$1.15 billion left two of the largest tokenized Treasury funds in the world in 28 days.

Same allocator. Same credit line. BlackRock’s BUIDL and Janus Henderson’s JTRSY, both drawn down between 24 August and 21 September 2026.

If all you saw was the flow data, you would call it a run.

Then you check what that same allocator contributed to protocol revenue over the quarter. It went up.

That gap, between what a tokenized fund redemption looks like and what it actually means, is the most useful thing buried in the Sky Ecosystem Q3 2026 Quarterly Report.

Here is what happened, and why it should change how you read onchain outflows.

What a $1.15 Billion Tokenized Fund Redemption Actually Looks Like

The figures come straight from the Q3 report, published 29 October 2026.

• $600 million of BlackRock BUIDL redeemed
• $550 million of Janus Henderson JTRSY redeemed
• $831 million of USDS drawn through the same allocator, repaid
• Window: 24 August to 21 September 2026

Position sizes tell the rest. Janus Henderson exposure through Grove fell from $1,242.5 million at the Q2 close to $703.2 million. BUIDL fell from $714.0 million to $119.3 million.

Combined, $1,956.5 million became $822.5 million. A 58% cut to tokenized fund exposure inside a single quarter.

For scale: BUIDL carried around $2.2 billion when the instant-redemption rails for it went live in May 2026.

A $600 million exit by a single allocator is not a rounding error in a fund that size.

Why Redeem BlackRock BUIDL and Janus Henderson JTRSY in the Same Month?

Because something better was already built and waiting for the capital.

On 15 July 2026, Grove committed USDS capital as warehouse lender behind a $500 million senior secured facility for Galaxy Digital.

Galaxy underwrites, originates and services institutional loans backed by BTC and ETH. Anchorage Digital and BitGo hold the collateral.

By the Q3 close that facility had financed $375 million of loans. Seventy-five percent of capacity, in under three months.

Total Galaxy exposure through the allocator went from $27.6 million at the Q2 close to $393.1 million.

One detail from the case study stuck with me: a $320 million loan can go from underwriting to funded over a weekend.

Warehouse credit funded mortgage and auto originators for decades. This version runs on a stablecoin instead of a bank balance sheet, and it settles while the bank is shut.

So the sequence reads: redeem the short-duration tokenized Treasury positions, repay the credit line, write secured credit at a wider spread.

Redemption was step one of three. Step two was repayment. Step three was a $500 million warehouse facility that was already 75% drawn by quarter end.

The Detail That Makes This Redemption Story Almost Too Neat

Here is the part I enjoyed.

In May 2026, Grove launched Basin, a facility offering up to $1 billion in daily stablecoin liquidity for instant redemptions out of tokenized funds.

The two launch products? BUIDL and JTRSY. Both managers joined as launch partners.

Basin existed to close the settlement gap, the multi-day lag between an approved redemption and cash actually landing, in a tokenized Treasury market then worth around $15 billion.

Four months later, the same allocator redeemed $1.15 billion out of those exact two funds.

You can read that cynically. I read it as the only honest version of infrastructure: you build the exit, then you prove it works by using it at size.

Most redemption rails in this market have never been tested above nine figures.

The One Number That Breaks the Panic Reading of Tokenized Treasury Outflows

Forced exits shrink the revenue line. This one grew it.

Grove contributed $26.68 million to Gross Protocol Revenue in Q3 2026, up from $24.38 million in Q2.

That was the largest contribution of any allocator in the Sky Agent Network, out of $52.65 million from agents in total.

Hold the two facts next to each other:

• Tokenized fund exposure down 58%
• Revenue contribution up 9.4%

Distress does not look like that. Distress looks like selling the liquid positions to cover the illiquid ones, with the revenue line falling alongside.

Sky Protocol closed the quarter with $99.29 million in Gross Protocol Revenue and $37.19 million in Net Revenue, up 11.8% from Q2.

Sixth consecutive positive quarter. The Net Revenue margin widened to 37.5% as total expenses fell 16.0%.

What Onchain Redemption Flows Show That Traditional Fund Flows Hide

This is the part most allocators still underrate.

In traditional markets you learn about a billion-dollar rotation weeks later, in aggregate, in a filing, with no counterparty detail. By then the trade is cold and the reasoning is gone.

Onchain, the redemption, the repayment and the redeployment are three separate movements, each timestamped, each traceable to one mandate. You can reconstruct the decision rather than guess at it.

Which is why the Q3 report could publish counterparty allocations outright:

• Janus Henderson, $703.2 million
• Anchorage, $210 million
• PayPal USD, $197.2 million
• BlackRock BUIDL, $119.3 million
• Securitize STAC, $103.4 million

Try getting that level of disclosure, quarterly, from a money market fund complex.

Context helps here. Tokenized US Treasuries sat near $14.7 billion in late October 2026, roughly 37% of a $39.3 billion tokenized real-world asset market.

Flows of this size used to be invisible. They are not anymore, and that visibility cuts both ways.

Traditional fund flows tell you what happened. Onchain redemption flows tell you what is happening, who did it, and what they did next.

How the Sky Savings Rate, USDS and sUSDS Actually Fit Together

Worth stepping back, because the mechanics explain the incentive.

Sky Agents are independent allocators, not subsidiaries. They borrow USDS under risk parameters set in public, deploy it, and pay fees for that access. Those fees are where the Sky Savings Rate comes from.

• Allocators borrow USDS. Sky Agent Vaults held $5.95 billion at the Q3 close, 52.2% of Protocol Collateral
• They deploy into secured credit, Treasury strategies and lending markets
• Fees flow back to Sky Protocol
• Sky Governance sets the Sky Savings Rate out of that revenue
• sUSDS holders receive it automatically, with no action required

The rate is variable and governance-set. It moved from 3.52% in July to 3.60% in September, then to 3.80% on 7 October 2026. No emissions funding it. No trading volume propping it up.

That is why a rotation like this one matters to anyone holding sUSDS. Rate payments to sUSDS holders were the protocol’s single largest expense in Q3, at $41.94 million. Better deployment at the allocator layer is what keeps that payment fundable.

Circulating sUSDS reached $4.72 billion on 21 October 2026, up 118% year over year.

Is This Risk Discipline, or Concentration With Better Framing?

Fair challenge, and worth putting on the table rather than dodging.

Cutting exposure to one fund manager by 43% inside a quarter is diversification.

Taking exposure to a single counterparty from $27.6 million to $393.1 million in that same quarter is the opposite of diversification.

Both statements are true at once.

The counterargument runs roughly like this. The tokenized Treasury positions were liquid, rated and low-yielding.

The warehouse facility is senior secured, over-collateralized with BTC and ETH held at qualified custodians, and sized at $500 million against a Protocol Collateral base of $11.41 billion at the Q3 close.

Concentration risk is real. It is also bounded by parameters published before any capital moved.

Reasonable people land in different places on that trade. Publishing the position sizes is what makes the argument possible at all.

What to Watch in Q4 2026

Three things, if you are tracking this.

• Whether the warehouse facility fills past 75% of capacity
• Whether the rotation holds through Q4, or BUIDL exposure quietly rebuilds
• Sky Reserves, which closed Q3 at $85.08 million against a $150 million medium-term Solvency Reserve target

Two outside signals landed after quarter end. S&P Global affirmed its rating with a stable outlook on 1 October.

Moody’s assigned its first counterparty risk rating on 7 October, making Sky Protocol the first stablecoin protocol rated by both agencies. Standard Chartered initiated research coverage in September.

Galaxy also added $100 million of sUSDS to its corporate balance sheet and approved sUSDS as eligible collateral across a lending business carrying an average loan book of $1.4 billion and more than 1,600 counterparties.

Rune Christensen put the quarter this way in the foundation’s release:

For years, the industry asked when institutions would arrive onchain. Q3 showed that they are already integrating at the balance-sheet level.

Total Collateral stood at $13.60 billion on 11 November 2026, up from $11.41 billion at the Q3 close.

So: rotation, or retreat?

I read it as rotation, and the revenue number is hard to argue around. But I would rather hear the other side than agree with myself.

If you think a 58% cut to tokenized fund exposure in 28 days signals something I have missed, make the case below. I will take it seriously.

An Allocator Redeemed $1.15 Billion of Tokenized Funds in a Month. Deliberately. was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.

The most important line in the OCC’s final stablecoin rule is a line that is not in it.

On 21 October 2026, the Office of the Comptroller of the Currency finalised its rulemaking under the GENIUS Act.

Most of the coverage led with the yield provision. Fewer people noticed this one: the OCC asked whether every permitted payment stablecoin issuer should be limited to a single brand, read the comments, and declined to write the rule.

One issuer. As many branded stablecoins as it likes. In most cases, sitting on one shared pool of reserves.

That is bigger than it sounds.

What Is the One Brand Per Issuer Stablecoin Rule the OCC Just Walked Away From?

Answer first. There is no one brand per issuer stablecoin rule in the United States. The OCC considered one. It chose not to adopt it.

How it played out:

• In the March 2026 proposal, the OCC asked for comment on capping each issuer at a single stablecoin brand.
• The case for the cap was contagion. If one branded coin loses confidence and holders rush the exit, a shared reserve pool can drag the issuer’s other brands down with it.
• Commenters pushed back. Morrison Foerster’s read is that most saw added regulatory burden with little real risk reduction, and a requirement inconsistent with the statutory text.
• The OCC agreed on the law. Davis Polk notes it cited the burden on issuers, especially smaller ones and nonbank fintechs that depend on white-label distribution, plus the absence of a clear statutory mandate.

Then comes the part worth sitting with.

The OCC still said that a one-brand structure may better protect the rights of stablecoin holders, because it keeps each brand’s reserves legally separate and cuts contagion risk.

