[ARFC] Deploy Aave V4 on the Monad Network

Governance New Market
1

title: [ARFC] Deploy Aave V4 on the Monad Network

author: @TokenLogic

created: 2026-09-11

Summary

This proposal recommends deploying an Aave V4 instance on the Monad Network, with an initial market structure designed to support tokenized equities and cash equivalent assets.

The proposed instance would use a hub and spoke architecture to separate assets according to their underlying risk profiles while maintaining access to shared stablecoin liquidity. The structure is designed to limit cross exposure between assets with materially different volatility profiles and provide a framework that can scale as the tokenized equities market develops.

Motivation

The tokenization of RWAs has emerged as one of the fastest growing segments of onchain finance, with tokenized equities and other traditional financial assets increasingly becoming available on public blockchains. This expansion creates an opportunity for lending markets to provide additional utility around these assets.

Tokenized equities are particularly relevant as they provide onchain access to familiar financial exposures while retaining a direct connection to underlying traditional securities. As the range and adoption of tokenized assets continues to expand, there is an opportunity for Aave to establish infrastructure that can support these assets within a controlled framework.

Incentives Package

The initial $15M incentive budget allocated by the Monad Foundation under the Aave V3.7 deployment proposal will also be used to support the transition to and growth of Aave V4 on Monad.

The incentive budget is intended to support liquidity growth and adoption of the V4 deployment, with the specific allocation and distribution strategy to be determined as the deployment progresses.

Market Structure

The proposed Monad Tokenized Equities Hub is structured around three risk differentiated spokes for tokenized equities and cash equivalent assets. The objective is to keep assets with broadly similar market risk together while isolating the amount of hub liquidity that can be drawn by each risk tier.

The Core Assets Spoke contains lower volatility collateral, Growth Assets Spoke contains higher volatility names, and Emerging Listings Spoke is reserved for assets with limited history or exceptionally high volatility.

USDC and USDT0 provide the borrow liquidity across all spokes, while tokenized equities and cash equivalents are collateral only. The collateral factors shown below are indicative and are intended to reflect both observed price volatility and the risk of price gaps while the underlying market is closed over the weekend.

image
image2560×1200 207 KB

Figure 1. Proposed Monad Equities Hub and spoke structure, including indicative risk parameters and liquidity draw caps

Risk Segmentation Across Spokes

The spoke structure is primarily designed to group assets with similar risk profiles and, critically, to prevent cross collateralisation from transmitting liquidation risk between assets with materially different volatility. By separating higher volatility assets into distinct spokes, a sharp decline in one asset is less likely to create liquidation pressure on more stable assets within the same collateral pool. This allows each spoke to be managed against a more consistent underlying risk profile and limits the potential for contagion across asset classes.

  • Core Assets: The Core Assets spoke contains assets with one year realised volatility below 25%. SPYx and QQQx have annualised volatility of 13% and 20% respectively, while SGOVx is effectively cash equivalent.
  • Growth Assets: The Growth spoke covers assets with one year realised volatility between 25% and 60%. These assets therefore require materially more conservative parameters. As an asset matures, grows in size, and its volatility declines, it can graduate to the Core Assets spoke.
  • Emerging Listings: The third spoke is reserved for assets with less than one year of history or volatility above 60%. It is empty at launch, but provides a framework for onboarding new or exceptionally volatile assets without immediately introducing them into the more established risk pools.

image
image1920×1080 148 KB

Figure 2. Daily price moves over the last year. The distributions show the increasing dispersion of daily returns as asset volatility rises.

The daily return data illustrates why volatility is a useful first level of segmentation. For example, SPYx has a typical daily move of 0.8%, compared with 3.0% for TSLAx and 3.5% for EWYx. The largest observed daily losses also become materially larger across the risk spectrum, reaching 14.5% for TSLAx and 14.1% for EWYx. This supports separating assets into distinct collateral and liquidity risk buckets.

