# Stablecoin Payments Volume: What Transfer Charts Actually Measure

By Pawnie (@pawnie_ on X). Web3. Published 2026-11-08. Canonical page: https://www.w3match.com/blog/stablecoin-payments-volume-vs-transfers

Understand stablecoin payment volume, adjusted transfers and provider-reported payments, with a practical way to compare datasets without mixing their definitions.

Stablecoin payment volume measures activity classified as payments. Stablecoin transfer volume measures tokens moving between addresses. The second can include exchange operations, trading and repeated movements of the same balance. Filtering those movements improves the data, but an adjusted transfer total still needs a separate test of what counts as a payment.

That distinction changes what a chart can tell you about adoption. A large transfer number may show that the settlement infrastructure is busy. It does not, by itself, show how much businesses collected from customers or how many invoices were paid.

## Three measurements that answer different questions

| Measurement          | What it tries to count                           | What it cannot establish alone                               |
| -------------------- | ------------------------------------------------ | ------------------------------------------------------------ |
| Unadjusted transfers | Stablecoins moved in the covered onchain dataset | The purpose of each movement                                 |
| Adjusted transfers   | Activity remaining after a provider's filters    | That every remaining transfer paid for goods or services     |
| Identified payments  | Activity attributed to a stated payment use case | Complete coverage of every provider and offchain transaction |

The definitions belong beside the numbers. “Adjusted” is a methodology label, not a universal category shared by every dashboard.

The [Visa Onchain Analytics transactions dashboard](https://www.visaonchainanalytics.com/transactions) separates adjusted and unadjusted activity. It also categorizes use cases, rather than treating all adjusted activity as payments. Its published approach combines address labels and behavioral filters, so the result depends partly on what the data provider can identify.

This is useful for comparing activity within a defined dataset. It is not a receipt from the merchant at the end of every transfer.

## One invoice can create several transfers

Consider a fictional business paying a $1,000 invoice. It sends stablecoins to a processor. The processor moves the funds to a settlement wallet, then to a venue that converts the balance into the recipient's local currency.

If each onchain movement is $1,000, a dataset that adds all three movements records $3,000 of transfers. The customer paid one $1,000 invoice. This example illustrates the counting problem; it is not a claim about a measured provider or the exact filters used by Visa or Artemis.

The same issue appears when an exchange moves funds between its own wallets. The chain records a real transfer. Whether that movement belongs in a payments dataset depends on the question the dataset is designed to answer.

For a payment business, useful evidence would connect the invoice, processor and recipient. Onchain addresses alone may not expose that complete chain of relationships.

## What a bottom-up payments study adds

Provider data can reveal purpose that is difficult to infer from wallet behavior. A payroll processor knows which transfers settle payroll. A business-payments provider may know which transactions pay suppliers. That context is different from a label identifying an address as belonging to a company.

The [Artemis methodology for its 2025 bottom-up study](https://www.stablecoin.fyi/methodology) combines payment-firm data with estimates for additional providers. That page says data were collected in May 2025 and uses February 2025 as its latest month. It discloses incomplete coverage and possible duplication between connected providers. It describes a historical sample, rather than a current census of all stablecoin payments.

Use that distinction when reading its growth figures. Volume can rise because existing participants processed more payments, because a new provider entered the sample, or because the methodology gained visibility into activity that was already happening.

## Compare the denominator before comparing growth

Suppose a report states that payment activity doubled. Before treating that as customer growth, inspect four details:

1. **The period.** A month's activity multiplied by twelve is an annualized run rate. It is not twelve months of observed payments.
2. **The provider set.** A growing dataset can change the total even when the original participants are flat.
3. **The transaction purpose.** An invoice payment, a trading deposit and prefunding for a payment processor are different economic actions.
4. **The boundary.** An onchain dataset may miss payments settled inside a custodial platform. A provider dataset may see them, but include fewer firms.

This does not make one source unusable. It tells you which comparisons require a common scope.

## Adjusted volume and payments can move differently

A trading event can increase stablecoin settlement without an equivalent rise in merchant payments. A payment provider can grow while market-wide trading activity falls. A chain migration can shift activity between networks without adding a new end customer.

Artemis changed its Terminal and API transfer-volume methodology in July 2026\. Its [current methodology](https://www.artemis.ai/docs/data-reference/stablecoin-methodology) says the base measure now counts gross transfers, excluding mints and burns. The previous measure removed duplicates, intra-exchange transfers and specified trading activity. Its separate Artemis-filtered and peer-to-peer metric variants are being retired.

The API field names remained unchanged. A comparison built from an old snapshot and a new response can therefore mix two definitions under the same label. Keep the methodology version beside the data. This change does not turn gross transfers into payments, and it does not establish that every Artemis delivery surface follows the same update schedule.

For a report, keep the series separate. Plot transfers, adjusted transfers and identified payments with their own definitions. If you calculate a percentage between them, state whether their coverage and time windows actually match. Otherwise, the result can look precise while measuring the difference between two datasets.

## A useful stablecoin-payments report should disclose its limits

At minimum, a reusable data table needs the period, covered chains and tokens, classification method, treatment of internal transfers, and whether values are observed or estimated. It should also say how it handles providers that route payments through each other.

For a country comparison, ask whether the location describes the sender, recipient, provider registration or wallet attribution. For a business comparison, distinguish payment value from the provider's revenue. Processing $1,000 does not mean earning $1,000.

If the report cannot establish the purpose of a flow, call it a transfer. If it covers selected providers, name the sample. A smaller, well-defined number is more useful for a business decision than a larger total whose boundary changes between paragraphs.

## Where agent payments fit

Protocols such as x402 provide a way to attach payment to a request. That is a payment mechanism, not proof that every stablecoin transfer on the supporting network came from an AI agent.

To estimate agent-payment usage, the evidence would need to connect paid requests, successful settlement and a stated set of services. Read [how x402 handles API payments](https://www.w3match.com/blog/x402-explained-ai-agent-api-payments) for the request flow. Keep the activity count separate from the chain's total transfer volume.

**Sources reviewed October 20, 2026.** This article explains measurement. It does not publish a new live estimate of global payment volume. The $1,000 example is illustrative. Dashboard definitions and coverage should be checked again when a report is updated.