So the regulator named the risk, conceded the fix would help, and left it optional.

Scope is worth being precise about. The rule covers the issuers the OCC licenses.

The Federal Reserve, the FDIC and the NCUA have their own proposals in flight, and they do not all agree. Both the GENIUS Act and this final rule take effect on 18 January 2027.

After that date, issuing a payment stablecoin in the United States without being a permitted payment stablecoin issuer, or holding a transition waiver, is unlawful.

A Multi Brand Stablecoin Issuer Is Not a Thought Experiment. It Is the Market.

This is not a rule about some future structure. It describes how stablecoins are already issued.

The brand on the token is frequently not the company holding the reserves.

• Paxos is the issuing entity behind PYUSD, SoFiUSD and USDG. Three brands, three distribution stories, one regulated issuer.
• Bridge, Stripe’s issuance arm, runs more than ten live branded coins through its Open Issuance product, including MoneyGram’s MGUSD and MetaMask’s mUSD.
• At least eleven platforms publicly market white-label or stablecoin-as-a-service issuance in 2026, among them Anchorage, BitGo, Brale, Agora, M0 and Coinbase.

Think about what that means at the counter. PYUSD is a PayPal product in every way a user experiences it. The reserve machinery sits with Paxos.

Coinbase does not issue USDC itself. The brand you trust and the balance sheet you are actually exposed to are often two different companies.

Reed Smith’s reading of the final rule is blunt. By declining to require a separate issuer per brand, the OCC preserved room for regulated issuers to support multiple branded products and widened the lane for white-label arrangements.

Scale sharpens it. DefiLlama put the total stablecoin market at roughly $290.8 billion on 10 November 2026. Two issuers hold most of it.

And counting is getting harder, not easier: 382 stablecoins were being tracked by mid-2026.

More brands. Fewer issuers. The same reserve pools underneath.

Reporting Is Not Segregation, and That Gap Is the Whole Story

Here is the detail that deserves more attention than it is getting.

Under the final rule, a multi-brand issuer has to report reserves at the brand level. It does not have to hold them separately.

The difference is everything.

• Reporting tells you what the issuer says sits behind each brand.
• Segregation decides what happens to your brand when a sibling brand breaks.

The OCC also split from its sister agency here. Skadden notes the FDIC had proposed requiring multibrand issuers to keep segregated reserve pools per brand absent written approval.

The OCC declined to follow, choosing to assess segregation case by case during licensing.

There is a second wrinkle. The final rule also made technical changes to accommodate issuers that use multiple special purpose vehicles.

That gives multi-brand structures more room, not less. Flexibility was the explicit goal.

The Federal Reserve has been flagging the underlying problem for a while. In an April 2026 FEDS Note, Fed staff wrote that complex intermediation chains in stablecoin service provision introduce significant contagion risks and transparency challenges.

So the structural question stays open. The regulator sees it. The rule leaves it to discretion.

Which moves the work somewhere specific: onto holders, to verify for themselves.

If Nobody Has to Separate the Reserves, Who Actually Checks the Collateral?

This is where onchain collateral transparency stops being a slogan and becomes a procedure.

A branded fiat stablecoin asks you to accept a chain of attestations. The issuer reports. An auditor attests. A regulator reviews on a cycle. You read a PDF and take it on trust.

Sky Protocol is built the other way round. The collateral backing USDS is verifiable onchain, continuously, by anyone, without asking an issuer for a number.

In practice:

• Protocol Collateral stood at $13.85B on skyeco.com at the time of writing, against a stablecoin supply of 8.61B.
• Every risk parameter, collateral type and debt ceiling is set through Sky Governance by SKY token holders, onchain, with a public record of who decided what and when.
• Capital is deployed by the Sky Agent Network: independent allocators including Spark, Grove, Keel, Obex and Osero that access USDS liquidity under governance-set limits. They are sovereign businesses, not subsidiaries taking instructions.
• Two public dashboards carry the evidence. financial.skyeco.com holds the balance sheet. info.skyeco.com holds the live onchain state. Any figure quoted on the site can be checked against them.

Architecture matters more than the promise here. There is no custodial balance standing between a holder and the collateral.

Users retain control of their own assets at all times, which changes the shape of the question entirely.

The OCC’s debate was whether brand separation protects holders. There is a second answer: verification does.

Where the Sky Savings Rate, sUSDS and USDS Sit in a Multi Brand World

Short version: different architecture, different question for the reader.

USDS is the base unit of capital in Sky Ecosystem. Overcollateralised, non-custodial, convertible 1:1 from major stablecoins through the Peg Stability Module with no fee and no slippage.

sUSDS is the yield-generating version. Supply USDS, receive sUSDS, and the position accrues the Sky Savings Rate programmatically. No lockups, no manual claiming.

The Sky Savings Rate is a variable rate set by Sky Governance. It was showing 3.80% APY on skyeco.com at the time of writing.

It is funded from revenue Sky Protocol accrues when independent allocators pay for access to USDS liquidity. Not from a yield-sharing side deal with a brand partner.

From the Q3 2026 quarterly report published by Sky Frontier Foundation on 29 October:

• Gross Protocol Revenue of $99.29M, up 14.2% year over year. A $397.2M annualised run rate.
• Net Revenue of $37.19M, up 11.8% quarter over quarter. The sixth consecutive positive quarter.
• Net Revenue margin widened to 37.5% from 31.0%.
• Protocol Collateral of $11.41B at quarter end, up 25.3% year over year.
• sUSDS supply of $4.52B, up 118% year over year, and $4.72B by 21 October.
• Sky Reserves of $85.08M after a $28.47M transfer in the quarter.

One data point lands harder than all of them. In Q3, Galaxy added $100 million of sUSDS to its corporate treasury and approved sUSDS as eligible collateral across its institutional trading business.

What to Check Before 18 January 2027

Whatever you hold, the questions are the same.

• Who is the legal issuer? Not the brand on the token. The entity behind it.
• How many brands share that issuer? One point of failure, or several.
• Pooled or segregated reserves? Brand-level reporting is not segregation.
• Can you verify it yourself? Or are you waiting on the next attestation.
• Where does the yield come from? A reserve book, a partner’s rewards programme, or protocol activity you can audit.

The OCC chose flexibility over structure. That is a defensible call, and the commenters made a real case for it.

It also quietly moved the burden onto you.

So here is the question I actually want answered below: did the OCC get this right, or did it hand the contagion problem to holders and call it optionality?

Tell me which side you land on.

One Issuer, Many Brands: The Rule the OCC Chose Not to Write was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.

Six days in October changed how onchain credit gets talked about.

On 1 October 2026, S&P Global Ratings affirmed a B- issuer credit rating with a stable outlook on Sky Protocol.

On 7 October, Moody’s Ratings assigned the same protocol a B3 long-term counterparty risk rating, also stable. It was the first rating Moody’s had issued on any stablecoin protocol.

On the published mapping tables, B3 and B- are the same rung. Two agencies. Two teams. Two frameworks. One answer.

Except they were not answering the same question.

The agreement is in the letter grade. The disagreement is in everything that produced it.

That gap is the most useful thing to come out of October for anyone trying to underwrite onchain exposure.

Most of the coverage reported the grade and moved on. The grade is the least interesting part.

What the two frameworks chose to count, and chose not to count, tells you far more about how traditional credit analysis is going to treat onchain balance sheets from here.

Here is what each framework measures, where they split, and the one thing they both landed on.

What Moody’s and S&P Global crypto ratings actually measured

Three separate published instruments are now attached to one protocol. People keep collapsing them into a single headline. They are not the same product.

• Moody’s assigned a counterparty risk rating, not an issuer credit rating. A counterparty risk rating asks whether a counterparty can meet its financial obligations. B3, stable outlook, 7 October 2026. Sky Frontier Foundation corrected its own announcement headline the following day to make that distinction clear.
• S&P Global assigned an issuer credit rating. B-, first assigned in August 2025, affirmed with a stable outlook on 1 October 2026. That one asks how creditworthy the issuer is.
• S&P Global also publishes a stablecoin stability assessment, a completely separate scale. USDS and DAI sit at 4, constrained.

Three questions. Three scales. One protocol.

Why a stablecoin stability assessment is not a credit rating

S&P says this plainly on its own product page, and it is worth repeating, because the market keeps blurring the two.

A stability assessment measures whether a token can hold its peg, factoring in both redemption and secondary market price.

It runs from 1, very strong, to 5, weak. It compares stablecoins with each other.

A credit rating measures whether the obligation gets redeemed on demand.

It compares stablecoins with bonds and other rated debt instruments. It explicitly does not reflect secondary market price volatility.

The mechanics inside the stability assessment are where it gets interesting.

S&P starts with an asset assessment covering credit risk, market value risk and custody risk.

Then it asks how far overcollateralization and liquidation mechanisms offset those risks. That produces a score.

Then five more factors get applied on top:

• Governance
• Legal and regulatory framework
• Redeemability and liquidity
• Technology and third-party dependencies
• Track record

Here is the part most people skip. Those five can only pull the score down. In S&P’s own wording, the final assessment can be in line with the asset assessment or lower. It cannot be higher.

The overcollateralization gap, and where the two rating methodologies split

This is the real divergence. It is not a disagreement about facts. It is a disagreement about what counts.

In S&P’s stability framework, overcollateralization is an explicit mitigant. Available overcollateralization and the robustness of liquidation mechanisms sit inside the asset assessment itself. Post more collateral, offset more asset risk.

Moody’s counterparty framework measures something else entirely: tangible common equity against tangible managed assets.

Reported coverage of the rating put that at roughly $90 million of equity against about $10 billion of managed assets as of September, a ratio near 0.9%.

Collateral a borrower posts against a loan is not equity. It secures a position. It does not absorb a loss at the protocol level.

