Felix feUSD: Hyperliquid Stablecoin, Collateral Design, and Redemption Risk

Pre-screen Decision

Decision: full research, not a quick note.

Felix feUSD / FEUSD deserves full-depth treatment because it is not just a small stablecoin ticker on a data aggregator. It is one of the most important native dollar instruments inside the Hyperliquid ecosystem, and its risk profile is structurally tied to HYPE collateral, HyperEVM execution, HyperCore spot liquidity, RedStone oracles, Liquity v2 mechanics, Stability Pool depth, and the broader effort to make Hyperliquid more than a perps exchange. Felix also touches adjacent products such as vanilla lending, spot equities, and Felix Perpetual Futures, but this memo focuses on the CDP Market and the FEUSD stablecoin. That scope matters because "Felix" can refer to a protocol, a possible HyperCore token symbol, multiple product lines, or unrelated assets on other chains. This report uses the Felix documentation, the FEUSD contract address 0x02c6a2fa58cc01a18b8d9e00ea48d65e4df26c70, the Hyperliquid spot token metadata for FEUSD, and the DefiLlama stablecoin listing as the identity anchors.

Local duplicate check was completed before live research. A read-only registry lookup against data/research-map/registry.json found no match for Felix, feUSD, FEUSD, felix-feusd, or surf:felix. A content search found the candidate backlog item surf:felix, a few incidental mentions of feUSD in Hyperliquid thought posts, and no existing Research MDX or Research Map card for Felix feUSD. The only local candidate record was Felix feUSD / FEUSD, marked as a high-priority Surf candidate. Because the user explicitly requested a new MDX and no high-confidence duplicate exists, this report proceeds as a new full research artifact while intentionally not editing Research Map cards, logo assets, registry files, package files, or session logs.

The pre-screen reason for full depth is fourfold. First, feUSD is a stablecoin, and stablecoins fail through details: collateral composition, liquidation design, redemption priority, liquidity depth, oracle design, admin controls, and secondary-market peg support matter more than marketing. Second, Felix is a native Hyperliquid application. Hyperliquid's HyperEVM inherits security from HyperBFT, uses HYPE as the gas token, and coexists with HyperCore's orderbook and spot-token system. That gives Felix distribution and composability, but also concentrates ecosystem risk. Third, the data has meaningful conflicts. DefiLlama's stablecoin API and the HyperEVM ERC-20 contract imply roughly 14.1 million FEUSD supply on the June 29, 2026 snapshot, while CoinGecko-style and DexScreener-style market-cap fields can show roughly 75 million dollars. That discrepancy must be explained rather than ignored. Fourth, Felix is using Liquity v2 architecture with Felix-specific changes: mint caps, admin-adjustable trove parameters, admin pausing, and upgradability before planned removal of admin privileges. That is a real risk tradeoff, not a minor footnote.

This report treats FEUSD as a utility and risk asset, not a token with equity-like upside. Holding FEUSD does not create upside from Felix growth in the way holding a governance or revenue-share token might. The holder's expected return comes from using the stablecoin in Hyperliquid markets, depositing into Stability Pools, earning borrower interest and liquidation proceeds, or deploying it in external liquidity venues. The research question is therefore not "can FEUSD 10x?" It is whether feUSD is a robust enough dollar primitive to underwrite borrowing, liquidity, and trading strategies in the Hyperliquid ecosystem.

Thesis / Executive Summary

Core thesis: Felix feUSD is a strategically useful but still young Hyperliquid-native CDP stablecoin. It is attractive because it converts HYPE, UBTC, and liquid-staked HYPE collateral into borrowable dollar liquidity without relying on an offchain custodian, uses Liquity v2 style user-set interest rates to price debt, and gives Hyperliquid users a native stablecoin that can circulate across HyperEVM and HyperCore. It is risky because its peg depends on redemption arbitrage, Stability Pool depth, collateral liquidity, oracle correctness, admin controls, and the willingness of arbitrageurs and depositors to act during stress. My final view is watchlist / selective use, not passive large-size parking.

Felix's own documentation is clear about the mechanism. The CDP Market overview describes feUSD as a money market connecting borrowers / minters with Stability Pool depositors and says it uses Liquity v2 architecture. Borrowers deposit collateral such as HYPE and UBTC, mint feUSD, pay an upfront borrow fee, and set their own interest rate. Stability Pool depositors buy or hold feUSD, deposit it into collateral-specific pools, and earn borrower interest, a share of upfront fees, and liquidation gains. The How it Works page says feUSD is over-collateralized, has no centralized custodian, and relies on liquidators and redeemers to keep circulating feUSD fully backed. That is the right mental model: feUSD is not USDC, USDT, or a bank-account stablecoin. It is collateralized debt plus liquidation infrastructure.

The most important design choice is redemption ordering. When feUSD trades below one dollar, holders can redeem feUSD for one dollar of collateral minus a fee. Redemptions are processed against open borrower positions starting with the lowest interest-rate positions. Felix's FAQ explains the core game: lower-rate borrowers get cheaper debt but are redeemed first, while higher-rate borrowers pay more to reduce redemption probability. This is inherited from Liquity v2 / BOLD style design, where borrower-set rates create a market for debt pricing. The bullish interpretation is that Felix can maintain the peg without a centralized issuer or governance-set interest rate. The bearish interpretation is that this design works only if there is enough liquidity, enough collateral quality, and enough arbitrage activity when FEUSD trades below peg.

The Hyperliquid relationship is central. Felix runs on Hyperliquid infrastructure. The Minting / Borrowing feUSD page says mainnet users can deposit HYPE and UBTC as collateral within the Felix protocol and borrow feUSD, and it specifically refers to transferring tokens on the Hyperliquid L1 network. The developer page for Market 1: feUSD CDP lists FEUSD, WHYPE, UBTC, kHYPE, and wstHYPE-related contract addresses, while Hyperliquid's own HyperEVM documentation explains chain ID 999, the EVM execution environment, and HYPE as the native gas token. Hyperliquid's HyperCore <> HyperEVM transfer documentation also matters because feUSD can be linked between the EVM contract world and HyperCore-style spot balances. This is why feUSD is not just "a stablecoin on an EVM chain." It is a native dollar instrument in an exchange-centered L1.

