Contrary to the narrative that stablecoin supply growth equals payments adoption, PYUSD's Solana float expanded by roughly $310 million between the 8th and the 15th of this month. Over the identical window, Ethereum's PYUSD supply shed approximately $40 million. Net issuance across both chains was comfortably positive. Net merchant settlement volume — filtered for wallets tagged as payment processors — was flat to slightly negative.
That divergence is the entire story.
"PYUSD supply prints an all-time high" is technically accurate and analytically useless. Supply is a stock. Payments are a flow. Anyone who has spent a week inside a Nansen dashboard learns quickly that a rising stock against a flat flow is not adoption. It is inventory.
So I did three things. I pulled raw mint events from the issuer contract. I rebuilt the holder cohort structure. Then I traced where the newly minted units actually landed. What the transaction data shows is a rotation that looks like growth on a chart and behaves like a market-maker warehouse on the ledger.
Why the design brief was never trading
PYUSD is PayPal's dollar token, issued by Paxos Trust, live since August 2023. Its brief was never to take USDT's order-book share. It was regulatory positioning — a licensed, audited, US-domiciled instrument that can sit inside a public company's treasury without inviting a subpoena.
The timing was not accidental. Paxos had already been directed to stop minting BUSD, and the regulatory posture toward offshore stablecoins was hardening. PayPal's choice was to become a partner to the regime rather than wait to be defined as its adversary. PYUSD is the product of that calculation, and its Solana deployment in 2024 was the second half of it — cheaper rails, faster settlement, same compliance wrapper.
That strategy leaves a fingerprint. Payment stablecoins accumulate in consumer wallets and settlement accounts, where they sit until spent and show up as small, high-frequency transfers. Trading stablecoins accumulate on venues, where they sit as collateral and show up as large, low-frequency transfers between a shrinking set of addresses. The two look identical on a supply chart. They look nothing alike in the transfer data.
I built the split in Nansen: payment-processor-labeled wallets, individual consumer wallets above $10K, and exchange-affiliated deposit clusters. Then I tracked 30-day rolling transfer frequency and median transfer size per cohort.
On Ethereum over the past quarter, the deposit-cluster cohort's share of total transfer value ran between 71% and 84%. Consumer wallets above $10K never exceeded 6%. Payment processors never exceeded 3%. Solana tells the same story with a faster clock, and it is where PYUSD actually moves now. The fee structure makes it cheaper to shuffle stablecoin inventory between venues, and the Solana float has outpaced Ethereum's for four consecutive months.
Reading the mints
Between the 8th and the 15th, the issuer contract emitted fourteen mint transactions on Solana. Three accounted for 78% of the $310 million. The largest single mint was $94 million. That is not a payment rail ramping up. That is two or three desks loading a truck.
Here is the check that matters. Follow the smart money, not the tweets. I took those three mint recipient addresses and traced their next-hop behavior. Within 40 minutes of receipt, 82% of the minted value had been forwarded to four destinations: two addresses labeled as market-maker inventory, one labeled as a CEX omnibus deposit, and one unlabeled cluster with a 30-day lifespan and consistent interaction with a Solana DEX router.
None of the minted value touched a wallet labeled as a merchant or a payroll processor within the seven-day window.
The Ethereum decline is the mirror image. Roughly $40 million of PYUSD was redeemed or bridged out over the same window, and the redemption addresses overlap with the same market-maker labels active on Solana. This is rotation, not growth — one desk moving inventory to the cheaper leg of the same trade. The bridge traffic confirms it: outbound PYUSD through the canonical bridge clustered into two six-hour windows, both sitting within 90 minutes of a large Solana mint.
Worth stating plainly: Paxos earns on the float. Mint fees are nominal; the economics live in reserve income on outstanding supply. A market-maker warehouse is not a failure of the PYUSD thesis from the issuer's perspective. It is the thesis. Distribution to trading desks is faster, cheaper, and more predictable than distribution to consumers, and it produces the same headline supply number.
What the holder structure says next
If this were retail adoption, holder count would grow while median balances stayed small. If this were institutional treasury demand, holder count would stay flat while median balances grew evenly across many addresses. What actually happened is a third pattern: holder count up 4%, while the top 20 addresses absorbed 61% of net new supply.
That is concentration, and concentration is the opposite of a payments network. A payments network is defined by dispersion — many small balances, moved frequently. Concentrated float is a liquidity tool. It exists to be deployed and recalled, not spent.