So a protocol can be heavily overcollateralized and thinly capitalized at the same time. Both frameworks are reading the same balance sheet and pricing different lines on it.

Overcollateralization is a mitigant in one framework. It is not capital in the other. Both statements are published, and both are true.

If you have ever argued with someone about whether a collateral ratio makes a stablecoin safe, that argument just got settled in the least satisfying way available. It depends which question you were asking.

Where Moody’s and S&P do agree: capitalization is the binding constraint

Two separate teams, two separate methodologies, and they named the same weakness.

Moody’s cited very low capitalization, confidence-sensitive stablecoin liabilities, and the operational, governance, legal and regulatory risks that come with a DAO structure and a permissionless stablecoin model.

S&P cited weak risk-adjusted capitalization, high depositor concentration, centralized governance, and regulatory uncertainty around decentralized protocols.

Both credited the same strengths: low historical credit losses, a material share of liquid assets, and profitability relative to asset risk.

Moody’s noted roughly 45% to 50% of assets held in stablecoins and tokenized money market funds, with around 25% in crypto-backed loans.

Moody’s then did something agencies rarely do so publicly. It printed the number.

• Upgrade considered if the capital ratio rises above 2.5% and is expected to stay there
• Review for downgrade if it falls below 0.5%, on consecutive quarterly losses, or on weakening liquidity

That is a scoreboard anyone can check.

What a rated protocol actually does about a flagged weakness

The Q3 2026 report from Sky Frontier Foundation, published 29 October, reads like a direct response to that scoreboard.

• Sky Reserves stood at $85.08 million at quarter close, against a $150 million medium-term Solvency Reserve target
• $28.47 million was remitted to reserves during the quarter, against a $2.64 million deficit in the same quarter a year earlier
• Up to 50% of monthly Net Revenue is routed to reserves until the target is met
• Gross Protocol Revenue reached $99.29 million, up 14.2% year over year
• Net Revenue was $37.19 million, a sixth consecutive positive quarter, with margin widening from 31.0% to 37.5%

Excluding draws made during the period, the report states Sky Reserves would have grown 38% in the quarter.

A published target, a published funding rule, and a live dashboard to check the build against.

Whether it closes the gap is a question for the next few quarters, not for this post.

How the Sky Savings Rate, sUSDS and USDS sit inside a credit frame

Worth separating the pieces, because the ratings do.

USDS is the base stablecoin and the unit of account for borrowing. sUSDS is the yield-generating version.

Supply USDS, receive sUSDS, and the position accrues at the Sky Savings Rate with no lockups and no exit fees. The rate is variable and set by Sky Governance, currently 3.80% APY.

The yield is not conjured out of nothing. Sky Agents borrow USDS liquidity under governance-set risk parameters and deploy it into diversified strategies. Revenue from those deployments is what funds the rate.

What the agencies rated is the protocol. Not the savings rate, not the token price. That distinction matters the next time someone forwards you a headline.

The numbers underneath, as of early November: $15.23 billion of total collateral against 9.59 billion of stablecoin supply.

sUSDS supply was $4.52 billion at Q3 close, up 118% year over year, and $4.72 billion by 21 October.

Galaxy added $100 million of sUSDS to its balance sheet in September and approved it as collateral across its institutional trading business.

Why one rated protocol in a $293 billion market is worth your attention

Total stablecoin market capitalization sat at roughly $293.8 billion on 3 November 2026, according to DefiLlama. Hundreds of tokens. One protocol carrying two agency ratings.

That is not a quality claim. It is an access claim.

A great many institutional mandates simply cannot hold an instrument that no recognized agency will score, regardless of how the instrument actually performs.

A rating does not make an exposure good. It makes it legible to a risk committee that needs a comparable number in a column.

Sky Frontier Foundation has said the long-term ambition is an investment-grade rating.

That is a long way up from B3 and B-, and both agencies have already published what would move it.

The question I keep turning over

If your framework treats borrower-posted collateral as a risk mitigant, you get one answer.

If it treats loss-absorbing equity as the only buffer that counts, you get a different one.

Crypto has spent a decade arguing the first position. Rating agencies have spent a century on the second.

Which one would you underwrite with? And if you run a risk framework of your own, where does borrower-posted collateral sit in it: mitigant, capital, or neither?

Tell me in the comments, and tell me why. I will argue back.

Moody’s and S&P Are Rating Crypto Now. Their Methodologies Disagree. was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.

Pull up a stablecoin conference agenda from this year. Count the panels on Washington. Then count the panels on Manila.

Now look at the volume.

Asia originated roughly $245 billion of the $390 billion in genuine stablecoin payments recorded in 2025.

That is about 60% of the world’s total, according to McKinsey and Artemis Analytics. North America came second at $95 billion. Europe third at $50 billion.

The money already moved east. The coverage did not.

Search behaviour tells the same story. Most English-language analysis of asia stablecoin payments still frames the region as emerging. It stopped being emerging around the time it passed half the world’s payment volume.

Asia is not an emerging stablecoin market. It is the market. Everywhere else is the long tail.

Why Does Asia Originate 60% of Global Stablecoin Payments?

Not speculation. Plumbing.

• Asia trades with itself. Intra-regional trade made up close to 60% of Asia-Pacific exports in 2024, per UN ESCAP, and almost all of it still clears through correspondent chains.
• The flows are enormous. APAC outbound cross-border payments hit $13.5 trillion in 2025 and are projected to reach $24 trillion by 2035.
• Remittances keep setting records. India received $135.46 billion in FY2024–25. The Philippines received $35.63 billion in 2025, equal to 7.3% of GDP.
• The legacy rail is expensive. The World Bank’s money transfer index sat at 5.52% in Q3 2025 against a UN target of 3%.
• And slow. A B2B payment between Taiwan, Japan and Korea typically takes two to five days, with fees stacking at every intermediary hop.

One detail makes the point better than any chart. Taiwanese manufacturers shipping AI hardware into Latin America have been paying up to 11% in local transfer fees, according to Capital Layer’s CEO Justin Wang, and Taiwan’s exporters are now fielding requests from overseas counterparties to settle in stablecoins. The buyer asked first. That is what a demand-side shift looks like.

Triple-A estimates about 327 million people in Asia now hold or use digital currencies. Roughly half the global total.

That is adoption driven by necessity rather than novelty. It tends to stick, and it does not reverse when the market turns.

The $10 Trillion Number Everyone Quotes Is the Wrong One

Three figures get used interchangeably in APAC stablecoin market coverage. They are not the same thing.

• $10.2 trillion in adjusted onchain volume over the twelve months to June 2026, per Visa Onchain Analytics. Mostly trading and treasury movement.
• $4.5 trillion in tracked Q1 2026 volume, with nearly two-thirds originating in Asia.
• $390 billion in genuine end-user payments. Money actually paying for something.

Asia leads all three. The third is the honest one, and B2B dominates it: roughly $226 billion, up 733% year over year.

Ten trillion is money moving. Three hundred and ninety billion is money paying. Neither number tells you what the dollar does in between.

The corridor data is sharper still. Triple-A’s APAC-origin flows to the Middle East grew 604% in 2025 and flows to Europe grew 205%.

Its APAC merchants now see 15% to 20% of processed volume arrive through stablecoin checkout.

Most of that is not card volume switching rails. It is business that card rails never reached.

Why No Regulated Asian Stablecoin Can Pay You Anything

Here is the part regional coverage almost always skips.

Every major stablecoin regime now live in Asia asks for a licensed issuer, full reserves and redemption at par. Every one of them also stops the issuer from paying holders anything for simply holding the token.

• Hong Kong. The Stablecoins Ordinance took force on 1 August 2025. The HKMA received 36 applications and granted exactly two licences on 10 April 2026, to HSBC and Anchorpoint. That is a 5.6% acceptance rate. Requirements include HK$25 million paid-up capital, full backing in high-quality liquid assets, and no yield or interest on balances.
• Japan. JPYC became the first licensed yen stablecoin in October 2025, capped at one million yen per transaction. SBI’s trust-bank-backed JPYSC followed in June 2026 without the cap.
• Singapore. The MAS single-currency framework is being written into the Payment Services Act.
• South Korea. Still no dedicated law. The next digital-asset bill is expected in parliament this month, with the bank-ownership dispute unresolved.

The United States did the same thing. Section 4(a)(11) of the GENIUS Act bars a permitted issuer from paying holders interest or yield solely for holding the coin. MiCA bans it too.

So the rails got regulated and the float got orphaned.

If the issuer cannot pay you, the return has to come from what you do with the token. That is now a structural feature of the market, not a loophole.

What Happens to a Dollar Between Two Payments?

This is the question Asian treasurers are actually asking, and almost nobody is writing about.

A Taiwanese exporter holds working capital across a dozen currencies. A Singapore payment institution pre-funds destination accounts before it can initiate anything. A Manila operator sits on float overnight.

On legacy rails, idle capital was a cost of doing business. Nobody could do anything about pre-funding, so nobody priced it. On stablecoin rails, the same balance sitting still is a choice rather than a constraint.

The market noticed. Yield-bearing stablecoins held in institutional treasury strategies grew from $9.5 billion to over $20 billion in the year to May 2026.

Total stablecoin supply held near record highs through the 2026 downturn, after falling more than 30% in the previous bear market.

Stablecoins stopped behaving like a crypto proxy and started behaving like infrastructure.

Where the Sky Savings Rate Actually Comes From

Sky Protocol is built for exactly that gap: the dollar that has arrived and has not yet left.

USDS is the fully backed unit of account. Convert into it through the Peg Stability Module at 1:1, with no fees and no slippage.

Convert USDS to sUSDS and the position accrues the Sky Savings Rate programmatically. No lockups, no manual steps, non-custodial throughout.