The current data snapshot is mixed. DefiLlama's stablecoin API showed Felix feUSD as a crypto-backed USD-pegged asset on Hyperliquid L1 with a price around 0.99976 and circulating supply around 14.09M on June 29, 2026, while an ERC-20 totalSupply() call to the FEUSD contract on HyperEVM returned roughly 14.09M FEUSD with 18 decimals. DefiLlama's protocol endpoint showed Felix TVL near 103.9M, down from roughly 117.7M seven days prior and 147.1M thirty days prior. The DefiLlama fees overview showed Felix CDP 30-day fees near 83,988 dollars and one-year fees near 4.37M dollars, with methodology tying fees to borrow interest, redemption fees, liquidation profits, and gas compensation. Those are real usage signals, but the contraction in TVL and the modest stablecoin circulating supply make the asset early and reflexive.

Secondary-market liquidity is a concern. DexScreener's FEUSD token endpoint showed multiple HyperEVM pools, including feUSD/WHYPE on HyperSwap, feUSD/USDT0 on Project X, feUSD/USDT0 on HyperSwap, feUSD/USDXL, feUSD/USDHL, and feUSD/USDC, but the visible individual pool liquidities were mostly small, with many pools in the low thousands to low tens of thousands of dollars. Hyperliquid's Core API showed a FEUSD/USDC spot market with a mid price around 0.9987 and daily notional volume around 529K, which is more meaningful than the small AMM pools, but the orderbook depth and withdrawal/redemption routes still need active monitoring. A stablecoin can be fully backed in theory and still be painful to exit in size if secondary liquidity is thin.

Investment view: feUSD is useful for active Hyperliquid users who understand CDP and Stability Pool risk, but it is not a generic cash parking asset. The bull case is that Felix becomes the native credit layer for HYPE collateral, feUSD grows into a reliable quote / lending / margin asset, and the CDP system keeps peg through redemptions and stability pools. The base case is that feUSD remains a useful but niche Hyperliquid stablecoin with modest supply, volatile TVL, and ecosystem-specific risk. The bear case is a HYPE drawdown, oracle failure, bridge / receipt failure for non-native collateral, thin Stability Pool coverage, or admin / smart-contract issue that breaks confidence faster than arbitrage can restore the peg.

Project Overview / What It Is

Felix is a DeFi protocol built around Hyperliquid-native lending, stablecoin borrowing, and trading products. The public site describes Felix as a modern financial account powered by stablecoins, with 24/7 derivatives trading, U.S. spot equities, and stablecoin lending markets. For this report, the relevant component is the CDP Market that issues feUSD. Felix documentation describes the CDP Market as a borrowing and stablecoin market built on Liquity v2, where borrowers deposit collateral and mint feUSD while Stability Pool depositors provide liquidation backstop liquidity.

The official CDP Market page identifies two main actors. Borrowers / minters deposit collateral such as HYPE, UBTC, and later liquid-staked assets, then mint feUSD and pay an upfront fee plus a borrower-selected interest rate. Stability Pool depositors provide feUSD liquidity that can be burned during liquidations. In exchange, they earn borrower interest, a share of upfront fees, and liquidation gains when undercollateralized positions are closed. This is closer to Liquity BOLD than to Aave GHO. It is not a pooled variable-rate lending market where a protocol algorithm sets a utilization curve. It is a set of collateral branches where borrowers choose the rate they are willing to pay and the system uses redemption priority to discipline underpriced debt.

FEUSD is the stablecoin. The token address in the Felix FAQ and developer documentation is 0x02c6a2fa58cc01a18b8d9e00ea48d65e4df26c70. The Market 1 developer docs list the same FEUSD address and the collateral registry at 0x9de1e57049c475736289cb006212f3e1dce4711b. The Hyperliquid Core metadata for the spot token also identifies a FEUSD token with full name Felix USD, token ID 0x88102bea0bbad5f301f6e9e4dacdf979, HyperCore index 241, and linked EVM contract 0x02c6a2fa58cc01a18b8d9e00ea48d65e4df26c70. This identity check is important because unrelated FELIX or FEUSD references can appear on aggregators, and this report does not mix them.

Felix has several adjacent products. Spot Equities are tokenized equities and ETFs accessed through Felix and routed through Ondo mint / redeem infrastructure. Felix Perpetual Futures are equity and commodity perpetual markets managed by the Felix team on top of Hyperliquid, with all settlement happening onchain on the Hyperliquid protocol. Hyperliquid's HIP-3 builder-deployed perpetuals explain the permissionless perp deployment model and the 500K HYPE staking requirement for deployers. These products matter to Felix as a brand and distribution layer, but they do not change the core FEUSD risk model. FEUSD is still a CDP stablecoin whose safety depends on collateral, redemptions, liquidations, oracle behavior, and stability-pool depth.

The basic user workflow is straightforward. A borrower holds HYPE, UBTC, or another supported collateral asset on Hyperliquid / HyperEVM. The borrower deposits collateral into Felix, chooses an amount of feUSD to borrow, selects an interest rate, approves transfers, and opens a position. The borrower can later repay feUSD to reduce debt, add collateral, withdraw collateral, adjust the interest rate, or close the position. If the position becomes undercollateralized, it can be liquidated, with Stability Pool feUSD burned against the debt and collateral distributed to pool depositors. If feUSD trades below peg, a redeemer can buy cheap FEUSD and redeem it for one dollar of collateral value, with debt cancelled starting from the lowest-rate positions.

The stablecoin has no offchain bank-account reserve in the way USDC or USDT do. Felix explicitly frames feUSD as over-collateralized, no centralized custodian, and dependent on liquidators and redeemers. This is the right design for a Hyperliquid-native DeFi dollar, but it gives users a different risk surface. The primary risks are collateral market risk, liquidation execution, Stability Pool solvency, oracle correctness, redemption pressure, smart-contract risk, and admin control. Those risks can be acceptable for active DeFi users. They are not the same as holding a short-term Treasury-backed stablecoin or a fully regulated fiat-backed instrument.

