Stablecoins Moved $891bn in Third Quarter After Excluding Exchanges and Bots

Stablecoin transfers totalled $891bn in the third quarter of 2026 once trading infrastructure and automated activity were stripped out, according to research announced by Coinbase on 20 October.

The figure is more than four times the $216bn recorded on the same basis in the first quarter of 2023. It points to substantial growth in stablecoin movement beyond the parts of the market that tend to inflate headline on-chain numbers.

What the research measured

The study was carried out by The Block Pro Research and commissioned by Coinbase. It relies on blockchain data supplied by Allium, and is intended to serve as a benchmark for how large, and how fast-growing, stablecoin transfers are once certain categories of activity are excluded.

Those exclusions are broad. The filter removes transactions involving exchanges, decentralised finance protocols, cross-chain bridges and other forms of crypto infrastructure. It also strips out automated transactions and the issuance or redemption of stablecoins themselves.

What remains is the $891bn figure. It is therefore not a count of every stablecoin movement recorded on public blockchains, which would be considerably larger, but a measure of the activity left over after those categories are set aside.

The report is careful about what that residual figure represents. It counts transfers, not confirmed purchases from merchants, and it does not identify how many distinct individuals or organisations are behind the activity.

Growth rates and projections

According to the report, adjusted stablecoin transfers have grown at an annualised rate of 49.9% since early 2023. That growth has not been smooth, varying from one quarter to the next.

Extrapolating forward, the researchers estimate that quarterly volume could reach nearly $1.5tn by the fourth quarter of 2027 if the current pace of growth holds.

That projection rests entirely on the assumption that recent growth persists. Compounding rates near 50% a year are difficult to sustain over long periods in any market, and the report frames the figure as conditional rather than as a forecast.

Why the filtering matters

Stablecoins are tokens designed to hold a steady value, usually pegged to the US dollar. They have become the main settlement asset within crypto markets, which means a large share of their on-chain movement reflects trading: deposits to and withdrawals from exchanges, collateral shuffled between lending protocols, and liquidity moved across chains.

Automated activity adds a further layer of noise. Bots that arbitrage prices or rebalance positions can generate very high transfer volumes without representing any underlying economic exchange between separate parties. Minting and burning of tokens by issuers likewise inflates raw totals.

Raw on-chain stablecoin volume has long been cited as evidence that the tokens rival established payment networks. Critics have countered that such comparisons overstate real-world use because so much of the activity is circular. By removing these categories, the Coinbase-commissioned research attempts to isolate something closer to genuine transfers between users, even if it stops short of claiming to measure payments for goods and services.

The growth from $216bn to $891bn on the filtered basis is notable precisely because it is not driven by the trading and automated activity that dominates headline figures. It suggests that stablecoins are increasingly being moved for purposes other than shifting funds around crypto venues.

Activity outside banking hours

A second finding concerns timing. Drawing on a separate dataset from Dune Analytics, the report concludes that 73% of screened stablecoin transfer volume took place outside US banking hours during the 12 months to 31 July 2026.

That result speaks to one of the main arguments made for stablecoins: that, unlike traditional bank transfers, they can settle at any hour, including evenings, weekends and holidays. A large share of after-hours activity is consistent with users valuing that round-the-clock availability, though the data alone does not explain motive.

The screening for this analysis differs from the quarterly figure. It excludes transfers below $1, transfers between a user's own addresses, and minting and burning. However, it explicitly retains automated transactions, which the Allium-based quarterly total removes.

The two analyses also cover different periods. The adjusted $891bn relates to the third quarter of 2026, while the timing analysis spans the year ending 31 July 2026.

Limits and caveats

Because the datasets, filters and time windows differ, the 73% share cannot be applied to the $891bn total to estimate how much money moved outside banking hours. The report presents the two findings as separate pieces of evidence, one pointing to growth in filtered transfers and the other to the prevalence of after-hours activity.

The inclusion of automation in the timing analysis is a particular consideration. Bots operate continuously and have no reason to favour banking hours, so their presence could push the after-hours share higher than it would be for human-initiated transfers alone. The report does not break out that effect.

There are other reasons for caution. The research was commissioned by Coinbase, a company with a commercial interest in stablecoin adoption, although it was conducted by a third-party research unit using external data providers. And while the filter removes many obvious sources of inflated volume, a transfer that passes it is not necessarily a payment, a remittance or a business transaction. It may still involve the same person or entity moving funds between addresses that the filter fails to identify as linked.

The absence of data on distinct users also limits interpretation. Rising transfer value could reflect broader adoption, larger transactions by a relatively small number of participants, or some combination of the two.

What to watch next

The findings are likely to feed into continuing debate over how stablecoins should be measured and how significant their non-trading use has become. Several questions remain open:

• Whether filtered transfer growth continues near the 49.9% annualised rate, or slows as the base grows larger
• Whether quarterly volume approaches the nearly $1.5tn projected for the fourth quarter of 2027
• Whether future research can separate human from automated activity in timing data, or link transfers to verified commercial use and distinct users

For now, the report offers a more conservative benchmark than raw on-chain totals. Even on that narrower basis, it shows stablecoin transfers expanding rapidly over the past three and a half years.