The rate is 3.80% APY, raised on 7 October 2026. It is governance-set and variable, not promised.

Where it comes from matters more than the number itself:

• Sky Protocol makes USDS liquidity available under risk parameters published onchain.
• The Sky Agent Network, made up of independent allocators including Spark, Grove, Obex, Osero and Keel, accesses that liquidity and deploys it into institutional-grade strategies.
• Returns flow back to the protocol. Sky Governance calibrates the Sky Savings Rate against what the system actually generated.

The Q3 2026 report published by Sky Frontier Foundation puts numbers on it. Gross Protocol Revenue of $99.29 million, up 14.2% year over year.

A Net Revenue margin of 37.5%, the sixth consecutive quarter in positive territory. sUSDS supply of $4.52 billion, up 118% year over year, climbing to $4.72 billion by 21 October.

As of today, Protocol Collateral stands at $15.23 billion against 9.59 billion in stablecoin supply.

Every figure sits on a public dashboard. If the mechanics are new to you, start here.

USDS is over-collateralised by design and live across Ethereum, Solana, Base and Arbitrum, which matters for a region that routes capital across chains as casually as it routes it across borders.

Galaxy placed $100 million of sUSDS on its own balance sheet in September and approved it as collateral across its institutional lending business. That is a treasury decision, not a trade.

When an allocator of that size moves a nine-figure position onto a yield-generating stablecoin rather than a money market instrument, it has priced the structure, not the headline rate.

What an Asia-First Stablecoin Treasury Policy Looks Like

Nothing exotic. Four moves.

• Benchmark against $390 billion, not $10 trillion. Compare your corridor to real payment volume.
• Price the float, not the fee. Days of locked receivables cost more than transfer fees in most hardware corridors.
• Separate the payment token from the yield position. The regulated coin moves. The yield-generating one waits.
• Verify instead of trusting. Collateral, rate and parameters are onchain. Check them before anything else.

Asia Stablecoin Adoption: What to Watch Through 2027

• South Korea’s bank-ownership dispute is the single largest unresolved item in APAC.
• Japan and Taiwan may become the first Asian corridor regulated on both ends.
• Local-currency coins stay small. The market remains roughly 99.5% dollar-denominated.
• Gulf to APAC becomes the first cross-regional pair with licensed local coins on both sides.

Asia already won the payment layer. Singapore, Hong Kong and Japan built the regimes, the corridors and the volume.

The open question is whether the region keeps the yield on those dollars, or exports that too, the way it exported settlement for forty years.

What is your corridor, and what does your money do between transactions?

Asia Stablecoin Payments Are 60% of the Global Total. The Coverage Does Not Reflect That. was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.

Corporate finance hands you two boxes. Equity and debt. Onchain protocols do not fit in either.

On 7 October 2026, Moody’s tried anyway. It assigned Sky Protocol a B3 issuer rating with a stable outlook, the first rating it has ever issued to a stablecoin protocol. S&P Global had already landed on B-.

Two agencies. Two methods. One shared conclusion: this thing has a balance sheet now, and the capital sitting under it is thin.

I have spent the last few weeks reading protocol financials properly. The capital structure question is the one almost nobody answers.

A protocol does not issue shares or file a bond prospectus. It still has creditors, a first-loss layer and a residual claim. They just live in smart contracts.

What is a protocol capital structure, and why is everyone suddenly asking?

Answer first. A protocol capital structure is the order in which claims on an onchain balance sheet get paid, and the order in which they absorb losses.

Same job as a corporate capital stack. Completely different instruments.

• Senior claims sit at the top. For USDS, that is whoever holds the stablecoin.
• Loss-absorbing capital sits underneath. Reserves, buffers, retained surplus.
• The residual claim sits at the bottom. That is the governance token.
• Nobody signs an indenture. The waterfall is code, plus a vote.

So why now? Because the numbers finally exist in a form an analyst can read.

Sky Frontier Foundation publishes a quarterly balance sheet for Sky Protocol. Credit teams picked it up. Then they graded it.

The timing is not an accident either. Stablecoins crossed $300B in supply this cycle. At that size, “trust me, it is overcollateralised” stops being an answer.

Where does the debt sit when a protocol has never issued a bond?

In the stablecoin itself.

USDS is the liability. Every unit is a dollar claim the protocol has to honour on demand.

The other side of the ledger is collateral. At the time of writing, $14.15B of Protocol Collateral against $11.48B of stablecoin supply.

That gap is the entire argument.

• Q2 2026 closed with $12.32B of collateral against $12.22B of obligations.
• The $90.26M difference is Protocol Surplus. Equity in everything but name.
• Protocol Collateral grew 45.5% year over year.
• USDS is not borrowed from a lender. It is drawn against approved collateral, or converted 1:1 through the Peg Stability Module.

What counts as equity when a protocol cannot issue shares?

Three layers, stacked.

• Protocol Surplus. Retained earnings under a different name.
• Sky Reserves. The dedicated loss buffer, at $80.36M on 8 October against a governance-set $150M solvency target.
• SKY. The residual claim. 17.08B staked at the Q2 close, or 73.3% of circulating supply.

Now the uncomfortable detail. Sky Reserves went down.

They stood at $93.53M at the end of August. Six weeks later, $80.36M.

A reserve that only ever rises is a marketing number. This one moves, in public, on a live dashboard.

Equity in a protocol is not a share certificate. It is the money that disappears first when something breaks.

Why did Moody’s put the capital ratio at 0.9%?

Because it measured loss-absorbing capital against exposure, the way it measures any financial institution.

Roughly $90M of tangible capital. Roughly $10B of managed exposure. That is 0.9%.

For reference, Basel III sets a 4.5% CET1 minimum, rising to 7% once the conservation buffer is counted.

Four things worth holding in your head at once.

• The ratio is not a claim about the USDS peg. Collateral still exceeds supply.
• Thin equity against a large book and insufficient backing are different questions.
• Moody’s named Sky Reserves as the central factor. The Foundation said publicly that it agrees.
• Both ratings carry a stable outlook. Neither is investment grade.

Sit with the strange part. A protocol published a number that makes it look thinly capitalised on a bank yardstick, then pointed everyone at the live dashboard where you can watch it move.

Most issuers would have buried it.

And this is not the first time a rating has rewritten the balance sheet. After S&P flagged the small surplus buffer last year, Sky Governance voted on 14 March 2026 to cut token buybacks from 75% of surplus to 7.5% and push almost everything else into reserves.

The analyst wrote a note. Governance rewrote the waterfall. That feedback loop barely existed in DeFi two years ago.

How do junior and senior risk capital work onchain?

This is the piece credit people recognise instantly and most crypto readers have never seen.

Allocators in the Sky Agent Network do not simply borrow and disappear. They post first-loss capital before they deploy anything.

• Junior Risk Capital is the allocator’s own treasury, or capital rented from outside investors. It burns first.
• Senior Risk Capital is the USDS drawn from the protocol. It is only exposed once the junior layer is gone.
• How much junior capital is required comes from a Capital Requirement Ratio that varies by asset class.
• Those ratios adapt their logic from the Basel Accords, scored across financial risk, smart contract risk and administrative risk.
• Fall below the requirement and the monthly settlement cycle applies penalties.

That is securitisation tranching, running on a public ledger, with the first-loss position disclosed rather than buried in an appendix.

For context on how rare this still is: onchain tranching accounts for less than 1% of DeFi lending deposits today.

Where does protocol surplus go before anyone gets paid?

There is a published answer, ratified on 13 August 2026. Every settlement cycle splits surplus like this.

• 50% to the Surplus Buffer
• 22.5% to SKY buybacks funding staking rewards
• 22.5% to USDS staking rewards
• 5% to SKY buy and burn

Read the order rather than the percentages. Loss absorption is funded before distributions. Sky Governance can change that split, and has already done so twice this year.

For scale: Sky Protocol generated $107.35M in Gross Protocol Revenue in Q2 2026, and $231.66M across the first half.

Where do sUSDS and the Sky Savings Rate sit in the stack?

Near the top. That is the whole proposition.

sUSDS is how you hold the Sky Savings Rate. Supply USDS, receive sUSDS, and the exchange rate compounds in your favour.

What makes that claim senior:

• The rate is funded from aggregate protocol revenue, not from one strategy or one borrower.
• You are not exposed to any single agent, pool or collateral type.
• It is governance-set and variable, published live rather than promised in advance.
• No lockups, no exit fees. Convert back to USDS whenever you want.

Then look at the three rates the protocol prints at once.

• sUSDS at 3.52%
• stUSDS at 5.07%, which lends against SKY collateral
• Staked SKY at 6.57%

That is not a product menu. That is a yield curve for seniority. The safest claim pays the least, exactly as it should.

And the stack is still being built. On 8 October 2026, governance launched the stUSDS Keeper, formalising a rung that did not exist a year ago. New instruments, added by vote, mid-quarter.

Is institutional capital actually pricing any of this?

Follow the balance sheets instead of the press releases.

• Galaxy added $100M of sUSDS to its corporate treasury on 23 September 2026 and approved it as collateral across its institutional trading business.
• sUSDS closed Q2 2026 at $5.52B, up 149% year over year.
• Sky Protocol is, so far, the only stablecoin protocol formally rated by both Moody’s and S&P.

Nobody puts $100M into a yield-bearing stablecoin because the APY looks nice.

They do it because they can finally locate themselves in the capital structure and know who absorbs the loss ahead of them.

What should you actually watch from here?

Four numbers. All public, all updated without waiting for a quarter to close.

• Protocol Collateral against stablecoin supply. That gap is the cushion.
• Sky Reserves against the $150M target. Loss absorption before distributions.
• Gross Protocol Revenue, quarter over quarter.
• Whether the surplus split changes, and who voted for it.

