technology · Solana · Payments · AI Agents · Scaling
Solana Payment Channels Pass 1M Payments per Second in Test
The benchmark compresses many signed payment authorizations into an onchain open-and-settle cycle; it is not a claim of one million base-layer transactions per second.
What the test measured
The Solana Foundation says a payment-channel proxy processed more than one million payments per second while testing 100,000 unique wallets.
The architecture does not settle each payment individually on Solana. A user first deposits a maximum amount into onchain escrow, then authorizes spending off-chain through signed messages. The final amount is settled onchain once, with unused funds returned to the primary wallet.
That distinction is essential. The benchmark measures authorized activity inside payment channels, not conventional base-layer transaction throughput.
Why payment channels exist
Per-request settlement becomes expensive and operationally awkward when software agents make hundreds or thousands of small purchases in a short period. Variable-cost services—such as token-based AI inference or metered computation—also make it difficult to know the exact charge before work is completed.
Payment channels move those repeated authorizations away from the settlement layer while preserving a spending ceiling and a final onchain record. The Solana Foundation compares the mechanism to opening a tab: collateral is committed first, usage accumulates and the balance settles later.
What the number does and does not prove
The Foundation estimates the test could represent more than 80 billion payment authorizations over 24 hours, at an extremely low processing cost per payment. Those figures come from the project’s benchmark environment, not a live measurement of open-network demand.
The test demonstrates that signature authorization and metering can scale far beyond one-transaction-per-payment designs. It does not establish production reliability, economic demand, security under adversarial conditions or equivalent throughput on Solana’s base layer.
The next evidence boundary
The specification, program and load-testing template are public, making independent reproduction possible. The stronger test will be whether third parties reproduce the results and whether live deployments maintain secure custody, correct settlement and predictable dispute handling.
Payment channels may become valuable infrastructure for machine-to-machine commerce. Their credibility will depend less on the headline throughput and more on the integrity of the open, meter and settle lifecycle.
Sources and disclosures
Primary source: Solana Foundation, “Payment Channels: 1 Million Payments Per Second,” published September 3, 2026. Benchmark results are attributed to the Foundation. The distinction between payment authorizations and base-layer TPS is Blue Rogues analysis.
Related analysis
- [Solana Votes to Cut Issuance Faster](/news/solana-votes-to-cut-issuance-faster)
- [OpenSea Adds Solana NFT Trading](/news/opensea-adds-solana-nft-trading)