Managing Weekend Price Gap Risk

Daily volatility alone does not fully capture the liquidation risk of these assets. Because the underlying securities trade only during the market session, their reference prices remain effectively unchanged for roughly 60 hours over the weekend. During this period, xStocks will be neither mintable nor redeemable. A material change in the value of the underlying can therefore appear as a single Monday opening gap, before normal market based liquidation mechanisms can act.

image
image1920×1080 151 KB

Figure 3. Historical Friday close to Monday open price gaps over five years, with the drop side 99th percentile highlighted.

This makes the historical weekend gap distribution an important input into the collateral parameters. The CF ranges presented in the market structure diagram reflect both the asset’s realised volatility and the tail, represented by the 99th percentile, of its historical weekend drop distribution. The parameters are intended to provide sufficient protection for a maximally leveraged position to withstand a weekend gap materially worse than anything observed over the five year sample, including the liquidator’s bonus, before bad debt arises.

As an additional layer of protection, Aave v4 enables collateral factors to be adjusted dynamically. CFs could therefore be temporarily reduced while markets are closed to limit borrowing capacity. While this would not trigger liquidations for existing positions, users whose positions exceed the temporarily lower factor would be unable to withdraw collateral or borrow further. The CF could then be restored when the market reopens.

Taken together, the proposed structure provides a framework for managing tokenized equities with different risk profiles. By separating assets into distinct spokes, the design limits cross exposure between risk tiers while allowing assets to move between spokes as their risk characteristics evolve. This enables the market to scale while maintaining clear risk boundaries between assets with different characteristics.

Specification

This section will be updated upon receiving feedback from various stakeholders in the lead-up to the deployment.

Hub and Spoke Configuration

The proposed initial Monad deployment will activate one Liquidity Hub and three Spokes: a Core Assets Spoke, a Growth Assets Spoke, and an Emerging Listings Spoke.

Hub Assets
Monad Tokenized Equities Hub SPYx, QQQx, SGOVx, NVDAx, SMHx, TSLAx, EWYx, USDC, USDT0
Spoke Collateral Borrowable
Core Assets Spoke SPYx, QQQx, SGOVx USDC, USDT0
Growth Assets Spoke TSLAx, EWYx, NVDAx, SMHx USDC, USDT0
Emerging Listings Spoke None at launch USDC, USDT0

Oracle Configuration

The proposed deployment plan is to use Chainlink’s 24/5 oracle infrastructure for xStock price feeds during the initial deployment. These oracles stitch together the regular, extended and overnight session into a continuous price during the week, with no price updates over the weekend. Chainlink Labs is working on a 24/7 pricing solution that will extend these price feeds to provide continuous pricing updates through the weekend gap as well. We are collaborating closely with them on this design and more details on this will be shared by Chainlink Labs in the near future.

Next Steps

  1. Gather community feedback on the proposed Monad V4 deployment and market structure.
  2. Finalise the initial asset configuration and risk parameters.
  3. Progress the proposal to the ARFC Snapshot stage.
  4. Subject to a successful ARFC Snapshot, submit an AIP for final confirmation and activation.

Disclaimer

TokenLogic is an active service provider to the Aave DAO, the beneficiary of stream 100086 and the KPI as outlined in this publication. The scope of this engagement is available via this forum proposal.

TokenLogic supports and maintains an independent delegate voting platform within the Aave community.

TokenLogic and associated entities have no undisclosed material conflicts of interest at the time of submission.

Copyright

Copyright and related rights waived via CC0.

2

A measurement note from a different chain, offered as feedback rather than as a claim about any asset here.

I have not measured Monad or xStocks. What follows is from Robinhood Chain (Uniswap v4, USDG-quoted), where the same tickers trade as a different issuer’s wrapper — so it bears on the choice of axis, not on your parameters.