Equity and debt were invented for entities that can be sued, wound up and liquidated. A protocol cannot be any of those things.

So the industry is rebuilding the stack from first principles, in public, with a rating agency now reading over its shoulder.

Here is where I would like your take.

If a protocol’s collateral comfortably exceeds its liabilities, does a 0.9% capital ratio actually matter? Or is that a bank metric being forced onto something that was never built to be a bank?

Tell me below. I read every response, and I will argue with the good ones.

Protocol Capital Structure: Equity, Debt, and Whatever Onchain Protocols Are Building Instead was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.

Here’s something most stablecoin holders never think about.

The digital dollar in your wallet is probably helping fund the US government.

Not in a conspiracy-theory way. In a plain, legal, slightly boring way. When you buy a reserve-backed stablecoin, the issuer takes your dollar and parks most of it in short-term US Treasury bills.

Do that across a market worth roughly $300 billion and you’ve got a buyer of US debt that barely existed six years ago.

So what does stablecoin growth mean for US debt markets? Stablecoins have become a steady, rules-driven source of demand for short-term Treasuries. Inflows can nudge T-bill yields lower, outflows can push them higher, and in most cases the holder funding all of it earns nothing.

That last part is the one I want to dig into.

Key takeaways

• Reserve-backed stablecoins keep most of their reserves in cash, repos and Treasury bills, so their growth turns straight into demand for short-term US debt.
• Stablecoin issuers still hold less than 1% of all Treasuries outstanding. They move the front end of the curve, not the whole market.
• BIS research links a $3.5 billion stablecoin inflow to 3-month T-bill yields falling up to about 4 basis points within 10 days.
• Most stablecoin holders earn 0% while the reserve income stays with the issuer.
• USDS and sUSDS take a different path. Protocol revenue flows back to holders through the Sky Savings Rate, a variable rate set by Sky governance.

How Do Stablecoins Buy US Debt? The T-Bill Pipeline Nobody Sees

I can explain the pipeline in about ten seconds.

1. You swap $100 for 100 stablecoins.
2. The issuer owes you $100 and holds $100 in reserves.
3. Rules push those reserves into cash and very short-term government debt.

In the US, those rules come from the GENIUS Act, signed in July 2025. The OCC’s proposed rules list what counts as a reserve: cash, bank deposits, Treasury bills maturing in 93 days or less, certain repos, government money market funds, and tokenized versions of these.

The law kicks in by January 18, 2027 at the latest.

So every stablecoin minted under these rules is a fresh bid for T-bills. Every redemption is a sale.

Stablecoins and US Debt by the Numbers

Quick context. The US national debt crossed $40 trillion in August 2026. In the first nine months of fiscal 2026, net interest alone cost $827 billion. That’s more than the US spent on defense.

Washington needs buyers. Stablecoins have turned into one.

• $24 billion to $270 billion. USD stablecoin supply growth from 2020 to June 2026, per a new paper by Nellie Liang and Brent Neiman.
• Nearly $35 billion. What stablecoins spent on T-bills in 2025, roughly matching the largest government money market funds, per the BIS.
• Less than 1%. The share of Treasuries outstanding held by stablecoin issuers, according to Treasury’s own borrowing advisory committee.
• $800 billion to $1 trillion. Extra T-bill demand Standard Chartered sees if the market hits $2 trillion by end-2028.

Treasury Secretary Scott Bessent has floated similar numbers. I’d treat them as scenarios, though. Growth has cooled this year alongside crypto trading volumes.

Do Stablecoins Actually Move Treasury Yields?

A little. And now we can actually measure it.

The BIS working paper Stablecoins and safe asset prices studied daily data from January 2021 to March 2026.

A $3.5 billion inflow lowered 3-month T-bill yields by 0.71 basis points on impact and up to 4 basis points within 10 days. Longer maturities barely moved.

An IMF working paper found the same direction. When stablecoin market cap jumps 1%, 1-month and 3-month yields fall. Small moves, but they’re real.

The part that gets less airtime is the exit. An earlier version of the BIS study found outflows pushed yields up two to three times as much as inflows pushed them down.

Money that arrives slowly can leave all at once. That asymmetry is the part of stablecoins and US debt that keeps regulators up at night.

There’s a fair counterpoint too. The Kansas City Fed points out that if people fund stablecoins out of bank deposits, banks may buy fewer Treasuries. So net new demand could be smaller than the headlines suggest.

The Catch: Stablecoin Holders Fund the Debt, Issuers Keep the Yield

T-bills pay yield. Somebody collects it.

With most reserve-backed stablecoins, that somebody is the issuer. The GENIUS Act bars payment stablecoin issuers from paying you yield, so the reserve income stays on their side of the table.

Put simply: you supply the dollar. The issuer buys the T-bill. The issuer keeps the yield.

And a lot of those dollars just sit there. By Sky’s estimate, roughly 80% of stablecoins aren’t deployed into any source of yield. That’s money quietly funding Treasuries without earning its holder a cent.

Which is why “who keeps the yield?” matters as much as “who buys the debt?”

USDS, sUSDS and the Sky Savings Rate: Where the Yield Goes Instead

Sky.money gives you non-custodial access to Sky Protocol, which handles the same dollar differently.

USDS is Sky Protocol’s fully backed stablecoin, and it’s overcollateralized. On October 23, 2026, Sky.money showed $16.24 billion in collateral behind 10.02 billion USDS, about $1.62 for every dollar. You can check the mix yourself on the Sky financial dashboard. Here’s what backs a stablecoin if you want the full checklist.

Where does the revenue come from? Sky Agents tap USDS liquidity and put it to work across collateralized loans, US Treasury bills and lending markets. What they earn flows back into Sky Protocol.

For you, it works like this:

1. Supply USDS and receive sUSDS.
2. sUSDS accrues the Sky Savings Rate (SSR) automatically.
3. Exit to USDS anytime. No lockups, and no fees through Sky.money.

The SSR is a variable protocol rate funded from aggregate Sky protocol surplus. SKY holders raise or lower it through public onchain votes.

One honest detail: sUSDS isn’t exposure to any specific Agent, strategy or T-bill. You get the rate, not a slice of one position.

The numbers are public. In Q2 2026, Sky Protocol recorded $107.35 million in Gross Protocol Revenue, and $53.91 million went to sUSDS holders through the SSR.

sUSDS supply closed the quarter at $5.52 billion, up 149% year over year. It’s the world’s largest yield-generating stablecoin.

• Track record: the core protocol has run since 2017 without an exploit, from a team that’s spent close to a decade on stablecoin infrastructure.
• Rated: in 2025, S&P Global gave Sky Protocol the first credit rating ever assigned to a DeFi protocol.
• Risk discipline: Sky Frontier Foundation’s risk frameworks, among the most systematic in the space, tie each Agent’s access to capital to its performance.
• Built for steadiness: the SSR is engineered for strong risk-adjusted yield, not the biggest number on a leaderboard.

Risks still apply. Smart contracts can fail, and the rate can change. Need USDS first? You can convert 1:1 through the Peg Stability Module.

Lifeline or Liability? What Stablecoin Growth Means for US Debt Markets

Honestly? It depends who you ask.

The case for:

• A steady buyer at the short end can make government borrowing a bit cheaper.
• Dollar stablecoins spread dollar use abroad, and much of that demand lands in T-bills.

The case against:

• Money leaving bank deposits for stablecoins could mean less bank lending.
• A fast wave of redemptions could force T-bill sales right when markets are stressed.
• Leaning harder on short-term bills raises refinancing risk when rates jump.

Both sides can be right at once. That’s why what makes a stablecoin reliable matters more as this market grows: backing, liquidity, exits and transparency.

Stablecoins won’t fix a $40 trillion debt. They’re becoming part of how it gets financed, though.

Sky’s also working on ways to make capital movement data speak the same language in real time, whatever the vehicle or currency.

Quick Answers on Stablecoins and US Debt

Q. Do stablecoins hold US debt?

A. Yes. Reserve-backed stablecoins keep most reserves in short-term Treasury bills, repos and cash, making issuers a growing buyer of US debt.

Q. Does sUSDS pay T-bill yield?

A. No. sUSDS accrues the Sky Savings Rate, a variable rate funded from aggregate Sky protocol surplus. Treasury bills are one of several strategies feeding that revenue.

Could stablecoins hurt the Treasury market?

In a big run, possibly. Outflows move T-bill yields more than inflows do. That’s why reserves you can verify matter.

. . .

Over to you. Should stablecoin holders get a share of the yield their dollars generate, or is 0% a fair price for a stable dollar? Tell me in the comments. I read every one.

If you’d rather your stablecoins didn’t sit still, put them to work with sUSDS.

Stablecoins and US Debt: Who Keeps the T-Bill Yield? was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.

Twenty-one of the largest financial institutions on earth spent a year quietly building a dollar together.

Bank of America. Citi. Goldman Sachs. Deutsche Bank. UBS. Santander. MUFG. Wells Fargo. Fidelity.

They announced the plan on 1 September 2026. They are targeting the first half of 2027.

What they do not have yet: a name, a CEO, or a chosen blockchain.

What they will never have: permission to pay you anything for holding it.

That last part is not an oversight. It is written into the law the entire launch timeline was built around.

What Exactly Is the 21-Bank Stablecoin Launching in 2027?

The group started small. In October 2025, ten globally systemic banks opened an exploratory phase to test whether a bank-issued stablecoin was even workable.

By September 2026 those ten had become twenty-one, spanning five regions.

The plan, stripped to its parts:

• Form a jointly owned company in the second half of 2026
• Issue a 1:1 reserve-backed US dollar token on public blockchains
• Target market launch in the first half of 2027
• Follow with a euro token, then other G7 currencies
• Build to GENIUS Act standards, and to MiCA where it applies

The use case they keep naming is settlement. Cross-border commercial payments. Digital asset settlement. Moving value between counterparties faster than existing rails allow.

