Fact 01
Stablecoins are increasingly used as a transaction currency
On-chain stablecoin volume now exceeds Bitcoin's, despite a far smaller supply, and it turns over far more often. Transfers have also shifted toward small, everyday amounts as costs collapsed — by 2025, more than a third of stablecoin transfers were under one dollar.
36% of stablecoin transfers were under $1 in 2025
Circulation: stablecoins vs. Bitcoin
End-of-month circulation, USD billions. Source: paper, Fact 1.
Monthly on-chain volume: stablecoins vs. Bitcoin
Monthly on-chain transfer volume, USD billions. Source: paper, Fact 1.
See more — velocity, transfer sizes, and falling costs
Velocity — how often the money turns over
Monthly turnover = volume ÷ circulation. Bitcoin sits near 1; stablecoins are far higher, consistent with use as a medium of exchange rather than a store of value.
Transaction sizes are shifting smaller
Share of transfers by US-dollar size bucket, stablecoins vs. Bitcoin.
Cheaper rails, more micropayments
Left: average cost of an Ethereum stablecoin transfer (an upper bound — other chains are cheaper). Right: count of sub-$10 stablecoin transfers across all chains, in millions.
Fact 02
Stablecoins have a global footprint
Activity is spread roughly evenly across the Americas; Europe, the Middle East and Africa (EMEA); and Asia-Pacific (APAC). Using a method that infers a wallet's region from the time of day it transacts, the study finds most volume actually settles within regions — the standard industry approach overstates cross-border flows by 15–22 points. And balances are migrating toward Asia-Pacific.
APAC's share of balances rose 16% → 42% from 2022 to 2025
Where stablecoin volume originates, over time
Share of wallet-to-wallet volume by sender region, monthly. Source: paper, Fact 2 (Bayesian attribution).
See more — corridors, the cross-border correction, and the shift to Asia
Region-to-region corridors
Gross flows as a share of all wallet-to-wallet volume that year. Rows = sender, columns = receiver. The diagonal (same-region) dominates.
Most volume stays within a region
Cross-regional share of volume. The standard "visitor-share" method (grey) overstates cross-border activity versus the two timing-based models in the paper.
Balances are migrating toward Asia-Pacific
Annual-average share of end-of-month stablecoin balances by region.
Fact 03
Stablecoin wallets behave like cash accounts
Wallet owners hold a buffer of balances, spend them down, and top up when they run low — the same pattern households show with checking accounts. A classic cash-management model from monetary economics fits the data well. Since 2023, access to on-chain rails has been getting cheaper and more frequent.
Wallets hold roughly 4–5 days of spending as a buffer
Balances held, in days of spending
Balance-to-consumption ratio: average stablecoin balance ÷ daily outflow. Higher = a thicker cash buffer.
How often wallets top up
Replenishment frequency — deposits per year. Rising since 2023 as on-chain access improves.
See more — the cash-management model and who holds stablecoins
Access to the rails (model parameter p)
Free-adjustment intensity: how often, per year, a wallet can replenish at no cost. Rising since 2023 indicates improving deposit technology.
Who holds stablecoins
Wallet-level statistics by balance size, pooled 2020–2025 (Ethereum EOAs, 1/16 sample).
Fact 04
On-chain credit markets are expanding
Stablecoins increasingly serve as the settlement asset for on-chain lending — secured, over-collateralized credit has grown several-fold, and a newer unsecured, real-world market has expanded fast. And contrary to a common worry, this isn't draining bank credit: stablecoins mostly substitute for cash, and the dollars backing them recycle straight back into the financial system.
Secured on-chain credit: $10B → $41B
Outstanding on-chain credit
Total outstanding borrowing, USD billions. Secured (left axis) is over-collateralized DeFi lending; unsecured (right axis) is newer real-world credit. Different scales.
Does the growth of stablecoins destroy bank credit?
A common worry is that stablecoins pull money out of bank deposits and shrink lending. The evidence points the other way — these exhibits show why.
Stablecoins are replacing cash, not deposits
US currency growth by denomination (% per year, bars) vs. stablecoin issuance ($bn/yr, line). Transactional $1–$20 notes have stopped growing as stablecoins surged; the cash still growing — $100 notes — is held mostly abroad. Source: FRED CURRCIR + Federal Reserve denomination tables.
Why credit isn't destroyed
- Deposits are relabeled, not drained. Converting a deposit to a stablecoin moves the dollar into the issuer's reserves — no loan is touched.
- Stablecoins substitute for cash. The marginal holder is often abroad, using dollars for payments and savings once served by banknotes — bypassing the domestic deposit base.
- Banks can lend more, not less. A digital outside option pushes deposit rates up, drawing funding in; even 1-for-1 displacement cuts lending by only about a quarter.
- Credit is also being created on-chain, while reserves recycle into Treasuries that stay in the system.
Net effect on bank deposits: between −$10B and +$25B per $100B of stablecoins issued — under 0.15% of the US deposit base, with the central estimate a slight increase.
Fact 05
Stablecoins enable programmable, machine-initiated payments
Because stablecoins are software, value can be moved by code. Today most stablecoin value already flows through smart contracts — programs that execute payments for lending, settlement, and custody at machine speed. And a newer frontier is emerging: autonomous AI agents paying for services over open protocols like x402, almost entirely in USDC.
69% of stablecoin value moved through smart contracts in 2025
Most stablecoin value moves through code
Monthly share of USDC + USDT transfer activity on Ethereum with a smart contract on at least one side — by transaction count and by dollar volume. The volume share reaches ~80% late in the sample.
The agentic frontier: x402
Monthly x402 transactions (bars) and volume (line). x402 lets software pay per request; 99.9% settles in USDC. A late-2025 burst, then a rebuild from a lower base — small in absolute terms, but a working machine-payment rail.
See more — who transacts with code, what it does, and machine micropayments
Transacting with code is now normal
Share of active stablecoin wallets (EOAs) that transacted with at least one smart contract during the year — up from 8% in 2020 to 50% in 2025.
What the contracts do (2025)
Share of contract-involved stablecoin volume by function. Lending — the Fact 4 credit market — is the single largest workload; machine-native commerce is still tiny.
A market for true micropayments
Share of x402 transactions that are micropayments (under $0.01). The median such payment was a fraction of a cent — orders of magnitude below card-rail minimums.
What x402 payments were for
Monthly x402 transactions by use case, millions.