Executive Summary
We propose raising the JAAA reserve-level LTV from 81% to 87% and LT from 86% to 89% on Horizon, while keeping the 5% liquidation bonus. The reserve tier would then sit one percentage point below the JAAA GHO E-Mode, which operates at 88% / 90% / 5%, on both LTV and LT.
The change follows the fund’s settlement record, which Centrifuge has confirmed. Between 10 October 2025 and 5 October 2026, the fund processed 93 redemption orders totalling USD 717.3M across all JAAA vaults. Of these, 91 were paid on the day of the request or the next business day, and the remaining two were paid T+2.
We size the parameters against a three-business-day exit window, one day longer than the slowest settlement observed. Over such a window, the loss a liquidation can carry is the decline in collateral value between seizure and redemption. At 89% LT and a 5% liquidation bonus, the 6.55% solvency margin covers the worst decline of the Palmer Square CLO Senior Debt Index (CLOSE), our stress proxy, over any window of up to six trading days (5.99%), twice the assumed exit window.
The fund’s portfolio has not deteriorated since our original evaluation. Its holdings remain floating-rate AAA CLO tranches, its weighted average legal maturity has shortened with time, and diversification has recovered to 35 positions across 22 managers after the Q1 2026 outflows.
Background
JAAA is the tokenized share class of the Janus Henderson Anemoy AAA CLO Fund, issued through Centrifuge. The fund holds AAA-rated tranches of collateralized loan obligations, and its price on Horizon is derived from the fund NAV, which Chainlink publishes on-chain through the JAAA NAV feed.
Our initial evaluation in December 2025 set the reserve-tier parameters at 81% / 86% / 5%. At that time, redemptions were described as serviced within T+3 with the possibility of longer processing, and the fund had no settlement record under heavy outflows. We therefore sized the LT so that the solvency margin covered the full 8.33% maximum drawdown of CLOSE, which assumes a liquidator may have to hold seized JAAA through an entire sell-off. The 88% / 90% configuration was granted only in the JAAA GHO E-Mode.
Over the past twelve months, the fund has settled more than USD 700M of redemptions, including a single order equal to 43.9% of fund AUM, and Centrifuge has confirmed the settlement times behind them. The exit duration of seized collateral is now supported by an observed settlement record, and we size the reserve-tier parameters against the decline over a three-business-day exit window.
Redemption Settlement
Centrifuge reports the following settlement record from on-chain data across all JAAA vaults, covering 10 October 2025 to 5 October 2026.
| Settlement | Orders |
|---|---|
| T+0 | 63 |
| T+1 | 28 |
| T+2 | 2 |
| T+3 or later | 0 |
| Total | 93 (USD 717.3M) |
The largest redemptions in the period were:
- USD 326.9M on 9 March 2026 (43.9% of fund AUM), priced at the NAV of the request day and paid T+2.
- USD 300.1M across six orders between 6 and 16 January 2026, each paid T+0 or T+1.
- USD 200.0M across two orders on 21 and 22 September 2026, on Base and Solana, each paid T+1.
Centrifuge executes redemptions on the day of the request whenever possible, and states that it can redeem USD 300M within T+1. We use T+3 as a conservative exit window in the stress analysis below.
All redemptions settled within 0.5 bps of published NAV on a payout-weighted basis. The March event represents the strongest available real-world stress test. The fund liquidated USD 322.9M par across 18 CLOs at a par-weighted price of 99.87, while AUM fell 61% from USD 1,022M to USD 396M over Q1 2026. Despite this, token NAV rose from 1.0209 to 1.0274.
On 22 and 23 September 2026, the pool manager burned 152.3M shares (USD 160.0M at the published NAV of each day) held by a single address, outside the standard redemption request process.
