Frax v3: Fully Collateralised, RWA-Backed, and the Frax Universe
Executive summary
Frax v3 is the most consequential pivot in the algorithmic-stablecoin literature: a public, deliberate retreat from fractional-algorithmic backing to fully collateralised RWA-anchored design, executed by the team that had argued most loudly for the algorithmic model. The paper is half technical specification and half post-mortem, reflecting on what the 2022-2023 stablecoin collapses (Terra, sUSD partial depegs, Iron, MIM) had taught the team about reflexive instability. It introduces the AMO (algorithmic market operations) framework as the new control surface, formalises the FraxBP and Curve-pool dependency graph, and announces Fraxchain and frxETH as the productisation vectors. DI rates the paper a seven on technical novelty and a nine on epistemic honesty: very few stablecoin teams have published a document this candid about prior model failure.
Background and motivation
Frax launched in 2020 as the first credibly fractional-algorithmic stablecoin: FRAX was partially backed by USDC and partially backed by FXS (the governance and value-accrual token), with the collateralisation ratio adjusting algorithmically based on FRAX's market price. When FRAX traded above $1, CR decreased (more algorithmic backing); when below, CR increased (more USDC backing). The model was elegant in theory and worked passably in practice for the 2020-2021 bull market. The 2022-2023 stablecoin collapses changed the strategic landscape decisively. Terra's UST imploded in May 2022, vaporising $40B+ in days and discrediting the pure-algorithmic model. Iron Finance, Basis Cash, MIM, and other partial-algorithmic systems suffered severe stress events. Even Frax, which never lost peg, saw its CR drift from 90% to 100% as governance voted to over-collateralise in response to market fear. The fractional model had become a liability. Sam Kazemian announced the v3 transition publicly in 2023 with a frankness that was remarkable for the space: the team had concluded that algorithmic backing under stress was a reflexive amplifier rather than a stabiliser, and that future Frax design would assume worst-case correlated drawdowns. The v3 documentation is the paper trail of that pivot. It is structured less as a single whitepaper than as a constellation of FIPs, documentation pages, and technical addenda; this review treats them as a coherent corpus reflecting a single design intent. The motivation section of v3 explicitly cites the cost of dollar-loose money (USDC's 2023 SVB-related depeg, Tether's perpetual reserve-quality questions) as the macro context: stablecoins must carry their own yield to compete with USD-denominated alternatives, not merely peg to dollars while paying zero. This framing — stablecoin as yield instrument rather than payments rail — is the strategic core of v3.
Core technical contributions
The AMO framework is the technical crown jewel of v3. AMOs are smart-contract modules with permission to mint or burn FRAX into specific protocol activities — Curve LP pools, Aave lending, Frax Lend, MIM-style strategies — without changing the system's balance-sheet collateralisation. Each AMO must be approvable by veFXS governance and must provably maintain CR after operations. The effect is that FRAX can deploy substantial protocol-owned liquidity (POL) to defend its peg, capture LP fees, or earn lending yield, all while keeping the dollar peg intact and the backing fully collateralised. The AMOs that have shipped include Curve AMO (FraxBP liquidity provision), Lending AMO (Aave/Compound deposits), Liquidity AMO (Uniswap v3 concentrated positions), and the Investor AMO (RWA T-bill exposure via partner custodians). Reasoning about Frax's collateralisation in 2026 requires reasoning about the AMO graph rather than a static reserve account. frxETH is the second major contribution. Frax operates Ethereum validators with ETH deposited by users; frxETH is the receipt token, sfrxETH is the auto-compounding yield-bearing variant. The split is intentional: frxETH does not earn yield directly — instead, validator revenue is concentrated into sfrxETH holders, producing a higher APR than competing LSTs (Lido stETH, Rocket Pool rETH). frxETH is then deeply integrated into FRAX's Curve liquidity and used as gas for Fraxchain. This three-way integration (validator revenue → sfrxETH yield → FRAX peg subsidy via Curve LPs) is the deepest revenue-internalisation in the stablecoin sector. FXBs are zero-coupon FRAX bonds: users buy them at a discount, redeem at par at maturity, and the discount represents the prevailing FRAX-denominated rate. They give Frax a native yield-curve and provide on-chain duration without external counterparties.
