The traditional ratings establishment has arrived onchain. In October 2025, S&P Global Ratings began delivering its Stablecoin Stability Assessments directly into DeFi protocols and smart contracts through Chainlink, the first time institutional stablecoin risk analysis was readable by onchain applications in real time. In June 2026, Moody's embedded live credit ratings into tokenized securities on Solana through a partnership with Alphaledger, after finalizing a dedicated stablecoin methodology earlier in the year and standing up the first rating-agency node on the Canton Network. Two of the three names that have defined creditworthiness for a century now publish where the assets live.

A hundred-year-old trust signal crossing onto public chains is institutional validation of the tokenized economy. Allocators building on these networks now get standardized, independent analysis they can reference programmatically instead of pulling from a terminal.

But we want to be precise about what actually crossed over, because the framing that is settling into the market conflates two different things. The agencies did not build onchain-native ratings. They made offchain ratings readable onchain. Where a rating lives matters far less than how it is produced, and on the how, almost nothing changed. Here is where Webacy innovates: what the agencies are rating is not what is actually happening onchain, and the tokenized economy needs a solution.

What the agencies are actually rating

S&P's Stablecoin Stability Assessment scores a coin's ability to hold its peg on a scale of 1, very strong, to 5, weak. It weighs asset quality, including credit, market-value, and custody risk, alongside governance, redeemability, liquidity, and track record. S&P is careful to say the SSA is not a credit rating; it is a peg-stability opinion. The firm currently assesses roughly ten stablecoins this way, including USDT, USDC, and Sky's USDS, and separately it has begun rating DeFi protocols outright, assigning Sky Protocol a B- in August 2025 while citing depositor concentration, centralized governance, and weak capitalization.

Moody's approach centers on the credit quality of the reserve assets behind a stablecoin, plus market-value risk, liquidity, operational resilience, and technology risk. A coin qualifies only if its reserves are effectively segregated from the issuer's balance sheet so holders can be made whole in bankruptcy. Algorithmic and synthetic designs, the ones without direct fiat collateral, sit outside the methodology entirely.

Both are well-built frameworks. Both share three characteristics that matter enormously once you put them next to a live chain. They are periodic, produced and refreshed on an analyst review cycle. They are issuer-informed, built substantially from data the rated entity supplies or discloses. And they are backward-looking by construction, weighing reserve composition, track record, and governance as they stood at the last assessment.

The stablecoin market roughly doubled in a year to more than $300 billion, and Moody's own 2026 outlook describes stablecoins as core market plumbing settling trillions of dollars and evolving into digital cash for liquidity, collateral, and settlement. When the institutions that name that shift also decide their assessments need to live where the assets live, they are conceding something important: the terminal is no longer where risk gets consumed. The chain is. The open question is whether a signal built for the terminal era is the right one to carry across unchanged.

Delivering those assessments onchain changes their distribution, not their nature. When S&P's assessment becomes automatically referenceable inside a smart contract, what you get is real-time access to a periodic opinion. The access is instant. The opinion still moves at the speed of an analyst committee. The blockchain, in reality, moves much faster.

What is actually happening onchain

Onchain assets do not fail on the reserve-composition axis a periodic rating is built to watch, or at least not only there. They fail structurally and behaviorally, in hours or days, through connections a snapshot cannot see.

Consider Stream Finance in early November 2025. On October 30, Stream published a proof of assets of roughly $400 million. Days later, an external fund manager disclosed a $93 million loss, and the structure came apart. Stream had been running recursive leverage of more than four times, looping around $160 million of deposits into a claimed half-billion in assets, an accounting mirage that held only until someone looked closely. Its xUSD stablecoin lost its peg and fell toward pennies. Because several lending markets had hardcoded xUSD's oracle price at one dollar to prevent cascading liquidations, the illusion of stability persisted onchain even as the backing evaporated.

