Stablecoin settlement at million-transaction-per-day scale is no longer theoretical. Circle's CCTP V2 has moved more than $30B in cumulative cross-chain volume since launch, Bridge.xyz processes thousands of B2B payouts per minute for Stripe, and Solana clears 1,000+ effective stablecoin transactions per second during peak hours. Throughput is the new battleground, and the answer depends on three things at once: which chain you settle on, how your provider batches, and how it recovers from errors.
This benchmark walks through the real numbers for six leading providers (Circle CCTP, Bridge.xyz, BVNK, Coinbase Institutional, Anchorage Digital, Fireblocks), the per-chain TPS ceilings they inherit, and the batching and retry behavior that determines whether a 1M-transaction day actually clears.
What does "stablecoin throughput" actually measure?
Stablecoin throughput is the sustained rate of successful settlements (in transactions per second, or TPS) a provider can deliver end-to-end, including signing, broadcast, inclusion, and finality. It differs from raw chain TPS because providers add batching, queueing, and compliance checks. A useful figure combines effective TPS with a 99th-percentile latency target.
Three layers stack to produce the observable number:
Chain-level TPS ceiling (the physics).
Provider batching and queue logic (the software).
Compliance, policy, and signing latency (the workflow).
What are the chain-level TPS benchmarks for stablecoin settlement?
Chain throughput sets the upper bound. Ethereum L1 caps out near 15 TPS for ERC-20 transfers, while Solana sustains 1,000+ effective non-vote TPS, and Base routinely clears 2,000+ TPS during peak retail flows. L2 rollups dominate the top end because they batch calldata back to Ethereum.
Chain | Effective stablecoin TPS | Block time | Finality | Median fee (USDC transfer) |
Ethereum L1 | ~15 | 12s | ~13 min (2 epochs) | $0.30 to $4.00 |
Arbitrum One | ~4,000 theoretical, 40 to 80 sustained | ~250ms | ~7 day challenge, 1 min soft | $0.02 to $0.10 |
Base | ~2,000+ peak | 2s | ~12 min L1 settle, 2s soft | $0.01 to $0.05 |
Solana | 1,000 to 3,000 effective non-vote | 400ms | ~13s | $0.00025 average |
Polygon PoS | ~700 theoretical, 35 to 50 sustained | ~2s | ~256 blocks (~10 min) | $0.001 to $0.01 |
Tron | ~2,000 theoretical, ~70 to 200 sustained | 3s | ~57s (19 blocks) | $0 to $1 (energy model) |
"Effective" matters. Solana's 65,000 marketing TPS includes consensus votes; real non-vote TPS during heavy DeFi or memecoin days lands between 1,000 and 3,000, per public Solana validator dashboards. Arbitrum's 4,000 figure is a theoretical sequencer ceiling; sustained stablecoin TPS over a full hour rarely tops 80.
How do the major settlement providers compare on throughput?
Provider throughput is rarely published in TPS. It is implied by daily transaction counts, batch sizes, and webhook delivery latencies in their docs and audited disclosures. The table below normalizes those into a single comparison.
Provider | Settlement model | Reported daily volume / capacity | Effective TPS (estimate) | Stated error rate |
Circle CCTP V2 | Native burn and mint, instant cross-chain | $30B+ cumulative cross-chain, ~$200M+ daily peak | ~30 to 60 sustained per supported chain | <0.1% per Circle status page |
Bridge.xyz (acquired by Stripe, Oct 2024) | Real-time onchain payouts, off-ramps | Powers Stripe stablecoin payments across 70+ countries | ~50 to 100 sustained at peak | ~0.5% (Stripe stablecoin payments docs) |
BVNK | Multi-rail, batched fiat-to-stablecoin and back | $15B+ annualized payment volume (2024 disclosure) | ~20 to 40 sustained | Not publicly disclosed |
Coinbase Institutional / Prime | Custodial settlement, internal ledger then onchain | $185B+ institutional quarterly trading volume (Q1 2024) | Internal off-chain effectively unbounded, onchain ~25 TPS | <0.01% internal settle |
Anchorage Digital | Qualified custody with policy-engine signing | $50B+ AUC, batch settle windows | ~5 to 15 onchain (policy review adds latency) | 0% reported lost transactions since 2017 |
Fireblocks | MPC wallet infra plus Network co-signing | $6T+ cumulative across 1,800+ institutions | ~100+ aggregate across customers | <0.05% per Fireblocks SOC 2 reporting |
Numbers above are normalized from each provider's public docs, status pages, and audited disclosures as of Q1 2026. Internal ledger settlements (Coinbase, Anchorage's omnibus model) are effectively unbounded because they never hit chain. Onchain TPS is what matters when a counterparty demands proof of settlement.
How do batching strategies change the math?
Batching is the single biggest lever. The three dominant patterns are real-time per-transaction, micro-batched (sub-second windows), and daily-window net settlement. Each has different throughput and error semantics.
