Finnacl Research

R9 · free executive summary

The Spread Business in a Technology Costume

what a stablecoin issuer actually earns

The Spread Business in a Technology Costume

Executive summary · free edition · September 2026

Complete in itself. The finding is here, not held back to make you buy the report.

This edition expires on 18 January 2027, when the United States licensing regime takes effect and the central question changes. A successor edition, not a re-score, is required after that date.


What the business actually is

95.21% of the largest listed issuer's revenue is interest on short-dated dollar assets.

$667.7m of $701.3m. The reserves are cash plus a fund whose longest-dated holding is three months. The token, the network and the technology are the distribution mechanism.

This is a short-rate spread business, and its own accounts say so.

The number that decides it

From the company's own interest-rate sensitivity table:

A 100 basis point rise in short rates
Reserve income +$737m
Distribution costs paid away −$360m
Net +$377m — 180% of annualised pre-tax profit

Run it backwards. A fall of 55.5 basis points erases the entire annualised pre-tax profit.

The whole enterprise sits about half a percentage point from zero. That is not a technology company's risk profile. It is a money-market fund's, without the disclosure conventions.

The divergence in the title

Over the year
Onchain transaction volume +151%
Tokens in circulation +19.46%
Revenue +6.57%

The gap is a 66 basis point fall in the rate earned on the reserves. Usage more than doubled and revenue barely moved, because the company is not paid for usage.

Where the money actually goes

Follow one dollar of reserve income:

Reaches the token holder 19.82 cents
Stays at the distribution partner 28.79 cents
Stays at the issuer 51.39 cents

The issuer paid a single distribution partner $324.6m — 48.61% of reserve income and 46.28% of all revenue. Total distribution costs are 58.81% of revenue.

And it reconciles from the other side. That partner's own filing discloses one counterparty at 26% of its total revenue, and the figure reconciles to the issuer's disclosure in all four periods examined.

Both halves of this arrangement are public. Neither has been put next to the other.

Growth does not improve the split. Of the $737m from a 100 basis point rise, $360m goes straight out — 48.85%, within a quarter of a point of the average share. Marginal equals average.

An issuer paying nearly half its gross margin to one distributor is a supplier, not a platform. That is a description, not a judgement, and it is visible in two sets of filings.

Why the holder gets so little, and it is the law

The statute forbids the issuer paying yield to holders — read at the primary text, section 4(a)(11).

That single provision explains the entire structure. It is why the distribution partner pays rewards instead of the issuer, and why another company registers its yield-bearing product as a face-amount certificate rather than as a stablecoin. Who pays whom is a regulatory artefact, not a commercial choice.

What the incoming regime actually does

Effective 18 January 2027, confirmed at the statute and in the Treasury's own rulemaking footnote
A later provision 18 July 2028
Reserve maturity cap 93 days
Issuer paying yield to holders Forbidden
Rules finalised None. Every one is still proposed, with 131 days to run

The reserve cap formalises what the reserves already are — the largest issuer's longest holding is three months today. The regime does not change the business model. It writes the current one into law. That is a different and less exciting claim than the sector makes for it, and it is what the statute says.

The funds named for this

Fund What its largest holdings actually are Net assets
The one named for stablecoin technology Four cryptocurrency products at 29.38%, against 5.25% in the payment company its name points at $2,133,265
A tokenisation fund 29.27% cryptocurrency products on a banks-and-exchanges portfolio $1,020,959
A digital-payments fund Zero stablecoin exposure $180,240,820
The plain Treasury-bill fund $19.42bn of Treasury bills, weighted average maturity 36.35 days — exactly what it says 9,103× the first fund

Two of the largest cryptocurrency fund issuers registered stablecoin funds in July and September 2025. Neither has launched. Effectiveness was re-designated to 11 and 13 September 2026, after at least fourteen and eleven months of successive delays.

The entire investable stablecoin-equity shelf is two funds holding $3,154,224 between them. The honest product in this category is a Treasury-bill fund, and it is nine thousand times the size of the one carrying the theme's name.

And the issuers say it themselves. From one provider's own prospectus, describing its own fund's risk: "the performance of Stablecoin & Tokenization Companies may not be primarily driven by the adoption, volume or stability of any particular stablecoin or tokenization initiative."

A correction to our own prior work

An earlier Finnacl screen recorded one of these funds at 22.96% and 4.03%, taken from a secondary aggregator. The issuer's own file contradicts both — the correct figures are 29.38% and 5.25%.

The finding survives and is stronger. Our numbers were wrong. They are corrected here and in the source document, with the correction marked in place. It is the clearest argument we have for reading issuers' own files, and we learned it the expensive way.

The tokenised-Treasury figures, reconciled

An earlier pass found three published figures that could not be reconciled and refused to pick one. The reason is now established: they were never contemporaneous.

Live and measured: $2.786bn of a $15.87bn market — 17.56%, replicated at an independent source 1.13% apart. And one widely quoted figure is not what it appears: a regulatory filing showing $5.136bn is cumulative amount sold, not assets under management, across 28 investors.

What we cannot tell you

  • Rates may simply not move, in which case the sensitivity is true and inert. We print what is currently priced — +0.51 points implied one year forward (R5 printed +0.52 by its own method; the difference is method, not disagreement) — and say plainly that forwards are a poor guide, because R5 established they lost to assuming rates stay put at every horizon tested.
  • One quarter of one issuer carries much of the analysis. It is the largest listed one, and that is stated rather than generalised.
  • The distribution figure is a disclosed contract that can be renegotiated.
  • Every rule is proposed. Nothing here forecasts what final rules will say.
  • A statutory drafting defect exists in the section examined — one subsection cross-references another that does not exist — so at least one provision may be interpreted differently than we read it.

Verification. 612 automated checks, none failing. 125 planted defects across five mutation rounds, all caught. Forty-one errors were found and corrected inside the pack, three of them material. Thirty-three defects were found in the pack's own checker — 24 blind spots and 9 parsing faults — all documented rather than repaired quietly.

Access. A government full-text endpoint was CAPTCHA-gated and was not bypassed; the printing office's own service supplied the same document. Columns marked "login to unlock" on one site were not accessed and no account was created.

Disclosures. The author holds no position in any company, token or fund named here. No issuer payment, no advertising. Impersonal research, identical for every purchaser, tailored to no one. No statement about what any reader should hold, and no forecast. Finnacl Research is not an investment adviser.

The full report adds the nine-route snapshot series, the complete statutory reading, the reconciliation chapter and the evidence appendix.

That was the whole finding, not a teaser. The full report adds the evidence behind it: every figure with its source and date, the Tier-1 snapshots, a plain-language version you are licensed to hand to a client, and the quarterly re-scores for as long as the edition is live.