Then it spread. Elixir's deUSD had parked roughly 65% of its collateral in Stream through private Morpho vaults. When xUSD collapsed, deUSD's backing went with it, and the synthetic dollar fell about 98%, from a dollar to a cent and a half, before Elixir wound it down. The contagion continued into lending protocols including Euler, Morpho, and Silo, and Compound paused certain markets to contain it.

Now ask what a periodic, issuer-informed, reserve-focused rating would have made of that on November 3. The honest answer is that it would very likely have rated the structure as sound. The proof of assets was fresh. The reserves, as disclosed, looked adequate. Every fact that actually mattered, the recursive leverage, the offchain fund-manager exposure, the concentrated cross-protocol dependency, the oracle hardcoding, was either invisible to a reserve snapshot or actively obscured by one. And deUSD, as a synthetic design, would have fallen outside a reserve-collateral methodology's scope in the first place. The assets most prone to this kind of failure are frequently the ones a traditional framework excludes.

USR made the same point at a different tempo. On March 22, 2026, Resolv's USR lost 94% of its value in under an hour after an unbacked minting exploit. No review cycle convenes fast enough to matter against a clock like that.

These are not exotic edge cases. They are the representative failures of onchain finance, and they share a pattern that periodic reserve analysis is structurally blind to: the break is a live, observable change in structure and connection, not a slow drift in reserve quality.

Onchain-native is a method

If a rating is going to protect anyone holding an asset that settles onchain, it has to be built the way onchain assets behave. Three properties are non-negotiable, and none of them is satisfied by moving an existing assessment onto a chain.

It has to be continuous. Onchain state changes in minutes, not quarters. An address that was clean this morning receives contaminated funds this afternoon. A redemption queue open at the last review closes overnight. A peg holds all quarter and breaks in fifty minutes. A rating that refreshes slower than the asset it judges is a photograph of a moving object, however quickly that photograph can be fetched. Ratings in this environment cannot be an event. They have to be a continuous signal.

It has to be explainable at the level of a condition. A single grade or a bare score is not actionable when an allocator has to decide, right now, whether to hold or exit. Every rating should trace to the specific structural condition driving it: redemptions are closed, collateral concentration crossed a threshold, an upgrade key is unguarded, holder concentration spiked, a contamination path opened to a flagged cluster. Explainability is what turns a rating from a verdict into a tool, and it is what lets a smart contract act on the rating, which was the entire promise of putting risk onchain.

It has to read the chain independently. Periodic frameworks lean on issuer disclosure, and the most dangerous conditions are usually the ones an issuer would not volunteer. Stream published its own proof of assets. An independent, onchain-native rating does not wait to be told; it reads the position graph directly, including the recursive leverage and the dependency concentration that no attestation mentions. Independence has always been the hardest problem in ratings, and delivery onchain does nothing to solve it.

Where a rating is stored is a distribution decision. How often it updates, what data it is built from, and whether it can see structure the issuer omits are the decisions that determine whether it protects anyone. The agencies solved distribution. The method is the part still open.

What this looks like at the point of decision

The abstraction becomes concrete the moment an allocator has to act. A curator deciding whether to route capital into a vault, a treasury deciding whether to hold a stablecoin through the weekend, a protocol deciding whether to accept an asset as collateral: each is a decision made against live conditions, not against last quarter's assessment.

A condition-level rating is built for that moment. Our Vault Technical Risk Rating, as one example, is a composite score from zero to one hundred assembled from roughly twenty sub-scores spanning contract structure, upgrade and admin control, collateral and redemption mechanics, and behavioral signals. The composite is not the whole story. Hard floor overrides sit on top of it, so a single critical condition caps the score no matter how strong the rest looks. If redemptions are closed, the rating is floored, because an asset you cannot exit is not healthy regardless of how clean its code reads. A blended average can quietly average away the one condition that matters. A floor cannot.

The same logic runs across the surfaces we watch: stablecoin depeg dynamics, contamination propagation across connected assets, smart-contract and upgrade risk, and vault technical risk. Contamination propagation is the piece a traditional rating has no analog for. When one asset degrades, the rating has to answer immediately which other assets just inherited that exposure, because in onchain finance the second-order effect frequently dwarfs the first. deUSD looked like a sound synthetic dollar until the moment it was a cent and a half, and the entire distance between those two states was a dependency link to something that broke somewhere else.

