Abstract
We document five facts about payment stablecoins using granular blockchain data from 2020 to 2025. First, stablecoins are functioning as a transaction currency: their onchain transaction volume now exceeds Bitcoin's despite a far smaller outstanding supply, their velocity has risen sharply, and their transaction-size distribution has shifted toward small-value transfers. Second, we develop a Bayesian refinement of the standard visitor-share attribution methodology that uses transaction-timing data to help geolocate individual wallets. This approach shows that stablecoin activity is distributed broadly across the Americas, Europe, and Asia, but that standard methods substantially overstate the cross-regional share of flows. Third, we estimate a structural model of cash management on stablecoin transaction data and find that the typical stablecoin wallet operates on quantitative terms comparable to traditional cash holdings in terms of wallet replenishment opportunities and estimated transaction costs. Fourth, stablecoins are becoming the settlement currency for a rapidly expanding onchain credit market in which lending terms are enforced by rule-based smart contracts rather than by intermediary discretion; balance-sheet accounting disciplined by reserve and currency data indicates that stablecoins' growth reallocates rather than destroys credit. Fifth, stablecoins support programmable, machine-initiated payments: smart contracts participate in roughly a third of stablecoin transactions and over two-thirds of transfer volume on Ethereum, and early agentic-payment protocols such as x402 demonstrate machine-to-machine commerce at sub-cent transaction sizes. Together, these facts suggest that stablecoins are behaving as tokenized money. We discuss implications for traditional financial markets and the design of regulatory frameworks.