It is a serious project with serious balance sheets behind it.

It is also, by construction, a payment instrument and nothing more.

Why the GENIUS Act Makes Bank Stablecoin Yield Illegal

Here is the clause that shapes everything.

Section 4(a)(11) of the GENIUS Act says no permitted payment stablecoin issuer shall pay the holder of any payment stablecoin any form of interest or yield, whether in cash, tokens, or other consideration, solely in connection with holding, using, or retaining it.

The law was signed in July 2025. It takes effect on the earlier of two dates: 18 January 2027, or 120 days after federal regulators finalize implementing rules.

Now look at the consortium’s launch window again. First half of 2027.

The timing is not a coincidence. These banks are launching into the regime, not around it.

Europe reached the same destination by a different road. MiCA has banned interest on e-money tokens since 2024, and extends the ban to service providers as well as issuers.

So the two largest regulated markets on earth now agree on one thing. A payment stablecoin is a settlement instrument. It is not a savings one.

So Who Actually Keeps the Reserve Income on a Bank Stablecoin?

Every fully reserved dollar token sits on a pile of short-dated government paper and bank balances. That pile earns.

Under the yield prohibition, those earnings do not disappear. They simply stop at the issuer.

The split looks like this:

• You hold the token. You get settlement, transferability, and a stable unit of account.
• The issuer holds the reserves. The issuer keeps the income those reserves throw off.

For a payments business, that is a rational model. Float funds the operation and keeps transaction fees low.

For anyone parking a meaningful balance for longer than a few days, it is a quiet transfer. Your capital is working. The return lands somewhere else.

That is the missing piece. Not a missing feature. A missing claim.

The yield does not vanish in 2027. It just stops one step before you.

Is the Missing Piece a Bug, or the Entire Business Model?

This is where it gets contested, and where the numbers get loud.

A US Treasury advisory council flagged $6.6 trillion of transactional deposits as at risk from stablecoins.

Bank of America’s chief executive has put his own estimate near $6 trillion, suggesting 30 to 35 percent of transactional deposits could move if yield-bearing digital dollars became normal.

A December 2025 Federal Reserve staff note modeled a $1 trillion deposit drain with no recycling back into banks. Its estimate for the hit to bank lending ran from $600 billion to $1.26 trillion.

The American Bankers Association has argued that permitting yield could take the stablecoin market from roughly $300 billion to as much as $2 trillion.

Read those three numbers again.

None of them is an argument that yield is unimportant. Every one of them is an estimate of exactly how important it is.

Through 2026 the banking lobby pushed to close what it calls the stablecoin yield loophole in follow-on legislation.

At the same time, a coalition of major banks committed to a shared tokenized deposit network targeting that same first half of 2027.

Tokenized deposits sit outside the payment stablecoin definition. They can pay interest.

Both facts can be true at once. The question is which product you are actually being offered.

How a Yield-Bearing Stablecoin Is Built Differently

There is a second architecture, and it predates the 2027 cohort by years.

Instead of asking one token to do two jobs badly, it separates them.

USDS is the payment unit. Fully backed, soft-pegged to the dollar, convertible at 1:1 through Peg Stability Modules with no fees and no slippage, live across Ethereum, Solana, Base and Arbitrum.

sUSDS is the savings position. Supply USDS, receive sUSDS, and the position accrues value programmatically. No lockups. No minimum. Convert back on demand.

What sUSDS holders access is the Sky Savings Rate, a variable rate set by Sky Governance.

Here is the part that actually matters, because it is where the two models diverge.

The Sky Savings Rate is not reserve float being handed back out of goodwill. It is funded from Sky Protocol revenue generated by the Sky Agent Network: independent allocators including Spark, Grove, Keel, Obex and Osero that borrow USDS liquidity under risk parameters set in public and deploy it across diversified strategies. They pay for that access.

The income contributes to protocol revenue. Governance allocates a portion of it to the rate.

The holder’s claim is structural. Not discretionary.

One model asks who keeps the float. The other asks who is allowed to claim it.

What the Numbers Behind the Sky Savings Rate Actually Look Like

Describing a mechanism is cheap. Here is the audit trail.

• Sky Protocol generated $123.79M in Gross Protocol Revenue in Q1 2026, a record quarter, followed by $107.35M in Q2 2026
• Q2 2026 produced $33.29M in Net Protocol Surplus, the fifth consecutive positive quarter
• Protocol Collateral rose 45.5% year over year to $12.32B at the close of Q2, and the live counters on skyeco.com read higher today
• sUSDS grew 149% year over year to $5.52B over the same period
• In Q1 2026 alone, sUSDS added more than $2.5B in market capitalization, more new capital than the next four yield-bearing stablecoins combined, according to third-party research
• On 7 October 2026, Moody’s assigned Sky Protocol a B3 counterparty risk rating with a stable outlook, its first rating of any stablecoin protocol, following an affirmed B- from S&P Global (details here)

Every figure above is checkable on the public financial dashboard, updated continuously rather than attested monthly.

That is what an open balance sheet buys you. You do not have to take the claim on faith.

Five Questions to Ask Any 2027 Dollar Token Before You Hold It

Before you park a balance in anything launching next year, ask:

• Who receives the income the reserves generate, and is that written down anywhere you can read?
• Is the rate set in public, or at an issuer’s discretion behind a closed door?
• Can you verify the backing yourself, in real time, or only through a monthly attestation?
• Who controls the balance: you, or a counterparty’s ledger entry?
• If you want a return, must you leave the instrument entirely to get one?

For the 2027 bank cohort, question one already has a legally mandated answer. It is not you.

For a protocol model like Sky Protocol, the answers sit onchain under parameters set through Sky Governance. That is a different kind of promise.

Verifiable rather than guaranteed. Variable rather than fixed. And exposed to real risks, including smart contract vulnerabilities and market volatility.

The Argument Worth Having

2027 will be the year the digital dollar gets institutional distribution. Twenty-one banks, five regions, bank-grade governance from the first day. That is good for the category and probably overdue.

It will also be the year a lot of people discover that the most heavily regulated digital dollar in their wallet pays them nothing by design, while the reserves underneath it earn quietly for somebody else.

So here is the argument worth having in the comments.

Is the yield ban protecting ordinary holders from confusing a payment token with a savings product? Or is it protecting an interest margin that was never going to survive transparent, programmable pricing?

I lean toward the second. Tell me where that reasoning breaks.

Every Bank Stablecoin Launching in 2027 Has the Same Missing Piece was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.

Seven months ago, Stripe’s stablecoin cards worked in 18 countries. By the end of this year, the company expects them in more than 100.

That is not a pilot getting bigger. That is a payment rail being switched on.

Which raises a question almost none of the coverage asked. While those digital dollars sit on a card waiting to be spent, who is earning on them?

Short answer: not you. Not legally, anyway.

That single sentence is the most interesting thing in this story. Stripe did nothing wrong to cause it.

The rule that creates the gap was written by Congress, and it binds every regulated issuer on the planet.

What exactly did Stripe announce about its stablecoin cards?

Stripe told CoinDesk on 5 October 2026 that it expects its stablecoin card programs to reach more than 100 countries by the end of 2026.

The detail underneath the headline:

• The programme started with Visa in 18 nations, mostly in Latin America, with Europe, Africa, the Middle East and Asia Pacific next.
• It runs on Bridge, the stablecoin infrastructure company Stripe bought for $1.1 billion in 2024.
• Stripe has issued more than 400 million cards since 2018. That footprint is what makes 100 countries plausible rather than aspirational.
• Henri Stern, co-founder of Privy, now oversees stablecoins and crypto across all of Stripe.
• Current customers include Kraken, Ramp and the payments app Morse.

One caveat worth holding onto. More than 100 countries is a year-end coverage target, not a list of markets live today.

Why is stablecoin card spending tripling in twelve months?

Because people are actually spending with them.

Roughly $1.2 billion moved through stablecoin-linked cards in September 2026, about three times the figure a year earlier. That sits inside a digital dollar market now worth more than $300 billion.

Cards are the boring part of crypto, and boring is exactly the point. Nobody has to learn a new habit. You tap, the merchant gets paid in local currency, and the stablecoin settles behind the glass.

You can see why this lands hardest in the markets Stripe started with. If your salary arrives in a currency that loses value while you sleep, a dollar balance you can spend at the corner shop is not a crypto product. It is a defence.

But a card is a spending instrument. It is built for the second money leaves. It says nothing about the far longer stretch when money simply sits.

Payments solved how a dollar moves. They never answered what the dollar earns while it waits.

Why can a stablecoin card not pay you yield?

Because of one line in US law.

Section 4(a)(11) of the GENIUS Act bars a permitted payment stablecoin issuer from paying the holder any form of interest or yield, in cash, tokens or other consideration, solely in connection with holding the coin. Then it got tighter.

In February 2026 the OCC published a 376-page proposed rule adding a rebuttable presumption: if an issuer pays an affiliate or related third party, and that party pays yield to holders, the issuer is presumed to have breached the ban.

Comments closed on 1 May 2026, and the pushback from exchanges and wallet firms was loud. The direction of travel is not subtle.

What that means in practice:

• The reserves behind a payment stablecoin earn short-term rates.
• That income stays with the issuer and its distribution partners.
• The person holding the coin gets utility, not a rate.

So the more successfully Stripe spreads digital dollars across 100 countries, the more dollars sit in wallets earning their holders nothing. The rail gets better. The float question gets bigger.

Where does stablecoin yield actually come from, then?

From somewhere that is not the issuer.

This is the gap Sky Ecosystem was built around. It is not a payments company and it is not trying to be one.

