A modular AMM whose Monad liquidity today sits mostly in a handful of small, newer pools rather than the deep blue-chip pairs — proposed for Monad two years before Monad's mainnet even existed.
Balancer V3 is a modular AMM built around a single central Vault contract that holds all pool assets, with individual pool math plugged in separately so multiple pool types can share the same infrastructure. Weighted pools allow custom token ratios beyond the standard 50/50 — useful for index-fund-style baskets or projects that want to retain majority exposure to their own token while bootstrapping liquidity (an 80/20 split, for instance). Boosted pools hold yield-bearing wrapped tokens (for example, assets deposited into Aave) so idle liquidity earns lending yield in the background while still being available for swaps; V3 uniquely allows up to 100% of a position to be boosted, versus a lower cap in V2. Stable pools use the same StableSwap invariant Curve introduced — Balancer's own docs describe the design as directly based on Curve's approach — extended with a "StableSurge" hook that dynamically raises fees as a pool drifts from equilibrium, something Curve's base invariant doesn't do natively.
V3's central architectural shift is moving 18-decimal scaling and rounding-direction rules out of individual pool contracts and into the Vault itself, enforced consistently for every pool type rather than reimplemented per-pool. Vault-level "liquidity buffers" keep a small reserve of both the wrapped and underlying asset in boosted pools, rebalanced close to 50/50, so most trades skip calling the underlying lending protocol's wrap/unwrap functions entirely — a gas saving with no official published percentage figure, so treat any specific number you see elsewhere skeptically. V3 also adopts EIP-1153 transient storage for flash accounting, the same "Till" pattern concept Uniswap V4 uses, netting balance deltas across a callback chain and settling once.
Against Uniswap V4, the two are the closest architectural peers on Monad — both replaced per-pool contracts with a singleton custody layer plus hook-based extensibility — but Balancer's Vault was designed so pool contracts hold no custody logic at all, a stricter separation of concerns than V4, where a hook can still directly move funds. Against Uniswap V3, Balancer's edge isn't concentrated liquidity but multi-asset flexibility: up to eight tokens at custom weights in one pool, versus Uniswap's strictly two-token pools. Against Curve, Balancer's Stable pools use literally the same underlying invariant — a case of shared lineage rather than competition — with Balancer's dynamic-fee hook as the one meaningful addition Curve's base design lacks. Against Kuru's order book, Balancer remains a pooled-liquidity AMM with the standard tradeoff: permissionless liquidity without needing an active market maker to show up, at the cost of the tighter execution a real order book can offer when one is present.
On November 3, 2025, Balancer V2's Composable Stable Pools were exploited across nine blockchain networks via a rounding-direction bug in the invariant computation used by `_upscaleArray` — roughly 65 crafted micro-swaps drove pool balances down to single-digit-wei levels, at which point integer-division rounding errors let the attacker mint undervalued pool tokens and drain real assets. Reported losses range from $110M to $128.6M depending on the source and whether every affected fork is counted. Trail of Bits, which had flagged a related rounding issue in a 2021 audit without confirming exploitability at the time, states plainly: "We independently confirmed that Balancer v3 was not affected by this vulnerability," attributing V3's immunity directly to its centralized Vault-level rounding and scaling architecture — the exact change described above. The exploit was still consequential for the wider protocol: it was cited as the direct trigger for Balancer Labs, the corporate development entity, announcing its own dissolution in March 2026, citing ongoing legal exposure; the protocol itself continues under DAO governance, with a subsequent tokenomics overhaul ending BAL emissions, sunsetting veBAL, and raising the LP fee share from 50% to 75%.
A Balancer governance proposal to deploy on Monad was posted in March 2024 — well before Monad's mainnet even launched in November 2025 — pitching Monad's throughput as a fit for Balancer's pool types, with actual deployment timing left to later votes. Real on-chain Monad activity only became meaningful starting around February 2026. Pools found on Monad today are relatively thin and niche — small newer-token pairs rather than the deep WMON/USDC-style liquidity that mostly lives on other Monad DEXs — and Balancer has also been cited as a liquidity venue chosen by at least one Monad-native token community.
A hook is arbitrary code — a pool's actual behavior depends on whichever hook it's attached to, which may not be as scrutinized as Balancer's own core contracts. Boosted pools also inherit the risk of whatever lending protocol their yield-bearing wrapper depends on: a problem there can impair Balancer LPs even though the bug isn't in Balancer's code. V3's architecture is independently confirmed unaffected by the November 2025 V2 exploit, but that exploit is a live reminder that Balancer's Stable-pool math has been a recurring target across the protocol's history. And with Monad TVL concentrated in a handful of small, new pools, prices are easier to move and thinner liquidity means more price impact on a given trade size than the headline chain TVL might suggest.
Note: BAL's tokenomics are actively changing, not hypothetical: emissions have ended, veBAL is being sunset, and LP fee share rose from 50% to 75% as part of the DAO takeover following Balancer Labs' March 2026 wind-down.
No. The exploit hit V2's Composable Stable Pools specifically, via a rounding bug in code that V3 doesn't share — Trail of Bits independently confirmed V3's architecture, which centralizes rounding and scaling rules in the Vault rather than per-pool, was not affected.
They let LP capital earn lending yield (via ERC-4626 wrappers like Aave's aTokens) at the same time as swap fees, instead of sitting idle waiting to be traded against. V3 allows up to 100% of a position to be boosted, a cap V2 didn't reach.
Yes — Balancer Labs was the corporate development entity, not the protocol itself. It dissolved in March 2026 citing legal exposure from the November 2025 exploit, but the protocol continues under DAO governance ("Balancer OpCo"), with a subsequent tokenomics overhaul redirecting more fee revenue to LPs.
Balancer deploys chain-by-chain via separate governance votes rather than launching simultaneously everywhere — its Monad proposal was posted in March 2024, over a year and a half before Monad's own mainnet existed, and real activity only started around February 2026. New deployments begin with no incentivized liquidity by default.
Sources: BIP-571: Deploy Balancer on Monad — Balancer Forum, Boosted Pool — Balancer Docs, Modern DEXes: how Balancer V3 is made — MixBytes, Breaking down the Balancer hack — Certora, Balancer hack analysis and guidance — Trail of Bits, Balancer Labs will shut down — CoinDesk
Last reviewed 2026-07-08. More Monad research.
Monad DeFi board · Markets · Swap on NullTerminal