Product & Architecture / Mechanism

Felix feUSD architecture has four major modules: collateral branches, borrower positions, Stability Pools, and redemption / liquidation infrastructure.

Each collateral branch is independent. The Felix developer docs say protocol parameters are common across collateral branches but each branch is independent and does not share state. For WHYPE, the docs list a Zapper, Active Pool, Address Registry, Borrower Operations, Collateral Surplus Pool, Default Pool, Gas Pool, Price Feed, Sorted Troves, Stability Pool, Trove Manager, and Trove NFT. Similar module sets exist for UBTC, kHYPE, and wstHYPE-related branches. The independent-branch design is important because collateral risk does not necessarily contaminate every branch equally, but it also means each branch needs its own stability pool, oracle, caps, and liquidation health.

The borrower position is a Trove-like CDP. A borrower opens a trove by depositing collateral, minting FEUSD, and setting an annual interest rate. The Market 1 developer docs show functions such as openTrove, addColl, withdrawColl, withdrawFelix, repayFelix, adjustTroveInterestRate, and closeTrove. The borrower-set interest rate is not cosmetic. It determines debt cost, feeds Stability Pool yield, and determines redemption priority. Borrowers who set very low rates get cheaper debt but are first in line when FEUSD trades below peg and redemptions happen. Borrowers who set higher rates pay more but move later in the redemption queue.

The Stability Pool is the liquidation backstop. The Earning feUSD yield page says users deposit feUSD into Stability Pools to earn borrower interest and liquidation proceeds. During liquidation, deposited FEUSD is burned to cancel the borrower's debt, and the depositor receives the borrower's collateral at face value plus liquidation economics. This can be attractive when collateral remains liquid and the discount compensates for risk. It can be painful when HYPE, UBTC, kHYPE, or another collateral asset is falling quickly and depositors receive volatile assets that continue to decline. Stability Pool deposits are therefore not cash deposits; they are short convexity positions against collateral crashes.

Redemptions are the peg floor. The How it Works page explains that redemptions allow FEUSD holders to exchange FEUSD directly for one dollar of collateral at face value from the protocol, minus fees, processed against open positions in order of the lowest interest rate first. The FAQ specifies a redemption fee of min(0.5% + baseRate, 100%), with baseRate increasing with each redemption and decaying with a six-hour half-life. The same FAQ says redemptions can happen any time but are usually profitable when FEUSD trades below one dollar enough to cover the redemption fee. This means the peg is not maintained by a centralized issuer promising one-to-one fiat redemption. It is maintained by an arbitrage system that transfers collateral out of low-rate borrower positions.

Liquidations are separate from redemptions. A redemption can hit a healthy borrower position if FEUSD trades below peg and the borrower chose a low interest rate. A liquidation hits an unhealthy position when position health falls below threshold. The Minting / Borrowing feUSD page says position health below one triggers liquidation, with debt cancelled by burning FEUSD from the Stability Pool and collateral plus a penalty transferred to Stability Pool depositors. The FAQ says liquidations transfer borrower collateral plus a 5% penalty to pool depositors and that liquidators receive a small fee for calling liquidations.

Felix has an explicit contingency when Stability Pool liquidity is insufficient. The FAQ says liquidators can use just-in-time FEUSD to cover remaining debt and receive 105% of that value in collateral, or trigger redistribution where a liquidated position's debt and collateral are redistributed across other borrowers in the same collateral market in proportion to their collateral. This matters because empty Stability Pools are one of the main failure paths for CDP systems. If Stability Pool depth is strong, bad positions are absorbed cleanly. If it is weak, liquidation load can move to just-in-time liquidators or to remaining borrowers, which can create stress, confusion, and risk premium.

Oracle design is another key dependency. The Risk Management page says Felix uses RedStone Oracles as the primary pricing solution for crypto assets and is working with Anthias Labs on day-one risk monitoring and parameterization. The initial collateral table in the CDP overview lists HYPE at around 59% LTV and UBTC at 50% LTV, with RedStone as oracle. RedStone is a credible modular oracle provider, but an oracle dependency remains a dependency. Incorrect prices, delayed prices, liquidity mismatches between oracle reference markets and Hyperliquid execution markets, or oracle downtime can all affect liquidations and collateral solvency.

Admin controls must be treated as a real architectural input. The Smart Contract Audits page says Felix primarily changed three components from the Liquity v2 codebase: added mint caps, enabled trove parameter adjustments by admin, and enabled protocol pausing by admin. It also says Felix smart contracts retain upgradability upon deployment and that future changes before removal of admin privileges will undergo further auditing. This is not automatically bad. Young protocols often need caps and pauses to manage tail risk. But it means Felix is not an immutable, fully governance-minimized deployment today. Users are taking smart-contract risk plus admin-key / upgrade-process risk.

Mechanism walkthrough:

The design is coherent and battle-tested by lineage, but not risk-free. The right question is not whether Liquity v2 mechanics are elegant. They are. The right question is whether Felix's specific deployment has enough collateral diversity, oracle robustness, liquidity, stability-pool depth, and user discipline to survive a sharp Hyperliquid-specific stress event.

Token & Value Capture / Economics

FEUSD is a stablecoin, not a governance token or equity proxy. That single point prevents a lot of analytical confusion. The value of FEUSD should be approximately one dollar. There is no upside thesis in the token itself beyond holding a useful dollar instrument, earning yield through Stability Pools, and deploying the asset in Hyperliquid strategies. If Felix as a protocol grows, the immediate beneficiaries are borrowers with deeper liquidity, Stability Pool depositors with more interest flow, possible Felix protocol revenue recipients, integrations, and the broader Hyperliquid ecosystem. FEUSD holders do not automatically capture the upside of Felix growth as capital appreciation.