It is a savings and capital allocation network, and its product is the answer to what capital does between transactions.

The mechanism is plainer than the category usually sounds:

• USDS liquidity is made available by Sky Protocol under risk parameters set in public.
• The Sky Agent Network, independent allocators including Spark, Grove and Osero, draws on that liquidity and competes for it.
• Those allocators deploy into diversified strategies across credit, lending markets and short-term government debt.
• The fees they pay for access flow back to Sky Protocol as revenue.
• Sky Governance sets the Sky Savings Rate from that revenue, and it accrues to holders of sUSDS.

Five steps, every one of them visible onchain. The rate is variable and governance-set. It is not fixed, and it is not a promise.

Is this different from an exchange paying rewards on your balance?

Structurally, yes, and the difference is the whole point of the OCC fight.

• Exchange rewards are usually a share of issuer reserve income, routed through a partner. That is the arrangement the proposed rule targets.
• The Sky Savings Rate is not reserve income. It comes from fees paid by independent allocators for access to protocol liquidity.
• Rewards sit with a company that holds your balance. sUSDS is non-custodial. You hold it, you redeem it, nobody needs to approve the exit.

Different source, different custody, different counterparty. Worth knowing which one you are actually holding.

What is sUSDS, and why does it keep taking share?

sUSDS is the yield-generating version of USDS. You convert, the balance compounds at the Sky Savings Rate, and you convert back whenever you want with no lock-up and no exit fee. You keep the keys the whole time.

The adoption pattern is the interesting part.

• Yield-bearing tokens drove more than half of all net stablecoin supply growth in Q1 2026.
• sUSDS added more than $2.5 billion in that quarter alone, more new capital than the next four yield-bearing stablecoins combined.
• As of mid-2026, only a few billion dollars of a $300 billion-plus stablecoin market was classified as yield-bearing at all.

Read that last point twice. The overwhelming majority of digital dollars are still parked and doing nothing for the person holding them. Stripe is about to add a hundred countries of new parking.

Can anyone actually verify any of this?

Fair question, and it is the right one to ask of anything paying a rate.

• On 7 October 2026, Moody’s assigned Sky Protocol a B3 counterparty risk rating with a stable outlook, its first rating of any stablecoin protocol.
• S&P Global had already affirmed a B- rating. Sky Protocol is currently the only stablecoin protocol carrying both.
• Sky Protocol generated $123.79 million in Gross Protocol Revenue in Q1 2026 and $107.35 million in Q2, with a fifth consecutive quarter of Protocol Surplus.
• In September 2026, Galaxy added $100 million of sUSDS to its corporate balance sheet and approved it as collateral across its institutional trading business.
• Protocol Collateral, stablecoin supply and the live rate sit on public dashboards. Nothing above asks you to take anyone’s word for it.

Neither rating is investment grade, and Moody’s named the level of Sky Reserves as a central factor.

That is in the public report too, which is rather the point of having one. A yield you cannot audit is not a yield. It is a hope.

What the next 100 countries actually change

Stripe is doing the hard, unglamorous work of making digital dollars spendable where card infrastructure took decades to arrive.

That is genuinely good, and it will pull in people who have never typed the word blockchain.

Those people are going to end up holding dollar balances. Many of them for weeks at a time.

The payment layer has an excellent answer for the second the money moves.

It has no answer, and under current law cannot have one, for the long stretch when the money waits.

That is a separate job. It needs a separate rail, run in public, with the numbers on a dashboard anyone can open.

The rail that moves your money and the rail that pays you for parking it were never going to be the same rail. The law made sure of it.

Stripe just made the first one global. The second one is still the open question.

If a hundred countries get stablecoin cards next year, should the dollars sitting on them earn something, or is a payment rail entitled to keep the float?

I lean one way on this. Tell me where I am wrong in the comments.

Stripe Is Taking Stablecoin Cards to 100 Countries. It Had 18. was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.

In June, a market opened that let people lock a rate on sUSDS until a named date in November.

It passed $50 million in under two weeks.

At the time, that single market was roughly 4% of all the capital sitting on the venue hosting it.

Nobody modelled it moving that fast. Including, I suspect, the people who built it.

What Is Fixed Rate DeFi Yield, and Why Did 2026 Change Everything?

Fixed rate DeFi yield means you know your number before you commit.

You buy an instrument below par today, hold it to a stated maturity date, and redeem at par.

The gap is your return. No utilisation curve. No rate change halfway through. No surprise in month four.

That is ordinary in bonds. Onchain, it was mostly theory until this year.

Then it stopped being theory. Between May and September 2026, four separate protocols shipped fixed-rate, fixed-term lending: Morpho Midnight, Jupiter Offerbook, Kamino Fixed Rate Vaults and Nolus.

CoinDesk described the shift as DeFi quietly rebuilding the fixed-income stack for institutional capital.

Quietly is the right word. Most of crypto was looking somewhere else, at tokens and narratives, while the least exciting instrument in finance finally turned up onchain.

Why Did Fixed Rate Stablecoin Yield Suddenly Find Buyers?

Because the buyers changed.

Yield farmers do not want certainty. They want the highest number this week. Treasury teams want the opposite.

They want a figure they can put in a board pack in June and still be right about in November.

Three things lined up this year:

• Yield-bearing stablecoins grew from roughly $1.5 billion in early 2024 to about $20 billion by mid-2026, around 7% of the stablecoin market and the fastest-growing segment in it.
• Onchain real-world assets excluding stablecoins passed $33 billion in July 2026, close to four times their early-2025 level.
• Corporate balance sheets started holding onchain savings tokens directly rather than through a fund wrapper.

Once capital has a horizon, it needs a rate with a horizon. Predictability, not maximisation.

That is the whole story, and it is a very old story. Money markets did this in the 1960s. Crypto just took eighteen years to get there.

How Does a Variable Savings Rate Become a Fixed One?

You split it in two.

Supply USDS and you receive sUSDS, which accrues the Sky Savings Rate. That rate is variable and set by Sky Governance. It moves with protocol revenue and market conditions.

A yield-tokenisation venue then separates that position into two tradeable pieces:

• A principal token, which behaves like a zero-coupon note and redeems at full value on its maturity date. Buy it at a discount and the discount is your fixed rate.
• A yield token, which carries the floating stream and all the upside or downside in it.

One buyer locks a rate. Another takes a view on where rates go. Both trade freely against each other.

Here is the part people skip. You cannot strip a floating rate into fixed and floating legs unless the floating rate is credible, continuous and published.

A term market is always downstream of a reference rate. Which means the interesting question is not who built the fixed-rate product. It is whose rate was trusted enough to build one on.

What Did the sUSDS Fixed-Yield Market Actually Do?

It grew faster than the venue hosting it.

Sky Fixed Yield went live in June 2026 as a third-party integration using Pendle, with the Sky Savings Rate as the underlying yield source.

The opening fixed rate was 3.62%. Within two weeks the sUSDS pool had crossed $50 million.

By late August, the November maturity held around $82 million and was quoting roughly 4.81% fixed while the Sky Savings Rate sat at 3.52%.

That spread of about 129 basis points is the price of giving up flexibility. It is also a signal.

A positive spread means buyers expected the variable rate to sit lower, on average, between now and the maturity date.

That is a market forming an opinion about future rates, which is something onchain dollars simply could not do two years ago.

By late September, the 26 November sUSDS market was being published as one of the highest stablecoin fixed rates onchain, near 4.83%.

Worth being precise. Sky Protocol does not set, control or guarantee that fixed rate. The market sets it. And the rate is only fixed if the position is held to maturity.

Is the Sky Savings Rate Becoming the Reference Rate for Onchain Dollars?

The behaviour suggests yes.

Nobody builds a term curve on a rate they do not trust. They build it on the one everyone already prices against.

Look at where the capital went. sUSDS supply reached $5.52 billion at the end of Q2 2026, up 149% year on year.

In Q1 2026 alone, sUSDS added more than $2.5 billion in market capitalisation.

That was more new capital than the next four yield-bearing tokens combined, in a quarter when the broader crypto market fell more than 20%.

The engine behind it is not a promise. Sky Protocol recorded $107.35 million in Gross Protocol Revenue and a $33.29 million Protocol Surplus in Q2 2026, its fifth consecutive quarter in surplus.

Protocol Collateral currently stands at $14.15 billion against $11.48 billion of stablecoin supply.

A reference rate needs a revenue engine underneath it. This one has one, and every line of it is public.

Why Are Institutions Holding sUSDS Instead of Idle Dollars?

Because one balance can now do two jobs.

In September 2026, Galaxy Digital added $100 million of sUSDS to its corporate treasury and approved it as eligible collateral across an institutional trading business carrying a $1.4 billion average loan book and more than 1,600 counterparties.

Clients who post sUSDS against a loan keep accruing the Sky Savings Rate on the full position for as long as that loan runs.

Read that twice. The collateral keeps working while it is pledged. In most financing arrangements, posted collateral goes quiet.

That relationship did not start there either. Grove, a Prime Agent in the Sky Agent Network, provided Galaxy with a $500 million warehouse lending facility first. Galaxy later borrowed on Spark, a second Prime Agent.

Sky Protocol entered Q3 2026 with $5.41 billion supplied through independent allocators and into institutional tokenised funds, including anchor positions in BlackRock’s BUIDL and Janus Henderson’s JTRSY.

What Are the Real Risks of Locking an Onchain Rate?

Five, stated plainly.

• The rate is fixed only to maturity. Sell early and the market price decides your return, which may be above or below your entry.
• Liquidity is thinner than headline numbers suggest. That $82 million market showed only around $3.47 million of pool liquidity. Size a position around the exit, not the headline.
• You take on an additional smart contract. The term venue is a separate protocol with its own risk surface.
• You accept duration. If the floating rate climbs past your fixed rate, you left yield on the table.
• The Sky Savings Rate itself is variable and governance-set. It is not a promise and has never been presented as one.