Source: LlamaRisk, October 15, 2026
Portfolio Composition
We compare the tokenized fund’s own holdings across three snapshots. The first is Particula’s November 2025 rating report, the second is the fund’s holdings in May 2026 after the Q1 outflows, and the third is the holdings table in the fund’s Centrifuge pool metadata.
| Metric | November 2025 | May 2026 | September 2026 |
|---|---|---|---|
| Fund AUM | USD 1,013M | USD 409M | USD 577M (CLO holdings) |
| CLO positions | 54 | 20 | 35 |
| Largest position | 3.26% | 8.68% | 5.20% |
| Top-10 positions | 30.0% | 63.8% | 37.2% |
| Distinct CLO managers | Not disclosed | 13 | 22 |
| Largest manager | Not disclosed | 18.6% | 12.3% |
| Weighted average legal final maturity | 11.8 years | Not disclosed | 11.0 years |
| Share with legal final maturity beyond 12 years | 62.2% | Not disclosed | 30.0% |
The holdings remain floating-rate senior CLO tranches, so interest-rate duration stays close to zero. The ETF that follows the same strategy reports an effective duration of 0.04 years, against 0.09 years at the time of our original evaluation.
Weighted average legal final maturity shortened from 11.8 to 11.0 years, in line with the time elapsed between the two snapshots, and the share of holdings maturing beyond 12 years halved. The fund has not extended the maturity profile of its book.
Diversification narrowed during the Q1 2026 outflows, when the fund sold down to 20 positions and the top ten reached 63.8% of holdings. It has since recovered to 35 positions across 22 managers, with the largest position at 5.2% and the top ten at 37.2%. The book remains more concentrated than the 54-position portfolio of November 2025.
The mandate remains AAA CLO tranches, and the ETF reports 99.2% AAA exposure. Two positions totalling 3.5% of holdings are A2 tranches, a class that carries a rating below AAA in some CLO structures, and the holdings table does not show their ratings.
Overall, the portfolio has not deteriorated relative to our original evaluation in interest-rate duration, maturity profile or mandate. Concentration is higher than in November 2025 but well below the May 2026 level.
Stress Analysis
We use CLOSE as the stress benchmark, since the tokenized fund lacks a track record through a credit event of comparable severity. CLOSE tracks senior secured floating-rate CLO debt, the same asset class the fund holds. Over their overlapping history, 7-day returns of the Janus Henderson AAA CLO ETF, which follows the same strategy as the fund, showed a 0.76 correlation and a 0.98 beta to the index. The bounds below use daily CLOSE data from 29 May 2015 to 1 December 2025.
The buffer to LT is the NAV decline that takes a position opened at maximum LTV to liquidation.
The solvency margin is the decline, measured from the point a position reaches LT, that collateral can absorb while still covering the debt plus the liquidation bonus.
The margin from entry is the decline a position opened at maximum LTV can absorb before its collateral no longer covers its debt plus the bonus.
At 87% / 89% / 5%, the buffer to LT is 2.25%, the solvency margin is 6.55%, and the margin from entry is 8.65%.
Source: LlamaRisk, October 15, 2026
A liquidator that seizes JAAA at LT and submits a redemption request the same day carries NAV risk until the redemption is priced. In the March 2026 redemption, the order was priced at the NAV of the request day, which places the price exposure well inside the settlement window. We nonetheless size against the full three-business-day exit window, over which the 5% bonus covers the proxy’s worst decline with a 1.19% margin.
The 8.33% maximum drawdown built up over 27 days. In a decline of that shape, positions cross LT progressively and are liquidated and redeemed within days, so no single liquidation carries the full drawdown. If redemptions were suspended and liquidators could not exit, a position opened at the 87% maximum LTV at the start of the decline would still hold collateral worth more than its debt plus the 5% bonus at the trough, with a 0.32% margin. Redemptions have not been suspended or deferred in any of the fund’s 93 orders.
The token’s own NAV has moved far less than the proxy. Between 23 April and 14 October 2026, the largest daily decline in the Chainlink JAAA NAV feed was 0.07%, and the largest 3-day decline was 0.08%.