The economics are a three-sided market. Borrowers want cheap dollar liquidity against collateral. Stability Pool depositors want yield for taking liquidation and collateral exposure. FEUSD users want a stable, liquid dollar-like asset for trading, transfers, and strategy execution. Felix's job is to balance those sides. If borrower demand is high and Stability Pool demand is low, rates must rise or risk increases. If Stability Pool yield is attractive but borrower demand is weak, FEUSD supply stalls. If FEUSD secondary liquidity is weak, the peg can drift and redemptions can punish low-rate borrowers.

Borrower-set interest rates are the economic core. Felix docs say borrowers set their own fixed interest rate, and the FAQ advises users to compare against the median rate. Lower rates are cheaper but more exposed to redemptions. Higher rates cost more but reduce redemption priority. This creates a market-driven rate curve without governance setting a single borrow APR. Liquity's own BOLD and Earn documentation similarly emphasizes that user-set interest rates drive Stability Pool yield, while Liquity v2 borrowing docs explain that borrowers choose and adjust rates. Felix is importing that mechanism into Hyperliquid collateral markets.

Stability Pool yield comes from borrower interest, upfront fees, and liquidation proceeds. The Felix Earning feUSD yield page says users must deposit FEUSD into a Stability Pool to earn stablecoin yield, but also warns that their FEUSD can be burned during liquidations and converted into borrower collateral. DefiLlama's yield API showed a Felix CDP FEUSD pool with roughly 8.83M dollars TVL and about 7.53% APY on the June 29 snapshot. That is a meaningful yield signal, but it is not a risk-free stablecoin yield. It is compensation for acting as a liquidation backstop.

Protocol fees are real but modest. DefiLlama's fees overview showed the Felix CDP module with about 3.18K dollars in 24h fees, 22.45K in 7d fees, 83.99K in 30d fees, and 4.37M in one-year fees on the snapshot. DefiLlama's methodology says Felix CDP fees include borrow interest, redemption fees paid by borrowers, and liquidation profit, while protocol revenue is tied to borrow interest and supply-side revenue goes to Stability Pool participants and borrowers through fee mechanics. This is enough to prove the protocol is not empty. It is not enough to prove a mature cash-flow asset because FEUSD itself is not the protocol equity token.

The main "value capture" question is therefore split. Does the stablecoin capture utility value? Yes, if FEUSD becomes a convenient dollar collateral, quote asset, or liquidity leg in the Hyperliquid ecosystem. Does FEUSD capture protocol upside? No, not directly. Does Felix as a protocol capture value? Yes, through fees, product distribution, perps deployer economics, vaults, and related products, but that is not the same as FEUSD holder return. Does Hyperliquid / HYPE benefit? Potentially, because feUSD increases HYPE collateral utility and may create more onchain trading, borrowing, and fee activity.

This creates a subtle but important investment stance. FEUSD can be attractive for active users seeking Hyperliquid-native dollars or Stability Pool yield. It is less attractive for passive holders who simply want the safest possible dollar. A passive dollar holder would compare FEUSD against USDC, USDT0, USDe, USDHL, Aave GHO, BOLD, crvUSD, and other stablecoin options. FEUSD's differentiator is not scale or regulatory safety. It is native Hyperliquid composability and HYPE-backed credit creation.

The fee and incentive loop can be summarized:

The strongest bullish economic case is that FEUSD becomes the native borrowing dollar for HYPE holders. HYPE is the gas token, the main collateral, and the speculative asset around Hyperliquid's growth. If HYPE holders want leverage or liquidity without selling, Felix is a natural venue. The more HYPE collateral sits in Felix, the more FEUSD can circulate. The strongest bearish economic case is the same reflexivity in reverse. If HYPE falls, borrowers deleverage, liquidations rise, Stability Pools receive volatile collateral, and FEUSD demand may shrink exactly when confidence matters most.

Market & Traction

Data snapshot: June 29, 2026.

The first traction signal is TVL. DefiLlama's protocol API showed Felix TVL near 103.87M dollars, with the earliest listed point in February 2025 and current chain TVL categories including Hyperliquid L1 and Ethereum. The same snapshot showed TVL around 117.73M seven days earlier and 147.08M thirty days earlier, implying a material contraction over the month. This does not mean the protocol is failing. Stablecoin CDP TVL can fall when collateral prices move, borrowers repay, incentives rotate, or users migrate to other Hyperliquid strategies. But it means the current trend is not cleanly up-only.

The second signal is FEUSD supply and peg. DefiLlama's stablecoin API showed Felix feUSD as a crypto-backed, USD-pegged asset on Hyperliquid L1 with price around 0.99976 and circulating supply around 14.09M dollars. A direct totalSupply() call to the FEUSD ERC-20 contract on HyperEVM returned approximately 14.09M FEUSD. That alignment between an onchain EVM contract read and DefiLlama's stablecoin supply is the strongest supply anchor in this memo. It says the visible EVM supply is real and modest.

The third signal is HyperCore circulation. Hyperliquid's spotMetaAndAssetCtxs API identified FEUSD as HyperCore token index 241, full name Felix USD, with linked EVM contract address matching the Felix docs. The FEUSD/USDC spot market had a mid price around 0.9987, previous-day price around 0.9983, and daily notional volume around 529K. This is important because a stablecoin on Hyperliquid may circulate not only as an ERC-20 on HyperEVM but also through HyperCore spot mechanics. The market data suggests there is at least some real orderbook usage beyond small AMM pools.

The fourth signal is DEX liquidity. DexScreener's token endpoint showed FEUSD pools on HyperEVM with prices clustered near one dollar, including FEUSD/WHYPE, FEUSD/USDT0, FEUSD/USDXL, FEUSD/USDHL, and FEUSD/USDC pairs across HyperSwap, Project X, Kittenswap, Laminar, Hybra, and other venues. The problem is depth. At the time of the terminal snapshot, visible individual pool liquidities were mostly in the range of roughly 1K to 24K dollars, with Project X FEUSD/USDT0 near 19.9K, HyperSwap FEUSD/WHYPE near 23.6K, and HyperSwap FEUSD/USDT0 near 12.9K. Search-result snapshots for the same DexScreener pages can show larger short-term volume spikes, but the consistent point is that AMM depth is thin relative to the risk one would take holding millions of FEUSD.

