USD.AI: The Loan Book Catches Up

Below we frame 12-month scenarios for USD.AI and CHIP. The model calibrates to current fundamentals: yield-bearing capital at ~84% of TVL, reserve yield at 4.5%, and net APY at 85% of gross, matching today's 8.20%/6.98% print. Origination fees are assumed at 1% of new originations. FDV is held at ~$408M across scenarios to isolate the fundamental question.

Bear assumes Season 2 ends in October without an Aave listing, points capital exits, and pipeline conversion stalls; deployment rises mechanically as TVL shrinks against termed-out loans. Base assumes the signed term sheets convert with modest organic growth on top. Bull assumes the $1B origination target lands, funded by deposit growth, loops through a live Aave integration, and external facilities in the Bullish mold.

For depositors, net APY lands between 7.9% and 8.3% in every row, because falling TVL raises the deployment ratio against a termed-out book while rising TVL brings rate compression and dilution from capital that outpaces origination. Double-digit yield in these scenarios is a gross figure: the protocol's 11.7% projection is gross, assumes full pipeline conversion on today's TVL, and would imply ~10% net at the current 85% pass-through. We treat that as an upside case rather than a base case, since it requires all $312.1M of signed paper to fund without matching deposit growth. The realistic net range of 7.0% to 8.3% still ranks among the stronger hardware-backed yields onchain, and sits well above sUSDe's ~5% current carry, with credit and liquidity risk in place of funding-rate risk.

For CHIP, the same cenarios the fee multiple from ~15x to ~5x. Base-case ~8x lands at rough parity with ENA's ~7.5x, but the claims behind the two multiples have diverged: ENA's fee switch has passed and awaits only a supply milestone, while CHIP has no proposal scheduled. Paying a peer multiple for a token without a peer claim is the premium, even where the headline ratios match; CHIP screens cheap only in the bull column, and reaching it requires flawless execution.

On tracked protocol revenue, even the bull case is ~23x with no committed claim on any of those dollars. A token whose multiple ranges that widely on execution outcomes, with no claim on the revenue behind it, is a venture position in a liquid wrapper, and that is the correct frame for sizing it. Three developments would change our stances:

1. Fee routing to sCHIP: A governance-committed share of protocol revenue to stakers converts CHIP from a claimless governance token into a yield instrument and would move us from Neutral toward Overweight, cliff calendar permitting. Continued absence into Q1 2027, as the April cliffs approach, would move us back to Underweight.