Methodology and rigor
The Frax v3 documentation corpus is organised more like protocol engineering than academic research. The FIP series is rigorous in mechanism specification — each AMO has a documented integration with provable invariants — but the broader strategic claims (RWA-backed is structurally safer than algorithmic; AMO yield can subsidise peg defence indefinitely; Fraxchain capture economics will reward FXS holders) are asserted rather than modelled. There is no formal stress-test of the v3 design under correlated drawdown of all AMO venues simultaneously. The CR-equivalent calculations are static rather than dynamic: they assume each asset on the balance sheet trades at its quoted price and does not account for liquidation slippage if Frax needed to unwind AMO positions rapidly. This is the same methodological gap that pre-2022 stablecoin papers had, although the new fully-collateralised stance makes it less acute. Where the documentation is strongest is in operational transparency. Frax publishes a detailed monthly balance sheet, AMO-by-AMO performance attribution, and FXS revenue distribution — a level of disclosure that exceeds most stablecoin issuers and many traditional financial institutions. The forensic clarity makes external analysts' job possible in a way that USDT or USDC documentation frequently does not. The v3 paper's epistemic move worth highlighting: it explicitly retires the original algorithmic claims rather than rationalising them. Many stablecoin teams that survived 2022 quietly pivoted without acknowledging that their original design had been wrong. Frax says, plainly, that the fractional-algorithmic model was a mistake under stress and that v3 is a different system. This is methodologically unusual and intellectually credible.
Strengths
The v3 design's central strength is integrated revenue capture. Where most stablecoins (USDT, USDC, DAI, USDe) generate revenue from a single source — reserve T-bill yield, basis trade, RWA collateral — Frax generates revenue from at least four orthogonal venues: AMO LP fees, sfrxETH validator revenue, FXB yield-curve trading, and Fraxchain sequencer fees once that chain matures. Total protocol revenue per dollar of FRAX outstanding is the highest among major stablecoins by a meaningful margin in 2026. That revenue funds peg defence in a way that purely-collateralised stablecoins cannot match. The frxETH dual-token design has produced empirically higher LST yields than Lido's stETH for users willing to accept the centralisation tradeoff (Frax's validator set is smaller and more concentrated). Whether that tradeoff is worth the yield premium is an open user-level question, but the mechanism design is internally coherent and produces measurable yield differential. The retreat from algorithmic backing — uncomfortable as it must have been for the team — is itself a strength. Frax did not lose users during the transition; it gained them, because the new design was more legible and less reflexive. This kind of public strategic pivot is rare in DeFi, and Frax's ability to execute it without a death-spiral is evidence of underlying community trust. FraxBP's displacement of 3pool as the primary stablecoin Curve venue is a meaningful market-structure achievement: it gave Frax control over the gravitational center of stablecoin liquidity routing, with downstream effects on every Curve-dependent protocol.
Weaknesses and limitations
The principal weakness of v3 is concentration risk in the AMO portfolio. AMOs deploy FRAX into productive venues, but those venues are themselves DeFi protocols with their own smart-contract, oracle, and counterparty risks. A serious exploit at Aave, Curve, or Uniswap v3 — all of which are AMO destinations — would impair FRAX's collateral indirectly. The CR calculation does not haircut for these tail risks. The Curve depeg of July 2023, when CRV's price decline triggered a near-cascade across multiple stablecoin pools, was a near-miss: Frax was significantly exposed to Curve LP positions and would have absorbed loss had the cascade gone fully reflexive. Second, frxETH's yield premium over stETH is a function of validator-set concentration, which is itself a centralisation vector. If Frax's validators were censored, slashed, or compromised, sfrxETH would underperform — and frxETH's wider DeFi integrations (as Fraxchain gas, as Curve LP collateral) would propagate that underperformance into FRAX's broader balance sheet. Third, FXB has not achieved meaningful adoption as a yield product; redemption is forward-only, with no secondary market beyond shallow Curve liquidity. The intent — a native on-chain yield curve — has not been realised. Fourth, Fraxchain's commercial execution has been slower than announced, with mainnet activity well below the original 2024 projections. The sequencer-revenue economics that were supposed to redirect to FXS holders are smaller than expected. Fifth, FXS economics are still tightly coupled to FRAX growth; if FRAX supply stagnates, FXS revenue stagnates, and the vote-escrow incentives that anchor governance participation weaken. The endogenous-revenue model is impressive but has not yet been tested under a multi-quarter stagnation.