The fifth signal is fee generation. DefiLlama's fee overview showed Felix CDP one-year fees near 4.37M and all-time fees near 6.46M, with 30-day fees near 83.99K. This confirms that borrowers have paid for CDP usage. It also shows how cyclical the economics can be. One-year fees look material relative to current FEUSD supply, but current 30-day fees are modest, and the TVL downtrend suggests the recent operating rate may be lower than the best historical period.

The sixth signal is ecosystem breadth. Felix's own docs include spot equities, vanilla markets, perps, and CDP stablecoin products. A Hyperion DeFi press release about a strategic partnership with Felix described Felix as a HyperEVM protocol that had crossed one billion dollars in TVL across borrow / lend in September 2025, launched Felix CDP first, and scaled to tens of millions of FEUSD debt and over 250 million dollars in deposited collateral before expanding into other products. This is a company-issued partner press release rather than a neutral source, so it should not override current DefiLlama data, but it supports the view that Felix was a major Hyperliquid ecosystem credit project during 2025.

The seventh signal is risk-adjusted yield. DefiLlama's yield list showed Felix CDP FEUSD pool TVL near 8.83M dollars and APY near 7.53%. That is large compared with visible AMM depth and smaller than total Felix TVL, which indicates that Stability Pool deposits are meaningful but not obviously overbuilt relative to total collateral risk. If FEUSD supply is around 14.09M, a Stability Pool around 8.83M is a significant backstop. The risk is branch-specific: a global number can hide whether the HYPE pool, UBTC pool, or kHYPE pool is adequately covered.

Source Conflict Matrix

The overall traction read is "real but early." Felix has meaningful TVL, real fee history, a documented stablecoin mechanism, Hyperliquid-native integration, and nontrivial Stability Pool deposits. It also has a small current stablecoin supply, thin visible AMM liquidity, recent TVL contraction, and supply / market-cap discrepancies across data sources. For a stablecoin, those caveats matter more than narrative.

Collateral, Reserve, Peg, and Redemption Mechanics

Felix feUSD is not backed by cash reserves. It is backed by overcollateralized borrower positions and supported by a Stability Pool and redemption mechanism. That makes "reserve" analysis different from USDC or Tether. There is no bank account where every FEUSD maps to a dollar claim. There are collateral vaults, debt positions, stability pools, liquidators, redeemers, oracles, and smart contracts.

The initial collateral list in the CDP overview shows HYPE and UBTC, with HYPE around 59% LTV and UBTC at 50% LTV, using RedStone oracles. The developer docs add deployed parameters for WHYPE, UBTC, kHYPE, and a wstHYPE branch marked coming soon. The most important current collateral is HYPE / WHYPE because HYPE is the native gas and economic asset of Hyperliquid. This gives Felix a natural user base: HYPE holders who want dollar liquidity without selling. It also creates reflexive risk: if HYPE price falls sharply, collateral value falls, borrower health deteriorates, liquidations rise, and FEUSD confidence can weaken at the same time.

UBTC is different. UBTC represents Unit Bitcoin on Hyperliquid, with the developer docs listing the UBTC address and a Felix UBTC wrapper for decimals. BTC-like collateral is generally less reflexive to Hyperliquid-specific risk than HYPE, but it introduces bridge / receipt / wrapper risk. Felix's FAQ explicitly discusses bridge receipt failure as a general protocol risk: assets aside from HYPE and PURR are not native to Hyperliquid, and a failure of bridged versions of otherwise trusted assets could cause abrupt value loss. That warning should be taken seriously. A stablecoin can be overcollateralized and still impaired if the collateral representation fails.

Redemptions are the main peg-restoration tool when FEUSD trades below one dollar. A user can buy FEUSD below peg and call redeemCollateral, as listed in the developer docs, burning FEUSD and receiving one dollar of system collateral minus fees. This reduces FEUSD supply and creates buying pressure. It also cancels borrower debt from the lowest interest-rate positions first. The mechanism has two effects: it gives arbitrageurs a reason to lift the peg, and it pushes borrowers to set interest rates high enough that their debt is not the cheapest to redeem.

The redemption threshold is economic rather than mechanical. The docs say redemptions can happen at any time, but are likely when profitable, usually when FEUSD price is less than one dollar minus the current redemption fee. The user docs give a practical warning that FEUSD trading at or below 0.995 can trigger redemption pressure. That matters for borrower behavior. A borrower who chooses an aggressive low rate is effectively saying, "I accept being first in line if the peg breaks." In calm markets, this can reduce debt cost. In stress, it can force partial deleveraging.

Stability Pools handle liquidations, not ordinary redemptions. If a borrower position falls below its threshold, FEUSD in the relevant Stability Pool is burned to cancel debt and collateral is distributed to pool depositors. If collateral continues falling after depositors receive it, depositors can lose money. Felix's docs explicitly warn about liquidation exposure and bad-debt shortfall for Stability Pool depositors. That is good disclosure. It also means FEUSD yield should not be compared casually with Treasury yield. It is closer to insurance underwriting against borrower undercollateralization.

The peg can fail in several ways. First, secondary liquidity can dry up before redemptions are economically attractive. Second, gas, UX, or collateral withdrawal frictions can reduce arbitrage speed. Third, if HYPE or UBTC collateral is falling rapidly, redeemers may not want the collateral they receive. Fourth, if Stability Pool depth is thin, liquidations can move to JIT or redistribution paths that users may not understand. Fifth, if oracle prices lag or diverge from executable collateral prices, liquidations can occur too late. Sixth, if the protocol is paused, upgraded, or affected by an admin action during stress, users may distrust the peg.

The best case is that FEUSD becomes a stable, deeply integrated Hyperliquid dollar. Borrowers set rational rates, Stability Pool depositors demand adequate yield, redemptions are rare but credible, and HyperCore / HyperEVM liquidity lets users enter and exit near one dollar. The worst case is a synchronized HYPE drawdown where borrowers rush to repay or get liquidated, Stability Pool depositors receive falling collateral, secondary FEUSD sells off, redeemers hesitate, and the protocol must rely on admin controls. The current evidence supports "credible but still young," not "battle-tested at scale."