Horizon Exposure
JAAA supply on Horizon is $51.9M at a NAV of 1.0522, across eleven addresses. Three positions holding 9,772,596 JAAA (19.8%) use the JAAA GHO E-Mode and are unaffected by this change.
The remaining 39.5M JAAA (80.2%) sit in the reserve tier, where seven positions carry USD 45.0M of debt, of which USD 34.5M is RLUSD and USD 10.5M is GHO. These positions currently borrow 78.2% to 83.4% of their collateral value, close to the current 81% JAAA LTV, which indicates looped positions that are likely to re-lever toward the new limit. At the proposed LT, the health factors of these seven positions rise by 0.023 to 0.038, and the lowest moves from 1.048 to 1.071.
Maximum looped leverage on JAAA rises from 5.26x to 7.69x, following \frac{1}{1 - LTV}. The supply cap bounds the resulting exposure. At the full 60M cap and current NAV, JAAA collateral would be worth USD 63.1M and could support at most USD 54.9M of debt at 87% LTV.
Liquidation Considerations
Restoring a position from LT back to maximum LTV requires a partial liquidation, where x is the debt repaid.
At 87% / 89% / 5%, a liquidator must repay 26.0% of outstanding debt to bring a position from 89% back to 87%, down from 38.9% at the current 81% / 86% / 5%. For the largest reserve-tier position, which carries USD 20.5M of debt, this amounts to about USD 5.3M of RLUSD or GHO.
JAAA is permissioned at the token level, so every liquidator and recipient of seized collateral must be on the fund’s memberlist. A liquidator then exits by redeeming with the fund. The USD 5.3M restoration amount above is small relative to the USD 300.1M the fund redeemed within T+1 in January 2026.
The update carries the following trade-offs:
- Borrowers gain capital efficiency, with maximum looped leverage rising from 5.26x to 7.69x, and RLUSD and USDC borrowing against JAAA moves to within one percentage point of the GHO E-Mode terms.
- A position opened at 87% LTV becomes liquidatable after a 2.25% NAV decline, which the proxy exceeded over two trading days in March 2020.
- The 6.55% solvency margin relies on the fund continuing to settle within three business days. A suspension or deferral of redemptions would leave only the 8.65% margin from entry, which covers the proxy’s maximum drawdown by 0.32%.
- The largest reserve-tier position holds 38.9% of reserve-tier JAAA, so a single borrower’s management of its position drives most of the liquidation exposure. The new parameters would still allow the borrowers self-manage the positions to rebalance in face of drawdowns, according to historical drawdown velocity.
Recommendation
Based on the fund’s settlement record and the proxy decline over a three-business-day exit window, we recommend updating the JAAA reserve-tier parameters on Horizon as follows:
| Configuration | LTV | LT | LB | Buffer to LT | Margin from entry | Max leverage |
|---|---|---|---|---|---|---|
| Reserve, current | 81% | 86% | 5% | 5.81% | 14.95% | 5.26x |
| Reserve, proposed | 87% | 89% | 5% | 2.25% | 8.65% | 7.69x |
| E-Mode 10 (JAAA GHO), unchanged | 88% | 90% | 5% | 2.22% | 7.60% | 8.33x |
Specification
| Asset | Parameter | Current | Recommended |
|---|---|---|---|
| JAAA | LTV | 81% | 87% |
| JAAA | LT | 86% | 89% |
| JAAA | LB | 5% | 5% |
Next Steps
We will move forward and implement these updates via the Risk Steward process. The implemented changes can be reviewed on the LlamaRisk Risk Stewards Dashboard.
Disclosure
This review was independently prepared by LlamaRisk, a DeFi risk service provider funded in part by the Aave DAO. LlamaRisk is not directly affiliated with the protocol(s) reviewed in this assessment and did not receive any compensation from the protocol(s) or their affiliated entities for this work.
The information provided should not be construed as legal, financial, tax, or professional advice.