This is not an argument against the agencies

We are not claiming reserve analysis or peg assessment is worthless, or that bringing these products onchain was a mistake. Independent assessment of reserve quality and peg stability is valuable. S&P's segregation and liquidity lens is sound. Moody's reserve-quality discipline is strong. Making those signals machine-readable is a genuine contribution, and institutions should want them.

The argument is narrower and harder to refute. A periodic, issuer-informed assessment delivered onchain is a useful input to onchain risk. It is not the same thing as onchain risk. The tokenized economy now needs a second, native layer that runs continuously, explains itself at the condition level, reads the chain independently, and covers the structural and synthetic assets a reserve or peg framework sets aside. One layer answers what is supposed to back this asset and how stable it has been. The other answers what this asset is doing, and what it is about to do to everything connected to it, right now.

That second layer is the category we build. Webacy's Digital Asset Ratings score stablecoins and vaults as a continuous signal across more than 600 monitored assets, with over 300 detectors, sub-second APIs, and an average alert lead time of 36 minutes ahead of public disclosure. Every score traces to a specific structural condition. When USR broke on March 22, 2026, our system flagged a 38% dislocation and issued a critical alert nearly two and a quarter hours before Resolv's own announcement. The value of that lead time is not the detection. It is the window it gives an allocator to act while acting is still possible.

Frequently asked questions

What is the difference between a rating that is onchain and one that is onchain-native? A rating is onchain when it is stored or referenceable on a blockchain, regardless of how it was produced. S&P's Stablecoin Stability Assessments delivered via Chainlink and Moody's ratings embedded on Solana are onchain in this sense: traditional assessments made machine-readable. A rating is onchain-native when it is produced from continuous, direct observation of onchain state and behavior, updates as that state changes, and explains itself through specific structural conditions. The first is about location. The second is about method.

Does delivering an assessment onchain make it real-time? It makes access real-time. A smart contract can fetch the current assessment instantly. It does not change how often the underlying analysis is refreshed or what data it is built from. An assessment produced on an analyst review cycle from issuer-supplied information is still periodic and issuer-informed after it is written to a chain.

If S&P's assessment focuses on depeg risk, isn't that already onchain risk? Peg stability is one of the most important onchain risks, and assessing it is valuable. The gap is cadence and inputs. A depeg can develop in under an hour, faster than any analyst-driven refresh, and the leading indicators are visible onchain in real time rather than in a disclosure. Measuring peg risk continuously from onchain data is a different exercise from opining on peg stability periodically, even when both are readable onchain.

Would a traditional rating have caught the Stream Finance and Elixir collapse? It is very unlikely. Stream published a roughly $400 million proof of assets on October 30, 2025, days before a $93 million offchain loss unwound the structure and dragged deUSD down about 98% through shared collateral. The determining factors were leverage, offchain manager exposure, and dependency concentration, none of which a reserve snapshot captures, and deUSD's synthetic design would likely have sat outside a reserve-collateral methodology's scope entirely.

Is continuous monitoring the same as a rating? They are converging. Monitoring produces alerts; a rating produces a standing, comparable judgment. An onchain-native rating is monitoring with the discipline of a score behind it: continuously computed, explainable, and consistent across assets, so an allocator can compare a vault to a stablecoin to a tokenized fund on the same terms.

The rating going onchain is just the bignning

The arrival of the major agencies onchain is the clearest signal yet that ratings belong there. We agree with the destination. We disagree that arriving is the same as being built for the place.

The tokenized economy inherited its trust signals from a world of quarterly filings and audited balance sheets, and it has now made those signals portable and machine-readable. That is real progress. The next step, the one that decides whether the signal actually protects the people relying on it, is to produce ratings the way onchain assets demand: continuously, transparently, independently, and across everything that can break, including the assets the traditional frameworks leave out. That is the standard the assets themselves are setting. The ratings should meet it.