Hyperliquid / HYPE Ecosystem Relationship

Felix's relationship with Hyperliquid is not superficial. It is the main reason the protocol matters. Felix is built for the Hyperliquid environment, uses HYPE / WHYPE as a core collateral asset, issues a stablecoin linked to Hyperliquid token infrastructure, and operates products that depend on Hyperliquid's orderbook and builder-deployed market design.

Hyperliquid is unusual because it combines a high-performance trading venue, HyperCore orderbooks, and HyperEVM smart contracts. The official HyperEVM docs say HyperEVM blocks are built as part of Hyperliquid execution and inherit security from HyperBFT consensus. HYPE is the native gas token on HyperEVM. Hyperliquid's dual-block architecture separates fast small blocks from slower large blocks to balance confirmation speed with large contract deployment. For Felix, this means the protocol can be an EVM lending application while still living next to a centralized-exchange-like onchain orderbook.

The Core / EVM link is critical. Hyperliquid's Interacting with HyperCore docs describe precompiles and a CoreWriter system contract for sending actions from HyperEVM to HyperCore. Hyperliquid's HyperCore <> HyperEVM transfers docs explain how linked tokens can move between Core and EVM. Felix's FAQ gives a practical FEUSD transfer instruction to send FEUSD to 0x20000000000000000000000000000000000000f1 to transfer to HyperCore. This is the distribution edge: FEUSD can be useful both as an EVM stablecoin and as a HyperCore market asset.

HYPE collateral is the product-market-fit hook. Hyperliquid users often hold HYPE because it is the native ecosystem asset, gas token, staking asset, and speculative claim on Hyperliquid growth. Felix lets those users borrow a dollar asset without selling HYPE. This is exactly the type of loop that made MakerDAO and Liquity relevant on Ethereum: convert volatile asset collateral into stablecoin liquidity. The difference is that Hyperliquid is more exchange-centered and HYPE is more reflexive to the venue's growth narrative.

Felix also extends into Hyperliquid's builder-deployed market story. Felix Perpetual Futures documentation says the product is built on top of Hyperliquid and enables equity / commodity perpetual futures. Hyperliquid's HIP-3 docs explain that builder-deployed perps require deployers to maintain 500K staked HYPE, set market definitions and oracle specifications, and operate markets. This is not the same as feUSD CDP, but it matters because it shows Felix is trying to become a broader financial account on Hyperliquid, not only a stablecoin issuer.

The ecosystem relationship is a strength and a concentration risk. It is a strength because Hyperliquid has deep trader mindshare, HYPE collateral demand, native orderbooks, fast execution, and a user base that wants leverage. It is a concentration risk because Felix depends heavily on Hyperliquid infrastructure, HYPE price health, HyperEVM reliability, HyperCore bridging / token linkage, and ecosystem liquidity. A protocol on Ethereum can sometimes diversify across many venues and collateral types. Felix's current edge comes from being close to Hyperliquid; its current risk also comes from being close to Hyperliquid.

Team, Funding, Governance, and Security

Felix's public docs do not present a deep founder biography or tokenholder governance framework in the way some DAO protocols do. The protocol should therefore be evaluated primarily through shipped product, code lineage, audits, risk disclosures, and admin controls.

The base code lineage is Liquity v2. Felix's audit page says the Liquity v2 codebase has undergone audits by ChainSecurity, Dedaub, Recon, Coinspect, and Three Sigma, and formal verification by Certora. The page links to ChainSecurity's Liquity BOLD audit, Dedaub's Liquity v2 audit, Certora's formal verification report, Coinspect's BOLD audit, and related governance audits. That is a stronger base than a fully novel CDP implementation.

Felix-specific changes are where residual risk lives. The same audit page says Felix changed three components from Liquity v2: mint caps, admin-adjustable trove parameters, and admin pausing. These changes are practical for a new deployment with new collateral on a new chain. They are also centralization and upgrade risks. A mint cap can prevent overgrowth during risk events. An admin parameter adjustment can respond to collateral volatility. A pause can stop damage. But every such lever creates questions: who controls it, what multisig or governance process exists, how quickly can it be used, how transparent are decisions, and when will admin privileges be removed?

Three Sigma's Felix Protocol case study describes Felix as a Hyperliquid L1 CDP and shows an audit period in July 2025 with no critical or high issues and two medium issues. That is constructive third-party signal. It should be read alongside Felix's own statement that the protocol is not yet battle-tested and that contracts retain upgradability upon deployment. The right risk posture is "audited but young," not "audited therefore safe."

Risk monitoring is explicitly part of the design. The Risk Management page says Felix works with Anthias Labs on day-one risk monitoring and parameterization to manage bad-debt risk. This is important because collateral parameters on Hyperliquid are not static. HYPE liquidity, UBTC liquidity, LST market depth, and Stability Pool coverage change over time. A well-run CDP needs ongoing risk management, not only initial audits.

The absence of a clear FEUSD governance token also matters. Some protocols distribute governance and fee-switch decisions to tokenholders. FEUSD itself is not that instrument. The stablecoin holder's protection comes from mechanism, code, risk controls, and market discipline, not from governance rights. If Felix later has or expands a separate governance / equity / incentive token, that would be a different investment question. This report does not treat FEUSD as that upside claim.

Competitive Landscape

Felix feUSD competes in three overlapping markets: Hyperliquid-native dollars, crypto-backed CDP stablecoins, and yield-bearing stablecoin strategies.

Inside Hyperliquid, the main competitors are USDC, USDT0, USDHL, USDe, USDXL, and any future aligned quote assets. USDC remains the default high-liquidity dollar in many crypto venues. USDT0 offers Tether-aligned omnichain dollar distribution. USDHL is a Hyperliquid-native or ecosystem-native dollar product with a different yield / reward model. USDe brings Ethena's synthetic dollar model. FEUSD's edge is not scale. Its edge is being minted directly against Hyperliquid collateral and integrated into Felix lending / Stability Pool mechanics. Its weakness is that it has less global liquidity and less track record than USDC or USDT.

