DefiLlama and Forgd introduce institutional token grades, but warn that AAA does not mean risk-free

DefiLlama and Forgd have introduced an AAA-to-CCC rating system covering 128 of 149 listed tokens, with Uniswap currently holding the dashboard’s only AAA grade at a composite score of 60.80.

DefiLlama’s live Universal Token Rating dashboard places Uniswap first with disclosure and performance scores of 7.87 and 7.72, respectively. Meteora follows with an AA grade and a composite score of 58.48, while Curve DAO ranks third at 53.32.

Developed with token advisory platform Forgd, the system grades assets by combining what a project discloses with what trading data shows. Its disclosure assessment covers areas such as tokenomics, insider wallets and commercial arrangements, while the performance side examines liquidity, spreads, venue coverage and market-maker conduct.

DefiLlama Head of Research Ryan Celaj told crypto.news that both components are required because averaging them could allow strength in one area to conceal serious problems in another.

“We’re multiplying a project’s disclosure and performance scores deliberately, because they are both necessary conditions for credibility. And ‘necessary’ is the key word. It’s not that performance and disclosures both factor in. They’re required.”

Under the formula, a project with a disclosure score of 10 and a performance score of 2 receives a composite score of 20. Celaj said an average would give the same project a much less critical score despite its weak market performance.

DefiLlama token grades require strength on both axes

The two scores range from zero to 10 and are multiplied to produce a result out of 100. AAA begins at 60, meaning a token cannot reach the top category if either component falls below six, even with a perfect score on the other axis.

AA starts at 40, with narrower bands separating A, BBB, BB, and B as weaknesses increase. Celaj said the thresholds make the highest grades difficult to obtain while creating distinctions among assets further down the table.

Although the letters resemble grades used in conventional finance, Celaj said they do not estimate default probabilities and should not be treated as equivalents to ratings issued by a traditional credit-rating agency. The format was selected because institutional traders already understand the AAA-to-CCC scale.

The approach also links stated policies to observable results. A project may publish detailed market-making terms or token-distribution plans, but the performance score tests whether liquidity, trading activity and wallet behavior match those claims.

Uniswap founder Hayden Adams drew attention to the results after UNI received the only AAA grade. Referring to the ranking in an Aug. 27 X post, Adams called it “the result of a neutral, unbiased ratings system” and referred to past criticism of Uniswap as “crypto Twitter psyops and fud.”

Market-maker conduct can lower a token’s grade

Forgd founder and CEO Shane Molidor said private contracts do not prevent the platform from assessing whether a market-making arrangement has produced durable liquidity.

Forgd monitors more than 500 market-maker engagements through reports and application programming interface data, according to Molidor. Its system measures contributions to volume and depth, uptime, compliance with agreed targets, and each provider’s record across other mandates.

“We do not determine sustainability from the disclosed contract alone,” Molidor said. “Forgd already monitors market-maker performance through its platform, giving us access to market-maker reporting and API data for the over 500 engagements we track.”

According to the executive, Forgd compares first-party information with exchange and on-chain data, including spreads, two-sided depth, venue coverage, and organic trading activity. Analysts also examine how liquidity behaves during volatile periods, token unlocks, and the period after launch incentives end.

Such checks are designed to separate persistent liquidity from volume temporarily supported by token loans, options, or other incentives, Molidor said. A project does not have to publish every commercial term, but it must provide enough verifiable evidence for Forgd to understand the arrangement and the commitments being measured.

Acceptable evidence may include relevant contract provisions, amendments, token-loan terms, options, wallet identifiers, liquidity targets, uptime requirements, incentive structures, market-maker reports and API records. Forgd also offers its market-maker monitoring software free of charge, allowing a poorly rated project to submit more data for review.

Market quality has become an important issue as institutions increase their exposure to tokenized assets. On Aug. 27, Stellar’s RWA value was reported to have increased from about $785 million in January to more than $3 billion in July, yet slightly more than $2 million had entered Blend pools that accept RWAs. The figures showed a large difference between assets issued on-chain and the amount actively used in decentralized lending.

Project claims cannot directly determine the score

Claiming a profile gives a token issuer an opportunity to submit evidence, but Molidor and Celaj said the process does not allow the issuer to assign or control its rating.

Missing information counts against the disclosure score. A project that supplies favorable details while withholding weak areas cannot obtain full disclosure credit, according to Molidor.

“The downside is that some ratings will appear artificially low until a project provides the necessary disclosures,” Molidor said. “But the upside is that for projects, there is no downside to being transparent, and no upside to selective disclosure.”

The performance score adds a separate check by using exchange records, on-chain events, and Forgd’s monitoring tools. Its inputs include depth, spreads, volume, exchange coverage, derivatives conditions, tokenomics, and adherence to market-making targets.

Exceptionally strong performance in one category is capped, Celaj said, preventing one metric from cancelling persistent weakness elsewhere. The methodology also excludes venues regarded as unreliable from relevant calculations.

Ratings update continuously rather than relying on a single audit. Material disclosures that remain outdated for more than 60 days receive a penalty, while verifiable events such as token unlocks and exchange listings enter the performance assessment automatically.

Even with those controls, both executives acknowledged limits. Molidor said the system cannot prove that an undisclosed commercial relationship does not exist. It can identify missing information, inconsistent claims, and activity that does not match a project’s account, but its grade cannot guarantee that every relationship has been found.

Celaj similarly said that no grading model can be considered impossible to manipulate. DefiLlama has made its methodology and category-level results available so users can trace grades and challenge disputed information, while the team plans to adjust the system if projects find ways to exploit it.

An AAA token grade does not predict returns

Neither DefiLlama nor Forgd has gathered enough long-term evidence to claim that highly rated tokens suffer smaller drawdowns or fewer market failures.

Molidor said a high performance score necessarily corresponds with stronger measured depth, tighter spreads, and more extensive liquidity because the system uses those conditions as inputs. Price declines can still result from security breaches, governance failures, or market conditions that the rating does not assess.

“An AAA grade means that, at this point in time, a token demonstrates a strong combination of disclosure quality and observable market performance under the UTR methodology,” Molidor said.

“It does not mean the token is risk-free, that its price will appreciate, or that an institution can replace its own legal, technical and financial due diligence.”

A CCC grade identifies substantial problems in disclosure, performance, or both, according to Molidor. It does not establish that a project is fraudulent or certain to fail, but it points institutions toward areas requiring additional review.

Celaj described the rating as a screening and monitoring tool rather than an investment recommendation. In his view, the system creates a dataset that researchers can eventually use to test whether combining disclosures with market data produces a better predictive signal than assessing each category separately.

Institutional interest gives that test practical relevance, especially for tokens linked to real-world assets. On July 31, an article on Ondo Finance reported that tokenized securities exceeded $36 billion in 2026, including approximately $12.88 billion in tokenized U.S. Treasuries.

For American institutions, token grades may help organize preliminary market-structure reviews, but regulated tokenized products remain subject to separate custody, eligibility, and securities requirements. On Aug. 3, BlackRock launched two tokenized money-market products backed by cash, short-term U.S. Treasuries and Treasury-backed repurchase agreements, with transfers restricted to approved investors and compliant wallets.

UTR does not assess every risk attached to such assets. Celaj specifically said its methodology does not measure cybersecurity risk, which has caused some of crypto’s largest historical drawdowns.