How Fast Is a Crypto Payment, Really? Five Chains Timed From Our Own Nodes
Block intervals measured in production: Solana 376ms, Robinhood Chain ~0.1s, BNB Chain 0.45s, Base 2.0s, Ethereum 12.0s. What that means for the seconds your customer actually waits at checkout — and why detection architecture matters as much as chain speed.
The short answer: measured across the latest 120 blocks on each of our production nodes, average block intervals are — Solana 376ms, Robinhood Chain ~0.1s, BNB Chain 0.45s, Base 2.0s, Ethereum 12.0s. For a customer at your checkout, that translates to "paid" appearing in under a second on three chains, a couple of seconds on Base, and ten-to-twenty seconds on Ethereum — every one of which beats the card stack's two-day settlement at the layer that matters. The measured table, the honest math from block time to customer wait, and the part chain speed can't do alone: detection.
Measurement snapshot: 2026-08-26, 16:20 UTC — same run as the fee dataset; timestamps of the latest 120 blocks per chain, from our own full nodes.
The measured table
| Chain | Avg block interval (measured) | Customer sees "paid" in ≈ |
|---|---|---|
| Robinhood Chain | ~0.1s † | under a second |
| Solana | 376ms per slot | under a second |
| BNB Chain | 0.45s | ~1 second |
| Base | 2.0s | ~2–3 seconds |
| Ethereum | 12.0s | ~12–25 seconds |
† EVM block timestamps carry one-second precision; 120 blocks spanning 12 seconds means the chain produced blocks in a near-continuous cadence during the window — we report ~0.1s rather than implying sub-decisecond precision we don't have. Honest instruments show their error bars.
From block time to "how long does my customer stare at the spinner"
The translation is simple enough to do in public. A payment sent at a random moment waits on average half a block interval to be included in the next block, plus propagation (negligible when the watching node is yours). So:
- Solana: ~188ms average inclusion wait + sub-second detection = the checkout flips to "paid" while the customer's thumb is still near the send button. This is the tapping-a-card benchmark.
- BNB Chain at 0.45s deserves more credit than it gets: measured, it's firmly in the under-a-second club — relevant since it's where much of the USDT world lives.
- Base at 2.0s is "the page updated while I was reaching for my coffee" — indistinguishable from instant for any online purchase.
- Ethereum at 12.0s is the outlier, and context matters: one to two block intervals is 12–25 seconds — a long checkout pause, and also completely irrelevant on the four-figure invoices that are Ethereum's actual payment niche. Nobody wiring $20,000 minds twenty seconds.
The pattern worth naming: four of five chains now confirm faster than a chip-card handshake. The "crypto is slow" era ended the same way the "crypto is expensive" era did — measurably, while nobody updated their talking points.
What block speed cannot do alone
A 376ms slot time is worthless if the thing watching the chain polls a third-party API every two seconds through a rate limiter. Chain speed sets the floor; your gateway's detection architecture decides whether you ever touch it:
- Subscription to block production, not polling — our nodes see a transfer the moment it enters a block, which is what makes the sub-second numbers real at the checkout rather than theoretical on the chain.
- Detected is not settled — speed of seeing money must never be confused with finality of having it; the checkout can turn green sub-second while fulfillment correctly waits for the settled webhook.
- Settlement rides the same speed — because settlement executes inside the payment flow, the merchant's money arrives on the same clock the customer's confirmation does. There is no second, slower pipeline.
The comparison your finance team actually cares about
Card authorization takes ~2 seconds — competitive with Base, slower than Solana. Card settlement takes two business days, plus a months-long reversal window. The honest speed comparison isn't "crypto vs. card at the terminal" — cards are fine there — it's time-to-final-money: 376 milliseconds versus 48-plus hours, with no clawback clock ticking afterward. That's the number cash flow feels.
FAQ
Are these numbers stable, or a lucky snapshot?
Block cadence is among the most stable metrics a chain has — far steadier than fees. We re-measure with the fee dataset and update on the same living-page schedule.
Why does my wallet sometimes show "confirming" longer than your table?
Wallet UIs often wait for extra confirmations or poll on their own timers. Our numbers are chain cadence plus our detection; a third-party wallet's display cadence is its own choice.
Does congestion change the picture?
On the fast chains, congestion moves fees before it moves block times — and measured fees are currently at floor levels. Ethereum's interval is protocol-fixed at 12s regardless of load.
Which chain should I want customers to use?
For checkout feel: Solana first, everything on — the customer picks where their funds are, and even the slowest option beats every legacy rail where it counts.
How fast are Solana transactions compared to Ethereum?
Our measured cadence: 376ms blocks on Solana versus Ethereum's protocol-fixed 12.0s — roughly 32 block opportunities to Ethereum's one. At the checkout, that's "already paid" versus "watch the spinner."
Is Solana really sub-second for payments?
For detection, yes — measured on our own nodes: the transfer enters a block in well under a second and our detection flags it immediately. What wallets display afterward is their own polling choice.
How fast are Base and BNB Chain for payments?
Measured: ~2.0s blocks on Base, ~0.45s on BNB Chain — both firmly in the tap-a-card class. The fast-chain tier is a tier, not a single chain.
What about Bitcoin's speed by comparison?
Different design decade: ten-minute blocks, several for honest finality — which is why BTC-holding customers usually pay via the conversion route instead.
Does payment speed matter for online checkout, really?
Abandonment lives in the waiting — a page that flips to "paid" while the customer watches closes the loop before doubt opens it. It's the same psychology as one-click buying, delivered by block cadence.
How does crypto settlement speed compare to card settlement?
Two different clocks: cards authorize in seconds but settle to the merchant in days; fast chains authorize and settle in the same seconds. The comparison that matters is money-in-hand time.
Does the payment amount affect confirmation speed?
No — blocks don't weigh value: a $10 transfer and a $10,000 one ride the same cadence. Speed is a property of the road, not the cargo.
Why does Robinhood Chain show ~0.1s in your table?
That's the measured interval with the honest error bar noted in the methodology — at sub-second scales our sampling resolution matters, so we state it rather than round it away.
Can merchants rely on sub-second detection operationally?
That's precisely why we run our own nodes: detection speed you build workflows on can't depend on a public RPC's mood. The webhook fires on settlement, and settlement is seconds.
Will these speeds hold as chains get busier?
Block cadence is consensus-fixed on most of these chains — load moves fees before it moves timing, and fees have floor-room to spare. The living-page promise covers re-measurement either way.
Speed claims are the easiest in crypto to fake and the easiest to verify: create a sandbox payment, send it, and count — you won't get far. Then check the timestamp trail afterward: detection, settlement, webhook, all stamped, all yours to audit.