Against Liquity BOLD, Felix is an implementation / adaptation rather than a completely different primitive. Liquity v2's BOLD docs and borrowing docs emphasize user-set rates, stablecoin redemptions, and Stability Pool yield. Felix brings those mechanics to Hyperliquid collateral. The advantage is ecosystem fit. The disadvantage is that Felix inherits design complexity while using younger collateral markets and younger chain infrastructure.

Against Aave GHO, Felix has more market-driven borrower rates but less broad collateral governance and less Aave-scale liquidity. Aave's GHO docs describe GHO as minted by supplying approved collateral into Aave and borrowing against it. GHO benefits from Aave's brand, collateral base, and governance. FEUSD benefits from Hyperliquid-native positioning and redemption mechanics. GHO is more protocol-treasury style; FEUSD is more Liquity-style redemption and Stability Pool insurance.

Against Curve crvUSD, Felix has simpler liquidation mechanics but less mature market infrastructure. Curve's crvUSD overview describes a CDP-style stablecoin with LLAMMA soft-liquidation design, where collateral can be progressively converted. Felix uses Stability Pool liquidation and redemptions. crvUSD's advantage is Curve's liquidity infrastructure. Felix's advantage is Hyperliquid-native collateral and trader distribution.

Against Ethena USDe, the risk model is entirely different. Ethena's USDe docs describe a synthetic dollar backed by crypto assets and delta-neutral hedges. FEUSD is not delta-neutral and not exchange-hedged. It is overcollateralized debt. USDe scales through centralized exchange hedging and collateral operations; FEUSD scales through borrower demand and collateralized lending. USDe has funding / custody / exchange counterparty risks. FEUSD has collateral, liquidation, redemption, oracle, and Stability Pool risks.

FEUSD's defensibility is local rather than global. It is unlikely to beat USDC or USDT as a universal dollar. It does not need to. It needs to be a credible dollar inside Hyperliquid for HYPE-backed borrowing, Stability Pool yield, and ecosystem trading. That is a narrower but realistic target.

Catalysts

The first catalyst is collateral expansion with discipline. Felix docs mention HYPE, UBTC, kHYPE, and wstHYPE-related branches. Adding liquid-staked HYPE or other Unit assets can increase borrow demand and system scale. The bullish version is diversified collateral with conservative caps, robust oracle feeds, and deep Stability Pools. The bearish version is rapid onboarding of reflexive or bridge-risk collateral before liquidity and risk controls mature.

The second catalyst is FEUSD becoming an aligned or high-quality quote asset on Hyperliquid. Hyperliquid's permissionless spot quote asset docs define requirements around peg quality, quote / USDC orderbook depth, HYPE / quote liquidity, and a 200K HYPE staked deployer requirement for quote assets. If FEUSD can satisfy high-quality quote-asset conditions or become more deeply embedded in HyperCore markets, utility increases. If it cannot maintain depth and peg tightness, it remains a niche stablecoin.

The third catalyst is Stability Pool depth growth. FEUSD supply is only safe if liquidation backstop liquidity is adequate and branch-specific. A larger, healthier Stability Pool lowers bad debt risk and can make borrower rates more competitive. The key metric is not only aggregate SP TVL. It is SP coverage relative to debt in each collateral branch, especially HYPE.

The fourth catalyst is Hyperliquid ecosystem growth. If HYPE trading, HyperEVM apps, spot markets, and HIP-3 perps keep expanding, demand for native dollar liquidity can rise. Felix is positioned to benefit because it gives HYPE holders a way to borrow without selling. The reverse is also true: if Hyperliquid activity contracts, FEUSD demand can contract.

The fifth catalyst is admin-risk reduction. Felix docs say contracts retain upgradability upon deployment and that the protocol has plans to become immutable in the future. A credible timeline, multisig disclosure, timelocks, parameter transparency, and eventual removal of admin privileges would improve confidence. A surprise pause or opaque upgrade would reduce confidence.

The sixth catalyst is data clarity. FEUSD has supply and market-cap conflicts across DefiLlama, onchain ERC-20 totalSupply, DexScreener, CoinGecko-style fields, and HyperCore token metadata. A public dashboard from Felix that reconciles EVM supply, HyperCore supply, Stability Pool deposits, collateral by branch, debt by branch, redemption history, and bad debt would materially improve investability.

Risk Matrix

The largest risk is a synchronized ecosystem shock. FEUSD is useful because it is native to HYPE and Hyperliquid. That same native exposure means a Hyperliquid-specific drawdown can hit collateral value, stablecoin confidence, liquidity depth, and user risk appetite together. The second-largest risk is that FEUSD appears stable until size matters. A small user can exit through available liquidity near one dollar. A large user may need redemption or HyperCore depth. The third risk is operational: admin controls and upgradability are rational at launch but must be disclosed and eventually minimized.

Valuation / Importance Framework

FEUSD should not be valued with FDV upside logic. It is designed to trade at one dollar. The relevant framework is importance, safety, liquidity, and yield-adjusted risk.

Importance is medium-high inside Hyperliquid and low-to-medium globally. Within Hyperliquid, FEUSD gives HYPE holders a way to borrow dollar liquidity, gives Stability Pool depositors a yield product, and potentially gives HyperCore markets another quote / stable asset. Globally, FEUSD is tiny compared with USDC, USDT, DAI / USDS, USDe, or even many mid-sized stablecoins. Its market role is local.

Safety is medium-low to medium. The Liquity v2 lineage and audits are positive. Overcollateralization, redemptions, and Stability Pools are serious mechanisms. But Felix-specific admin controls, new chain infrastructure, young collateral markets, oracle dependency, bridge collateral, and limited battle-testing reduce confidence. FEUSD is safer than an undercollateralized algo-stable with no redemption mechanism. It is not as safe as a top-tier fully reserved fiat stablecoin for passive cash management.

