Kurtosis manages US$800K TVL for other quant funds, and is building layer-by-layer risk ratings for onchain yield positions, so fund analysts can see which layer of a position drives its risk and defend the allocation to their investment committee. We run vault strategies with USD$800K TVL too.
The problem
Funds can see a token’s yield. They can’t easily see the risk of the position they actually hold once that token is wrapped, lent, looped or turned into a principal token (PT). Analysts rebuild the diligence by hand from issuer docs, audits and dashboards, then have to defend it to a PM or investment committee.
What we’re building
Kurtosis Ratings rates the exact position, not the token name. Our “Russian Doll” method opens each position layer by layer and follows its dependencies back to the underlying assets and the route investors use to exit. We’re building the grading step: one overall grade per position, with the reasoning behind every layer, the evidence for it, and a flag wherever the evidence is thin. Analysts can then challenge the assumptions instead of trusting a black box.
On Arbitrum
Our worked example is sUSDai, PT-sUSDai on Pendle, and sUSDai looped on Morpho, all on Arbitrum. These are three positions on the same underlying asset, each with different risks. The PT adds Pendle market, maturity and exit-liquidity risk. The loop adds lending-market, oracle, LTV and liquidation risk.
Who it’s for
Analysts and risk leads at crypto-native funds who must explain an unfamiliar onchain position to their PM, CIO or risk committee before capital goes in.
Traction
1st and 2nd place globally at a global algo trading hackathon.
USD 800k in non-dilutive vault capital from Presto Labs, a quantitative trading firm
Proprietary data partnership with Elfa AI