Counting each transfer once, and four other corrections
A transfer between two venues we both track was counted twice. Fixing it removes $459M across 21 venues. A second fix takes $1.29bn of VALR's own money out of its deposit figures. Both restate the whole series.
- −11.7%
- Luno, the largest move
- $1.29bn
- VALR's own money
- $240,109
- Hisa, now Full coverage
most of its deposits come from another tracked venue
no longer counted as deposits
Solana rail captured
Five changes landed, four of them on 1 November. Two move numbers we have already published, and they move them in the same direction: down.
A transfer between two tracked venues was counted twice
When money moved from one venue we track to another venue we track, both scans recorded it. The receiving venue booked a deposit, and so, in effect, did our total.
Each transfer now counts once, attributed to the venue receiving it. That removes $459M across 21 venues.
The size of the correction at each venue is a measure of how much of its deposit flow arrives from somewhere else we already watch:
| Venue | Before | After | Change |
|---|---|---|---|
| Luno | $2.18bn | $1.92bn | −11.7% |
| Daya Pro | $29.90M | $26.82M | −10.3% |
| Quidax | $764.40M | $742.90M | −2.8% |
| Obiex | $11.80bn | $11.71bn | −0.8% |
| Busha | $161.49M | $160.22M | −0.8% |
| Divest | $85.77M | $85.22M | −0.6% |
| VALR | $24.43bn | $24.34bn | −0.4% |
| Monica | $206.85M | $206.57M | −0.1% |
| Spenda | $107.58M | $107.49M | −0.1% |
Luno moved most, by 11.7%, because a large share of what arrives in its deposit addresses comes from another venue in our coverage rather than from outside it. That is a fact about Luno's position in the market, and it only became visible once we stopped counting those transfers twice.
Chipper, Roqqu, Jeroid, Grey Finance, Accrue, Breet and Raenest each moved by less than 0.15%.
$1.29bn of VALR's own money
One address had been counted as a VALR customer deposit address since December 2025. It holds VALR's own money.
Its $1.29bn is no longer counted as deposits. VALR's lifetime deposits move from $24.34bn to $23.06bn.
This is the failure mode we watch for hardest, because it is invisible from the outside: an address that behaves like a deposit box, receives like a deposit box, and belongs to the venue. It is now in the exclusions ledger with its reason and its value, where anything we remove from a membership set goes.
The Index
Two changes, affecting grades only — no deposit figure moves.
VALR and Luno are now graded on the customer deposits we publish for them, not on the desk-sized deposits that are excluded from those figures and disclosed separately. Grading a venue on money we deliberately do not publish was wrong.
| Venue | Before | After |
|---|---|---|
| VALR, October | B 48.5 | BB 58.4 |
| Luno, October | CCC 25.4 | CCC 33.6 |
| Spenda, October | AA 80.0 | A 77.6 |
Spenda moves because the basis change alters where every venue sits relative to the others, not because anything about Spenda changed.
Evolution is removed from the Index from October 2026, after it stopped operating. Its standings up to September stand unchanged — a venue that was there is not retroactively erased.
Hisa reads Full
Hisa's Solana rail is now captured: 126 deposit addresses and $26,347 of deposits going back to 18 July. Its lifetime figure moves from $213,762 to $240,109, and its coverage reads Full instead of Partial.
Hisa did not collect anything on Solana until 8 October, which is why the rail only became measurable this month. When we said in October's coverage note that the published figure was a floor, this is the part that was missing.
What this means for a figure you already have
Anything published before 1 November may not match the same figure today, at any grain. The direction is down, and the largest single moves are Luno at −11.7% and VALR's $1.29bn.
We would rather restate in public and tell you exactly what moved than quietly carry a number we no longer believe.
Method
These changes took effect on 1 November 2026 and apply to the whole history, not only to periods since. Deposit figures are restated at every grain — daily, weekly, monthly, lifetime — so a figure published before that date may not match the same figure today. The double-capture fix compares each transfer against the deposit-address sets of every venue AfriFlux tracks and attributes it once, to the receiving venue; previously a transfer whose sender and recipient were both tracked was recorded by each venue's own scan. The VALR correction removes a single address from that venue's deposit-address set and records it in the exclusions ledger with its reason and its dollar value. The Index changes affect grades only, not the underlying deposit figures. Figures quoted are lifetime deposits, which at Luno and VALR include the per-client boxes above $1M that are excluded from published customer-deposit figures and disclosed separately.
How attribution works →Cite this
Free to cite with attribution. No embargo, no approval needed.
AfriFlux (2026). Counting each transfer once, and four other corrections. Published 2 November 2026. https://www.afriflux.xyz/feed/counting-each-transfer-once
Reproducing a chart? Use this line: Source: AfriFlux — on-chain intelligence for Africa's stablecoin economy. afriflux.xyz