Liquidity is medium-low. HyperCore FEUSD/USDC volume around 529K daily is useful. Visible AMM liquidity in individual pools is thin. Stablecoin liquidity quality should be measured by executable exit size at 0.1%, 0.5%, and 1% slippage, not by nominal market cap. Current data does not support treating FEUSD as deep institutional collateral.

Yield-adjusted risk is situational. A roughly 7.5% FEUSD Stability Pool APY can be attractive if HYPE collateral risk is acceptable, branch coverage is strong, and the user is active enough to manage received collateral. It is unattractive if the user wants passive stablecoin yield without exposure to liquidations and bad debt. The yield is compensation for underwriting system risk.

Practical sizing framework:

The investable conclusion is that FEUSD should be evaluated like a credit instrument in a young ecosystem. It has utility, but users should demand a risk premium, monitor peg and liquidity, and avoid confusing stable price target with low risk.

Bull / Base / Bear Scenarios

Bull case: Felix becomes the Maker / Liquity-style credit layer for Hyperliquid. HYPE holders trust the system, borrow FEUSD responsibly, and keep collateral ratios conservative. Stability Pool depositors deepen coverage because borrower rates offer attractive real yield. FEUSD gains better HyperCore orderbook depth and becomes more useful as a quote or margin-adjacent stablecoin. Felix reduces admin risk over time and publishes stronger dashboards. In this case, FEUSD does not appreciate, but its utility and safety improve.

Base case: FEUSD remains a useful but specialized Hyperliquid stablecoin. It circulates for borrowing, Stability Pool yield, and some trading routes. Supply fluctuates with HYPE price and incentives. USDC and other stablecoins remain dominant for deep liquidity. FEUSD is fine for active users but not the default treasury asset.

Bear case: A sharp HYPE drawdown reveals weak Stability Pool coverage or thin secondary liquidity. Low-rate borrowers are heavily redeemed, HYPE liquidations accelerate, FEUSD trades below peg, and users discover that exits require redemption into volatile collateral. Even if the protocol remains solvent, confidence can fall and supply can contract sharply. A smart-contract bug, oracle error, bridge receipt failure, or opaque admin action would worsen the outcome.

Confidence Score

Confidence is medium because the mechanism is understandable and evidence is sufficient, but the operating history and liquidity are still developing. I would raise confidence if Felix publishes a live risk dashboard with branch-level debt, collateral, SP coverage, redemptions, liquidations, oracle status, and admin controls. I would lower confidence if supply data remains inconsistent, FEUSD trades below peg for extended periods, or the protocol relies on admin pause / upgrades during stress.

Red-team Check

The strongest reason the positive thesis could be wrong is that FEUSD may be more reflexive than it looks. In calm markets, overcollateralized HYPE-backed borrowing works well. In stress, the same users who borrowed against HYPE may be forced to repay or be liquidated while HYPE price falls, FEUSD liquidity thins, and Stability Pool depositors receive collateral they do not want. The mechanism is elegant, but elegance does not remove market microstructure risk.

The most gameable metric is TVL. Felix can have high TVL because HYPE collateral appreciates or because users loop collateral, but that does not prove stablecoin safety. For FEUSD, better metrics are branch-level collateralization, debt by collateral type, Stability Pool coverage by branch, redemption activity, realized liquidation losses, and executable FEUSD exit depth. A protocol can show large TVL while stablecoin liquidity remains thin.

The value-capture failure path is that Felix grows as a protocol but FEUSD remains only a utility stablecoin with no upside and nontrivial risk. This is not a failure for the product; it is a failure for investors who misunderstand the asset. FEUSD holders are not buying a share of Felix. They are taking stablecoin and protocol risk. Stability Pool depositors earn yield but underwrite liquidation events. Borrowers get leverage but face redemption and liquidation.

The zero or permanent impairment path is a combination of collateral failure, oracle failure, and liquidity failure. If HYPE or UBTC collateral crashes, oracle prices lag, Stability Pool deposits are insufficient, and secondary FEUSD trades below peg, users may rush to redeem or exit. If admin actions are needed at the same time, trust can deteriorate. Even if the protocol ultimately covers debt, the stablecoin could shrink, integrations could leave, and users could migrate to USDC / USDT0 / USDHL.

The most important red-team conclusion is that FEUSD is not a generic stablecoin. It is a Hyperliquid-native credit instrument. That is what makes it useful; that is also what makes it dangerous if the Hyperliquid ecosystem hits a concentrated drawdown.

Monitoring Dashboard

Follow-up Triggers

Conclusion / Final Investment View

Final view: watchlist / selective use for active Hyperliquid users; not a passive core stablecoin allocation.

Felix feUSD is a serious attempt to build a native credit layer for Hyperliquid. The design is coherent: HYPE and other collateral back FEUSD debt, borrower-set interest rates price capital, Stability Pools absorb liquidations, redemptions help defend the peg, and HyperCore / HyperEVM integration gives the stablecoin a natural distribution path. The protocol has meaningful TVL, real fee history, official documentation, Liquity v2 lineage, audit references, and a clear reason to exist inside the HYPE ecosystem.

The risk is also clear. FEUSD is young, small, and ecosystem-concentrated. Its peg depends on secondary liquidity, redemption arbitrage, Stability Pool depth, oracle correctness, and collateral quality. The largest collateral narrative is HYPE, which creates strong product-market fit but also reflexive downside. Data providers disagree on market-cap / supply fields, which makes conservative sizing important. Admin controls and upgradability are disclosed, but they mean the system is not yet fully immutable.

For a borrower, Felix can be attractive if the position is conservative, the interest rate is not recklessly low, and the borrower understands redemption risk. For a Stability Pool depositor, the yield can be attractive if they are willing to receive volatile collateral in liquidations and actively manage it. For a passive stablecoin holder, FEUSD should be sized below USDC / USDT-style cash instruments unless there is a specific Hyperliquid use case.

The one-sentence conclusion: FEUSD is a useful Hyperliquid-native dollar primitive, but it should be treated as collateralized credit infrastructure with peg and liquidation risk, not as a generic low-risk stablecoin or an upside token.