Subsequent influence
Frax's AMO framework has been the most directly influential v3 contribution. Liquity v2's PIL (protocol-incentivised liquidity) module borrows the AMO pattern conceptually. Curve's crvUSD design uses an AMO-like LLAMMA mechanism to manage collateral. MakerDAO's Endgame-era Allocator Vaults are a structural cousin, deploying DAI into yield venues with collateral-ratio invariants. The intellectual lineage is clear: protocol-owned liquidity deployed via constrained mint/burn modules has become a standard stablecoin tool, and Frax shipped the cleanest reference implementation. The frxETH dual-token model has been less imitated, partly because LST market structure had stabilised around stETH by the time frxETH launched, but partly because the validator-set centralisation tradeoff was uncomfortable for most teams. Lido's eventual move toward DVT (distributed validator tech) is a structural rejection of frxETH's design; whether DVT or frxETH-style concentration produces better long-run yield-versus-decentralisation outcomes is an open question. Beyond mechanism design, Frax's epistemic honesty about the algorithmic-model failure has been cited approvingly in commentary on stablecoin design and used as a counter-example by Terra-style failure post-mortems. The documentation corpus has become a standard reading assignment in DeFi research curricula.
How it holds up in 2026
Two years after v3 publication, the Frax design has held up better than the team's prior fractional model would have. FRAX has maintained peg through multiple stress events (March 2023 USDC depeg, July 2023 Curve incident, 2024 Q4 yield-curve compression), and total protocol revenue has grown roughly 4x over the period. The v3 design's RWA exposure (sFRAX-style T-bill yield) has captured the elevated rate environment effectively, with sFRAX yields tracking benchmark T-bills minus a thin protocol take. Where v3 has underperformed is in productisation. Fraxchain has not become a major appchain; sequencer revenue has been modest. FXB has not become a yield-curve standard. frxETH has retained its yield-premium niche but has not displaced stETH as the dominant LST. Frax's market share in stablecoins has been stable rather than growing — circulating FRAX is in the same band as 2023, while USDT and USDC have continued to grow in absolute terms. Whether this constitutes failure depends on the counterfactual: a Frax that had not pivoted from fractional backing in 2023 would, by the team's own admission, have been at significant risk during the 2024 yield-curve volatility. The pivot was protective. DI's holistic assessment: Frax v3 is the highest-quality stablecoin documentation corpus published in the space, with an unusual combination of mechanism rigor, operational transparency, and epistemic candor. The design has held up technically; the productisation thesis has been partly vindicated and partly stalled. As a research artefact, a clear seven; as a strategic playbook for stablecoin teams considering pivots, a foundational reference.
Strengths
- AMO framework is the cleanest reference implementation of protocol-owned-liquidity-via-constrained-mint
- Multi-source revenue capture (AMO fees, sfrxETH yield, FXB, Fraxchain sequencer) exceeds peer stablecoins
- Operational transparency exceeds USDT/USDC industry baseline by a wide margin
- Public retreat from algorithmic model is intellectually credible and methodologically rare
- FraxBP successfully displaced 3pool as primary stablecoin Curve venue, capturing routing gravity
- frxETH dual-token design produced measurable yield-premium over alternatives
Weaknesses
- AMO concentration risk in DeFi venues not fully haircut in CR calculations
- frxETH validator-set centralisation is an unresolved tension with broader DeFi decentralisation norms
- FXB adoption disappointing; native yield-curve thesis not realised
- Fraxchain commercial execution behind announced timelines
- FXS revenue model untested under multi-quarter FRAX-supply stagnation
- Recursive dependencies (FRAX↔Curve↔frxETH) create reflexive paths that stress-tests have not exhaustively probed
Key contributions
- Fully collateralised v3 design: collateralisation ratio (CR) raised to 100% via RWA tokenisation (sFRAX, FXB), retiring the algorithmic-fractional model
- AMO (Algorithmic Market Operations) framework: programmable balance-sheet operations that mint/burn FRAX into productive venues without changing CR
- frxETH liquid-staking token and sfrxETH yield-bearing variant, capturing ETH staking yield to subsidise FRAX peg defence
- FraxBP base pool (FRAX/USDC) replacing 3pool as the dominant Curve stablecoin liquidity venue
- Fraxchain announcement: appchain rollup using frxETH for gas, capturing sequencer revenue for FXS holders
- FXS economics revised around veFXS lockups and AMO-revenue distribution, replacing the original mint-and-burn arbitrage flow
- FXB (Frax Bonds): on-chain zero-coupon bonds redeemable for FRAX at maturity, providing duration and yield products natively
TL;DR
Frax's deliberate retreat from algorithmic backing produced the cleanest AMO reference implementation in the stablecoin literature and the most epistemically honest pivot document in DeFi history.
Get DeFi Intel research in your inbox
Weekly long-form coverage of papers, incidents, jurisdictions, chains, tokens and the people building them. Free tier covers headlines; Pro adds the analyst-grade breakdowns.
Sources
External references gathered from the body of this brief. Last reviewed 2026-05-03.