Stripe Charge → daily settle: Stripe accepts thousands of card or stablecoin charges per second but settles to merchants on a T+1 or T+2 batch. Effective customer-facing TPS: 4,000+. Onchain settlement count: a few large transfers per day per merchant.
Bridge.xyz → real-time payout: Each B2B payout is a discrete onchain transfer. Throughput is bounded by chain TPS and signing latency, ~50 to 100 sustained across Stripe's Bridge backbone.
Circle CCTP V2 → instant burn-mint: Each cross-chain transfer is atomic: burn on source, attestation by Circle's off-chain verifier, mint on destination. Median latency dropped from ~13 minutes (V1) to under 30 seconds for fast-transfer chains in V2. Throughput per supported chain pair sits at 30 to 60 sustained TPS.
Anchorage policy-engine batching: Each withdrawal passes through quorum-based policy review, so onchain settle is batched into 1 to 4 windows per day per client. TPS is low but error rate approaches zero.
The trade is universal. More batching equals higher headline throughput and lower fees, but slower individual confirmation and bigger blast radius if a single batch fails.
What error rates and retry patterns should you plan for?
At 1M transactions per day, even a 0.1% failure rate produces 1,000 errors. Provider retry logic determines whether those become support tickets or self-heal. Three failure modes dominate: nonce collisions, attestation timeouts, and destination-chain reorgs.
Nonce collisions: Most common on Ethereum L1 when a provider signs ahead with multiple keys. Fireblocks and BVNK handle this with sequenced signing queues; retry success rate is reported above 99.9%.
Attestation timeouts: CCTP V2 requires a Circle attestation before destination mint. If the attestation service stalls, the transaction is recoverable but stuck. Circle's status page logs 4 incidents in 2025 with median recovery under 18 minutes.
Reorgs: Mostly a Polygon PoS and Tron concern. Providers typically wait 64 confirmations on Polygon, 19 on Tron. A reorg deeper than that is treated as a hard failure and requires manual reconciliation.
Retry behavior also varies. Bridge.xyz auto-retries with exponential backoff up to 7 attempts. Fireblocks routes retries through alternate signers via its Network. Coinbase Institutional uses internal ledger reservations so client-facing balances never reflect a failed onchain leg.
How do you actually hit 1M+ transactions per day?
One million daily transactions averages ~11.6 TPS sustained over 24 hours, but real traffic is bursty. Peak hours often run 3 to 5 times the average, so you need a system that can absorb 35 to 60 TPS for 4-hour windows without queue backup. Three patterns achieve this in production:
Multi-chain fan-out: Route 60% to Solana or Base, 30% to Arbitrum, 10% to Ethereum L1 for high-value or compliance-mandated transfers. This is how Stripe, via Bridge, distributes across 70+ countries.
Hot-warm signing infrastructure: Pre-fund 8 to 16 hot wallets per chain with sequenced nonce management. Fireblocks and Anchorage both expose this pattern.
Internal ledger plus periodic onchain reconciliation: Coinbase Institutional's default. Customers see instant balance updates; the chain sees a handful of net-settlement transfers per day.
Orchestration platforms like Eco's intent-based routing, LI.FI, and Squid coordinate across providers and chains so a single payment intent can land on whichever chain has capacity at that millisecond. This is increasingly the design pattern for treasuries that refuse to be locked to a single rail. See how bridging-based routes fail and what reduces error rates for the failure-mode deep dive.
Where does throughput become a contract obligation?
Enterprise contracts increasingly specify TPS floors, p99 latency ceilings, and uptime credits. BVNK and Fireblocks publish enterprise SLAs covering signing latency. Anchorage's qualified-custody charter requires audited settlement windows. Coinbase Prime's institutional agreement covers internal ledger uptime but not onchain inclusion. The contract language differs by provider, and so does the financial remedy. What enterprise settlement providers actually promise in their SLAs is the companion read.
For treasury desks moving size, throughput is also an execution-quality question. RFQ flows (request-for-quote) handle large blocks via a small number of high-touch transactions, while OTC desk integrations batch differently. Stablecoin OTC execution vs RFQ compares the two for treasury operators.
Methodology and sources
TPS figures combine each provider's public docs, status-page incident logs, audited disclosures, and chain-level metrics from public validator dashboards (Solana Beach, Etherscan, Arbiscan, BaseScan) as of Q1 2026. Volume figures from official press releases and audited financial statements. Error rates from public status pages where available, otherwise marked as not disclosed.
Sources: Circle Developer Docs (CCTP V2), Stripe Stablecoin Payments docs, BVNK 2024 disclosure, Coinbase Q1 2024 shareholder letter, Anchorage Digital qualified-custody charter, Fireblocks Network statistics, Solana Foundation validator data, DeFiLlama stablecoin volume API, Etherscan and BaseScan public RPC stats.