Visa has lowered adjusted stablecoin volume, but there is still insufficient basis to conclude that actual payment activity has declined. Transaction count fell by less than 2 percent, while volume was adjusted more sharply. This largely reflects changes in classification standards rather than actual on-chain transfer amounts.

On Sept. 27 (local time), blockchain media outlet CryptoSlate reported that Visa lowered adjusted stablecoin volume through an on-chain data update on Sept. 18 and reduced adjusted transaction count by less than 2 percent. According to Visa’s on-chain analytics change log, the update increased the number of labeled addresses included in Allium’s identity dataset to about 600 million from about 15 million.

Visa kept its definition of adjusted volume unchanged. It excludes from the adjusted metric transfers tied to labeled exchanges and contracts, infrastructure such as bots and bridges, and minting and burning. But as the number of identified addresses rose sharply, more transfers that had previously been included were excluded from adjusted volume. The update added heuristics to identify short-term routing and also changed how it distinguishes organic activity from payment activity.

Visa’s public materials also show inconsistencies related to the number of addresses. The current transaction methodology says there are "more than 3 million" labeled addresses, but the Sept. 18 change log states that the new full identity dataset expanded to about 600 million. Visa did not explain whether the earlier figure refers to a different subset or is wording that has not yet been updated. Therefore, the 3 million figure cannot be treated as the prior baseline for this data update.

The adjustment did not change the transfers recorded on-chain. Visa did not disclose figures comparing adjusted stablecoin volume under the same standard before and after the change, so it is not possible to calculate how much volume fell because of the adjustment. Data that can be compared over the same period and standard have also not been released for the revised scale on individual blockchains such as Ethereum, Tron and Solana, or for changes in chain-by-chain rankings.

The gap between volume and transaction count stems from how they are tallied. Transaction count treats each transfer as 1 transaction regardless of the amount, while volume is calculated based on the transferred amount. Excluding some large transfers can therefore sharply reduce volume while having only a limited effect on the overall transaction count.

Visa cited a case on Solana in which an automated program repeatedly moved the same stablecoin into thousands of one-time wallets. The program moved very large amounts with relatively few transactions, and this pass-through type of activity was excluded from adjusted volume across multiple chains. But this single case does not mean the same activity occurred on all chains, and it has not been confirmed how much volume was revised on Solana or across the overall market.

Caution is also needed in interpreting stablecoin transfer volume as actual payment scale. Visa distinguishes payments from DeFi, centralised exchange flows, investment and trading, store of value, minting and burning, short-term routing and infrastructure. Not all transfers included in adjusted activity are classified as payments. "Retail-sized" transactions, meaning transfers of less than $250, also do not automatically indicate purchases of goods or merchant settlement.

Research by the Bank for International Settlements (BIS) also points to these measurement limits. In a study published on Sept. 15, it analysed that the same stablecoin can be used differently depending on the blockchain, appearing more closely linked to smart contract interaction on Ethereum, and to off-smart-contract holdings and uses for transactions and store of value on Tron.

Another BIS study analysed 593 million transfer events related to USDT, USDC and PYUSD on Ethereum in 2025 and found that nearly 60 percent of all transfer events occurred within complex transactions. Because a single transaction can include multiple token movements and financial actions, treating one transfer event as an independent payment can overestimate real economic activity. The study added that the figure is limited to the Ethereum sample.

Ultimately, Visa’s data update alone makes it difficult to judge that stablecoin payments have declined. What can be confirmed is that as classification standards and address labels became more refined, high-value pass-through activity was excluded more from adjusted volume, and adjusted transaction count fell by less than 2 percent. Before-and-after comparative data compiled under the same standard are needed to verify actual payment usage or chain-by-chain changes in activity.