Weekly recap: $10.69M, and a record inside a slower week

$10.69M

This week

retail stablecoin volume

-2.4%

Week on week

from $10.96M

$5.07M

Quidax

a record, 47.4% share

68.4%

Top two venues

Quidax and Spenda combined

Retail stablecoin volume across the 13 venues we track came to $10,693,813 for the week of 27 July to 2 August — down 2.4% from $10.96M the week before. That is the first weekly decline since the run began in mid-July.

The interesting part is where the decline came from, because it did not come from the top.

Quidax set a record inside a falling week. It cleared $5,073,490, beating the $4.9M it set seven days earlier, and took 47.4% of everything we measured. Spenda added $2,243,843, up 10.2%. Together the two largest venues accounted for 68.4% of the market — the highest concentration we have recorded.

Everything below them fell. Busha dropped 22.9% to $766,208 and Roqqu 22.8% to $289,973, both in the same week. Onboard fell 70.1% to $172,109, the sharpest single move on the board.

So the market did not slow evenly. It slowed everywhere except the top, and the concentration we flagged in last week's Gradient read tightened rather than eased.

Deposits tell a different story than volume

Spenda processed 57,677 deposits from 20,811 active wallets. Quidax processed 12,053 from 3,178 — a quarter of the deposits and a sixth of the wallets, for more than twice the volume.

That gap is a difference in what the two venues are for, not in how well they are doing. Spenda's median deposit is small and frequent; Quidax's is large and occasional. A single number ranking them by volume hides that entirely, which is the argument the Gradient framework exists to make.

Native stablecoins: cNGN is now nine tenths of it

Transfer volume across the native African stablecoins we track came to $2,872,838, down 19.4% on the week.

cNGN accounted for $2,575,796 of it — 89.7%, and it was the only token that grew, up 19.3% across 13,855 transfers.

Every other token fell, several by more than 90%. We would caution against reading those as collapses. Transaction counts across the non-cNGN tokens held steady all fortnight at between 159 and 809 a day while their dollar volume swung by two orders of magnitude — the signature of a handful of large transfers moving a small base, not of activity disappearing. On a base that thin, one transfer is a trend.

The durable observation is the concentration itself: native stablecoin activity outside cNGN remains small enough that weekly percentages are not yet meaningful.

Coverage change: Noblocks removed

We removed Noblocks from coverage this week while we revalidate its attribution, and restated the full history on the remaining 13 venues.

We found the problem while mapping the Paycrest Gateway. The Gateway records which app each order belongs to. Checking those records against the wallets we had attributed to Noblocks, we found 63 different apps using the same relayer. Our original attribution identified Noblocks users through that relayer key — and the relayer belongs to shared wallet infrastructure, not to Noblocks.

The pipeline was not the issue. The wallet attribution was.

Restating on 13 venues moves prior weeks in both directions: the week of 20–26 July now reads $10.96M rather than the $10.6M we published at the time, because removing Noblocks and improving attribution on the remaining venues pull opposite ways. The prior peak, the week of 29 June, restates from $9.2M to $8.81M.

Previous Noblocks posts stay up and now link to this correction. The methodology page documents the change.

Attributed, not estimated. This one wasn't.