None of that is an argument against fixed rates. It is an argument for reading the maturity date before you read the APY.

What Happens Next in Onchain Fixed Income?

Scale, slowly, and from a very low base.

Fixed-rate-specific lending onchain is still around $130 million, roughly 0.24% of the $54 billion DeFi lending market. Traditional fixed income is about $143 trillion. The gap is almost comic.

But the plumbing already works. Pendle has settled $69.8 billion in cumulative fixed yield. The rails exist, the maturities trade, and the curve is thin but real.

Treasury desks do not need a deep curve to start. They need one date they can plan around.

What was missing was a dollar rate worth building a curve on. Something continuous, revenue-backed, published and governed in public.

That part arrived in 2026. The demand surprised everyone because it was never really about chasing a higher number. It was about certainty showing up in a market that had never been able to offer it.

Every figure here can be checked. Sky Protocol publishes its balance sheet, revenue and collateral openly, and the full product documentation sits on Sky Ecosystem Insights.

One question I keep turning over, and I would like your answer in the comments: if a credible onchain dollar rate now exists with a tradeable term structure around it, what is the actual argument left for keeping treasury dollars idle?

Fixed-Rate Onchain Dollars Arrived in 2026. The Demand Surprised Everyone. was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.

In August 2026, a crypto exchange wrote a $100 million cheque to buy GPUs.

Not a token swap. Not a treasury trade. A loan, secured against physical hardware humming inside a data centre.

Bullish provided that facility to USD.AI. By the end of that month, USD.AI reported $281.8 million of lifetime capital deployed into compute financing.

Read that again. Stablecoin liquidity is now buying NVIDIA chips.

This is the story most of crypto missed in 2026. AI compute financing stopped being a venture capital question and became a credit question. And credit, unlike equity, needs a lender.

So who is lending? And what does any of it have to do with the dollar sitting in your wallet?

More than you would think. Follow the money far enough and it ends up somewhere boring, like a savings rate.

What is AI compute financing, and why is it suddenly everywhere?

Short answer: it is debt raised against AI hardware and the contracted revenue that hardware produces. GPUs, server racks, cooling, power interconnects.

The reason it matters is arithmetic.

• Morgan Stanley Research estimates roughly $2.9 trillion in global data centre capital expenditure between 2025 and 2028.
• Around $1.5 trillion of that cannot be covered by corporate cash flow. That is the financing gap.
• Roughly $800 billion of the gap is expected to be filled by asset-based finance, meaning debt secured on the equipment itself.
• J.P. Morgan estimates hyperscaler capex alone reaches $697 billion in 2026.

Equity cannot absorb a number that size. Debt has to. And once debt shows up, somebody has to underwrite the collateral.

For context, the entire US subprime mortgage securitisation market in 2007 was around $2 trillion. We are building something of comparable scale, in public, in about four years.

How did a GPU become bankable collateral?

Three things changed, and all three are unglamorous.

• It produces cash. A GPU cluster under contract generates metered revenue, the way a toll road or a cell tower does.
• It can be verified. Lenders now require installation and independent verification before capital is released.
• It can be insured and repossessed. Hardware sitting in insured facilities behaves like equipment finance, not a punt.

The moment an asset throws off contracted cash and can be seized, bankers stop calling it speculation and start calling it asset-based lending.

The moment an asset throws off contracted cash and can be seized, bankers stop calling it speculation and start calling it asset-based lending.

The 2026 deal flow makes the point better than I can.

• Nscale signed a $1.4 billion GPU-backed term loan led by PIMCO, Blue Owl and LuminArx.
• IREN closed $3.65 billion tied to a Microsoft AI Cloud contract.
• Lambda raised a $1.0 billion first-lien facility.
• Fluidstack won approval to borrow more than $10 billion against its NVIDIA holdings.

Roughly $500 billion of debt now finances the AI build-out across more than 200 separate instruments. Most of it did not exist eighteen months ago.

Why is stablecoin liquidity showing up in AI infrastructure credit?

Because the settlement layer finally fits the asset.

Compute loans amortise monthly. They draw in tranches. They need capital available on a Tuesday afternoon, not after a two-week wire cycle.

That is precisely what onchain capital markets are good at.

• USD.AI funded a $98.1 million, three-year facility for 2,304 NVIDIA B300 GPUs managed by Hydra Host.
• It approved a facility of up to $500 million for Sharon AI across Australia and Asia-Pacific.
• Crucible Capital drew $26.82 million against 576 B300 units deployed in Washington State.

These are non-recourse loans, secured against hardware rather than a borrower’s balance sheet. Risk sits at the infrastructure level, not the corporate one.

One detail I keep coming back to: in the QumulusAI deployment, the hardware was installed and independently verified before any capital was released. That is an underwriting standard, not a crypto one.

Boring infrastructure lending wearing a crypto jacket. That is exactly why it works.

Where does Sky Ecosystem fit into onchain AI compute financing?

Here is the honest answer, which is more interesting than the hype version.

Sky Ecosystem does not originate GPU loans. It is a global savings and capital allocation network. It supplies the liquidity rail that independent allocators borrow from.

The structure runs in a loop:

• Sky Protocol makes USDS liquidity available under governance-set risk parameters.
• Independent capital allocators in the Sky Agent Network borrow that liquidity and deploy it.
• Returns flow back to Sky Protocol.
• Aggregate protocol revenue funds the Sky Savings Rate.
• sUSDS holders access that rate, auto-compounding, with no lockups.

In March 2026, Obex, the incubator agent, deployed $1 billion of USDS across a cohort of eight capital allocators.

The sectors covered mortgages, energy infrastructure, structured credit, tokenisation, and GPU and AI infrastructure lending.

USD.AI sits in that cohort. Worth being precise: cohort members are independent capital allocators, not Sky Agents. Obex is the connector, not the controller.

Institutions have noticed the rail. In September 2026, Galaxy added $100 million of sUSDS to its corporate treasury and approved it as eligible collateral across its institutional trading business.

By the close of Q2, Prime Agent Vaults held $6.84 billion, with roughly $2.58 billion deployed across counterparties including BlackRock, Janus Henderson, Anchorage, PayPal, Securitize and Galaxy.

None of those names showed up for a yield farm. They showed up for plumbing that reports quarterly and gets audited.

None of those names showed up for a yield farm. They showed up for plumbing that reports quarterly and gets audited.

Does the Sky Savings Rate depend on AI compute loans?

No. And that is the whole point.

sUSDS does not hold GPU paper. The Sky Savings Rate is funded from aggregate Sky Protocol revenue across diversified strategies, then set by SKY token holder governance.

AI compute financing adds one more real-economy category to a collateral base that already spans treasuries, tokenised credit and onchain lending. It widens the funnel. It is not a bet.

The published numbers, from Sky Frontier Foundation:

• Q2 2026 Gross Protocol Revenue of $107.35 million, up 10.5% year over year.
• Net Protocol Surplus of $33.29 million, the fifth consecutive quarter in surplus.
• sUSDS supply up 149% year over year, closing Q2 at $5.52 billion.
• Protocol Collateral now standing at $14.15 billion, against $11.48 billion in stablecoin supply.
• An annualised Gross Protocol Revenue run-rate of $429.4 million entering the second half.

sUSDS is the largest rate-bearing stablecoin in the world by supply. The rate is variable. Nobody guarantees it, and anybody who tells you otherwise is selling something.

What could go wrong when the collateral is compute?

Four questions worth asking of any yield that touches this sector.

• Residual value. What is a 2026 GPU worth in 2029? Amortisation schedules assume an answer nobody actually has.
• Offtaker quality. The loan is only as good as the counterparty paying for the compute. The contract repays it, not the silicon.
• Concentration. One chip generation, one neocloud, one hyperscaler customer. Correlated exposure hides inside diversified-sounding labels.
• The price of risk. CoreWeave’s GPU-backed facilities carried rates of 10.5% and 14.1%. Spreads that wide are telling you something.

Sky’s answer here is structural rather than promissory. Exposure limits are set by governance and published onchain.

Collateral composition is verifiable by anyone with a browser. You are not asked to trust a press release.

That is a different kind of comfort than a big number on a landing page. It is also the only kind that survives a bad quarter.

What should you actually watch from here?

• Whether more capital allocators bring compute-backed loan books onchain, and at what size.
• Whether residual value assumptions hold through the next chip generation.
• Whether governance expands or caps exposure to the category.
• Whether yields across the sector hold when the first offtaker misses a payment.

Diversification is not a marketing word in this context. It is the only thing standing between a saver and one sector’s bad year.

My own view: compute credit is real and it is here to stay, but the category is eighteen months old and has not seen a downturn. Treat it as one slice of a portfolio, never the thesis.

The bottom line on AI compute financing

Compute became collateral in 2026. That part is settled.

What is not settled is who funds it, on what terms, and whether the people holding the other end of that trade understand what they own.

Onchain credit rails are now a real answer to a real $1.5 trillion problem. Sky Protocol supplies the liquidity.

Independent allocators do the underwriting. Governance sets the limits. The Sky Savings Rate is what falls out the other side, funded by everything in the portfolio rather than any single sector in it.

Compute became collateral in 2026. What is not settled is who funds it, and whether the people on the other end of that trade understand what they own.

Now I want your take.

Would you be comfortable with AI infrastructure credit inside the collateral backing a dollar you hold? Or does compute belong nowhere near a savings rate?

Tell me in the comments. I read all of them, and I will argue with the good ones.

AI Compute Financing Is Quietly Becoming an Onchain Yield Source was originally published in StartupInsider on Medium, where people are continuing the conversation by highlighting and responding to this story.