Price and exit are different axes, and the Core spoke is sorted on one of them. SPYx and QQQx sit together in the Core Assets spoke, and on the price axis that is hard to argue with — 13% and 20% realised volatility. Measured at pinned blocks across 466 rounds on the other chain, the same two tickers are three orders of magnitude apart on exit:
median min rounds >= 99%

SPY · $100,000 99.7490% 99.2732% 100%
QQQ · $100,000 0.0314% 0.0314% 0% (about $31 on $100,000)
QQQ · $10,000 0.2626% 0.2626% 0% (identical in every round)
At those two sizes the QQQ figure is identical in all 466 rounds — zero variance. It is not a volatility event that a wider collateral factor absorbs, and there is a mechanism behind it: in the table I enumerated, QQQ was the only one of the nine tokens whose cheapest USDG route charged 70%, while the other eight tokens’ cheapest routes carried ordinary fees. A 70% fee caps a round trip at (1 − 0.70)² = 9.00% before any price impact. That is a standing venue-structure difference between two assets the tiering treats as one risk class.

The number is also method-dependent, which is the part I would most want a risk reviewer to see. The same block and the same $100,000 in QQQ give three answers depending on which pools the measurement can see: 0.03% against a pool table enumerated from a single factory, 99.74% through the pool that can actually serve that size, and 99.87% for the sell leg alone. The venue that mattered was created after my table’s freeze at block 53,983,886 and was never in it. Two disclosures travel with my figures: every cell reads “the best of the pools in that table”, not the best on chain; and recovery is not a fee — where a one-way cost is derived as 1 − sqrt(recovery), it contains fee and price impact together.

So I am not arguing that QQQx cannot be sold. The point is narrower and it is about sizing: a cap or a collateral factor set from a single depth number inherits the blind spot of whichever enumeration produced it.

Your text already proposes the right mechanism — collateral factors reduced while the market is closed. What that mechanism needs as an input is a depth measurement at the size a liquidation would have to clear, rather than a price series. On provenance: each round-trip round carries a canonical hash, the on-chain ledger for that canon was deployed after this window so its commitment coverage is whatever the contract itself reports, and the one-sided figures come from a pre-registered series with no rounds committed.

One question, because it decides what such an input should look like: for the Core spoke, is the binding size the largest single position that could be liquidated, or a multiple of typical depth?

Method, data, verifier and the project’s error archive: GitHub - dirkdiggler1026/executability-report: Independent measurement of executable depth across on-chain aggregators, tokenized stocks on Robinhood Chain, and prediction markets — data & verification scripts included. · GitHub

Disclosure: commissioned measurement is part of what I do; the method, the data and the corrections stay public either way.

— dirkdiggler1026

3

Thanks to TokenLogic for the proposal. A dedicated tokenized-equity Hub with volatility-tiered spokes is a sensible structure, and pricing in weekend gaps is very helpful to consider as a critical risk factor in tokenized equities based systems. Before this goes to Snapshot, I’d like to raise three points.

1. Projections. The ARFC sets out the structure but no caps, IRM, reserve factor, or expected volumes. For comparison, the Base Equities Hub published a $21M borrow cap and LTVs of 65–79% before going to vote. As a rough guide, if a Monad Hub reached similar borrowing at around 5% APR, the DAO would earn roughly $100–200k a year, depending on the reserve factor. That may be worth it strategically, but it would be good to see the case made publicly. Could TokenLogic share:

  • proposed caps and risk parameters
  • current xStock supply and DEX depth on Monad, and how xStocks reach Monad (native issuance or bridged)
  • 6 and 12-month projections for supply, borrows, and DAO revenue, plus the costs the DAO takes on (oracle feeds, risk-provider scope, operations).

2. Incentives between V3 and V4. The $15M cited appears to be the first-year budget Monad Foundation committed at the V3 Monad launch in July, now “also” to be used for V4. Can you confirm whether this is new funding or a reallocation? If it is reallocated, how much moves from V3 to V4, on what schedule, and what happens to V3 Monad, which is still early and incentive-supported? Is there a plan for V3 and V4 to coexist on Monad, or for V3 to migrate over time?