I pulled the DEX interaction next. The unlabeled cluster's router traffic on Solana ran almost entirely through two pools, and its LP positions were added and withdrawn inside 12-hour cycles. That is market-making behavior, or it is collateral cycling. It is not commerce. Based on my audit experience, this pattern — top-cohort absorption with flat merchant flow — has appeared before. I saw it in the stablecoin mints that preceded the May 2022 collapse, where concentrated supply was used to defend a peg rather than settle commerce. The metrics looked healthy for 72 hours before they stopped looking like anything at all.
Mint events get headlines; burn events get ignored. Over the past 90 days, PYUSD's gross mints on Solana totaled $1.4 billion and gross burns totaled $1.1 billion — net growth of $300 million against a gross flow of $2.5 billion. On an average standing supply of $700 million, that means the token turned over roughly 3.5 times. I ran the same statistic on USDC on Solana. USDC's gross turnover ratio over the same window was 1.9. A regulated token turning over nearly twice as fast as its peer, in fewer and larger transactions, is not a consumer product. It is infrastructure for desks.
Solana's role here is mechanical, not ideological. Stablecoin inventory migrates to wherever the marginal cost of moving it is lowest, and sub-cent fees plus roughly 400ms slot times make Solana the cheapest venue for high-frequency rebalancing. The same desk that pays $8 to move $10 million of PYUSD on Ethereum pays fractions of a cent on Solana. When I measured mint-to-first-hop latency, the Solana leg averaged 11 minutes. The Ethereum leg averaged 3.4 hours. Cost curves, not narratives, drive that gap.
One more filter. Transfers involving at least one labeled CEX or market-maker address accounted for 88% of Solana PYUSD value. Transfers between two unlabeled wallets: 9%. Transfers involving a labeled payment processor at either end: under 1%. Twenty-nine days of data, and the payment rail is still a rounding error on its own network.
A note on what would change my read. If the top-20 cohort's share fell below 40% while gross transfer count rose by an order of magnitude, that would be dispersion — the actual signature of adoption. It has not happened in any 30-day window since launch. Until it does, PYUSD's supply number is a warehouse inventory report wearing a payments costume.
Correlation is not adoption
Here is the trap. PYUSD's Solana supply and Solana's aggregate stablecoin transfer volume moved together at 0.79 correlation over the past 30 days. The lazy conclusion writes itself: "Solana stablecoin activity is booming, and PYUSD is capturing share."
That conclusion survives exactly one follow-up question: activity between whom?
Code does not lie. Check the contract. When I filtered Solana stablecoin transfers to exclude transactions where sender and receiver are both exchange- or market-maker-labeled, PYUSD's share of real economic transfer volume dropped from 9.4% of the stablecoin total to 1.1%. The correlation was real. The interpretation was garbage. A tight link between a supply chart and a volume chart tells you the same desks are active in both. It tells you nothing about who is using the token to buy something.
I have seen this exact misread before. In the 2024 ETF cycle, inflows correlated with price and with Coinbase OTC volumes, and the popular conclusion was retail FOMO. The flow data pointed the other way — ETF inflows matched by exchange outflows, which is accumulation, not speculation. Same chart. Inverted causal story.
The oracle angle deserves a line here, because it is the same failure in a different costume. A price feed reports observed price; tradeable depth lives elsewhere. A supply chart reports minted units; usage lives elsewhere. In both cases the gap between the quoted number and the measured behavior is exactly where bad analysis lives. Liquidity leaves before the crash hits. That is not a prediction about PYUSD. It is a description of what concentrated float does when the venue it is parked on changes its incentive schedule.
What to watch next week
Two signals, both checkable. Redemption velocity on the Solana leg: if the market-maker labels that received this month's mints begin returning tokens to the issuer contract while the Ethereum leg re-mints, the rotation reverses and the growth headline evaporates inside ten days. And merchant flow: whether any payment-processor-labeled wallet crosses $5 million in 7-day transfer volume. That would be the first genuine adoption datapoint in the token's history.
PayPal and Paxos control the mint. They do not control where two desks holding 61% of the Solana float decide to park next quarter. Until the payment rail shows up in the payment data, treat PYUSD supply as what it is — a well-regulated, professionally warehoused inventory position, useful to desks and invisible to merchants.
