just to keep payment transactions from contending with everything else on the network. Circle just launched Nanopayments, an entirely separate infrastructure layer, because per-transaction gas costs on EVM chains make agent-scale micropayments unviable at the base layer. Three different problems, all rooted in the same design choice.
On Bitcoin, each output is self-contained. Your escrow doesn't know or care about any other transaction on the network. The conditions evaluate against the output itself, not against shared state that millions of other transactions are also touching. The money is the contract.
The Role of Trusted Third Parties
Despite its bearer quality, programmable money also needs to accommodate the conditional outcomes described previously. Fraud, dispute resolution and compliance checks don't disappear because the instrument is smarter. The question is whether those services are bundled into the network you depend on, or composed into the instrument on terms you define.
Take Stripe. When a merchant pays 2.9% per transaction, here's what they're getting: fraud scoring on every authorization, dispute resolution when a customer files a chargeback, PCI compliance so the merchant doesn't have to handle card data, onboarding that turns a signup form into a live payment flow.
Now Stripe also operates Tempo. Which means Stripe wants to simultaneously be the company you choose for fraud scoring and dispute resolution, and the company that runs the network your settlement depends on, governs the blockspace your transactions compete for, and controls the chain your assets live on. If you don't like Stripe's fraud scoring, you can switch to a competitor. If you don't like Stripe's chain governance, you migrate your entire asset base to a different network.
Bearer instruments on a neutral settlement network keep those concerns separate. You can encode Stripe as an arbiter on an escrow spending path, because Stripe is good at arbitration. Chainalysis for compliance. Whatever fraud scoring service fits. These are services composed into the instrument on terms the parties define. The settlement is on Bitcoin, where no single entity controls finality, blockspace, or governance.
Towards a Neutral Settlement Layer
The stablecoin infrastructure buildout is currently fragmented across a range of private and public networks which undermines the network effect of these new assets. New consensus, new validators, hundreds of millions in funding just to bootstrap liquidity. We can avoid this inefficiency by returning to a neutral settlement system.
Bitcoin processes over $10B in daily volume. Its finality is backed by proof-of-work, not a validator set that could change governance next quarter. No corporate entity controls its blockspace or decides which transactions get priority. It is the only network that doesn't have a CEO.
Every new chain being built for stablecoins asks you to make a platform bet: trust their sequencer, migrate your assets, accept their governance. Stablecoins don't need another blockchain. They need a way to inherit the properties of the one that already won.
At Ark Labs, we’re bringing stablecoins and digital assets to life on Bitcoin using Arkade. By extending the output model, we can achieve offchain execution and onchain settlement to power instant, programmable transactions. Whether it’s a conditional escrow or an automated agent, programmable money shines best on the world's most secure network.
Marco Argentieri (@tierotiero) is co-founder and CEO of Ark Labs, building open, programmable finance on Bitcoin.