3. Why Monad? Aave already supports tokenized equities on X Layer (V3, xStocks) and Base (V4, Coinbase-issued stocks). xStocks are also issued natively on other chains. What makes Monad the right venue for a third equities deployment: existing xStock liquidity, user demand, the incentive package, or something else? A short comparison against the alternatives (for example, expanding the existing X Layer or Base markets) would help the DAO see why this chain, and why now.

1 Like
4

On the comparison against existing markets (the third point above): I can contribute the Base half of it, since the Base Equities Hub is the baseline cited here — and its exit side turns out to be uneven in a way a cap cannot see.

At one pinned block (52,170,281) I measured the Base V4 equities venues reachable through the two factories I confirmed on-chain (at least three further factories hold pools for these wrappers and are not measured). The published subset is 19 rows: executability-report/measurements/base-depth at main · dirkdiggler1026/executability-report · GitHub . The finding I would want read before caps are set: two pools for the same underlying, at the same block, differ by roughly 600×. NVDAc/USDC at fee 3000 returns 99.31% on a USD 100 exit; a second NVDAc/USDC pool at fee 500 cannot return more than 0.168% of a USD 100 entry, because that is all it holds (that row is published as an on-chain holdings bound, not as a walk result). TVL and 2%-impact figures see neither the difference nor its direction.

One scope note. My largest published tier is USD 1,000, while this proposal discusses a USD 21M borrow cap — but the argument above is scale-independent: it is about which pool a liquidator lands in, and it holds at USD 100 as well as at size. Larger tiers are measured but withheld: they depend on one code path — accumulation across more than one liquidity range — that I have not been able to validate against a real swap, because I could not find one in recent history that crosses two initialised ticks cleanly. If that validation lands, I will publish them; if it does not, that stays the stated reason. I have not measured Monad or X Layer.

On method. Two runs at the same pinned block disagreed on tick-heavy rows. I have since found the likely cause in my own reader, not in the chain: a failed ticks() read returned 0 and a failed bitmap read returned a sentinel, either of which silently reads as “no liquidity change here” or “nothing initialised in this direction” — exactly the two error directions I saw. Both are being changed to propagate the failure so the row drops out instead of carrying a wrong number. Separately, I verified this endpoint supports block-hash-pinned reads (a forged hash errors), so the next publication pins by hash and removes a different failure mode.

Would a Monad-side run on the same method (pinned block, USD 100 to USD 100,000, both legs against the same state) be useful here — or is depth already covered by someone else’s numbers in this thread?

5

On the first of the three points — supply, depth, and how xStocks reach Monad — I checked Monad mainnet rather than the proposal text, and the answer today is that there is almost nothing to measure yet:

Every wrapped xStock on Monad reports totalAssets() == 0. The seven tokens exist at byte-identical EIP-1967 proxy code (decimals 18, same addresses on BSC and Optimism), and their wrappers exist and point back at them — but the wrappers hold nothing, and I could not find a real pool for any of them. The only “QQQX/WMON” and “NVDAX/WMON” pools I could find trade near 1e-8 against a real QQQx price around 600: those are impostors, matched by symbol rather than by address, and should not be used as depth evidence.

The proposal itself contains no contract addresses at all — the specification section says it “will be updated upon receiving feedback”. An asset whose address has not been specified is one whose depth cannot be measured or independently verified.

For anyone who will verify on Monad: block-hash-pinned reads with requireCanonical work on rpc.monad.xyz, and state is pinnable to about a million blocks back, but eth_getLogs is capped at a 100-block range, so coverage is cheap in latency and expensive in request count.

Would it be useful if I re-ran this at the block where the Hub goes live, so the first numbers are measured rather than assumed?