Balancer, one of decentralized finance's earliest automated market makers, is heading toward an orderly shutdown. On September 14, co-founder Marcus Hardt posted a governance proposal to wind down the protocol and distribute its remaining $9 million treasury to BAL token holders. A Snapshot vote ...
"The product worked. It did not sell enough." — Marcus Hardt, Co-founder & former CEO, Balancer Labs
Balancer, one of decentralized finance's earliest automated market makers, is heading toward an orderly shutdown. On September 14, co-founder Marcus Hardt posted a governance proposal to wind down the protocol and distribute its remaining $9 million treasury to BAL token holders. A Snapshot vote is scheduled for September 25-29.
The proposal follows a $128 million exploit in November 2025 that drained Balancer v2 composable stable pools across Ethereum and multiple layer-2 networks. Monthly protocol revenue collapsed from $1.13 million in October 2025 to $56,781 by August 2026 — a 95% decline over ten months. Balancer Labs, the corporate entity, shut down in March 2026. The remaining DAO team was cut from approximately 25 to 12.5 full-time equivalents.
At its peak in 2021, Balancer held $3 billion in total value locked and its BAL token traded at $74.45, reaching a fully diluted valuation near $70 billion. Today, TVL sits at $58 million and BAL trades at $0.12 — a 99.8% decline from the token's all-time high. The current fully diluted valuation is approximately $8.1 million.
On November 3, 2025, an attacker exploited a rounding bug in the "upscale" function of Balancer's v2 vault contracts. The attacker deployed malicious smart contracts and counterfeit tokens to manipulate invariant inputs — the mathematical rules governing token swaps — enabling swaps at favorable prices that drained pool liquidity.
Total losses reached $128 million across Ethereum (approximately $100 million), Arbitrum, Polygon, and Base. Primary affected assets included WETH, osETH, and wstETH held in v2 composable stable pools.
This occurred despite full security audits from OpenZeppelin, Trail of Bits, Certora, and ABKD. The exploit coincided with a broader record year for crypto theft, with industry-wide losses exceeding $2.2 billion in 2025 according to on-chain security trackers.
Balancer was no stranger to security incidents. A September 2023 "Boosted Pools" exploit across Ethereum and Optimism cost $1.2 million. An August 2026 exploit of the v1 codebase drained an additional $200,000. But the November 2025 breach was categorically different in scale — roughly 100x the combined total of all prior incidents.
The financial deterioration was rapid and unrecoverable:
| Month | Monthly Revenue | |-------|----------------| | October 2025 | $1,130,000 | | November 2025 | $371,000 | | December 2025 | Continued decline | | August 2026 | $56,781 |
Revenue fell 67% in the first month after the exploit and never recovered. By August 2026, monthly revenue was 5% of its pre-exploit level.
The revenue problem was compounded by costs. In April 2026, the DAO approved cost-cutting measures: team reduction from approximately 25 to 12.5 full-time equivalents, operating budget cut by approximately one-third, redirection of protocol fees to the treasury, and a halt to token emissions. These measures slowed the treasury burn but could not reverse the revenue trajectory.
As Hardt wrote in the governance forum: "Continuing on the current path spends the treasury to arrive at the same place later. That treasury belongs to BAL holders."
Balancer v3 was designed as the protocol's second act. It introduced a different architecture from v2, including Boosted Pools and AutoRange Pools intended to generate new revenue streams and restore user confidence.
It did not work.
Hardt acknowledged the failure directly: "The November 2025 exploit hit legacy v2 pools. v3 is a different architecture, but the event followed the name into every conversation."
According to DeFiLlama, Balancer v3 held $25.8 million in TVL as of mid-September 2026, down 19.1% over the prior 30 days, spread across nine chains including Ethereum, Arbitrum, OP Mainnet, and Gnosis. The combined v2 and v3 TVL of $58 million represents a 98% decline from the protocol's $3 billion peak in 2021.
The v3 launch demonstrated a structural problem in DeFi: brand damage from a security exploit is not contained to the vulnerable version. Users did not distinguish between v2 and v3. The Balancer name itself became a risk signal, and capital flowed elsewhere.
The proposed shutdown follows a phased timeline:
Phase 1 — Governance (September 2026)
Phase 2 — Exit Window (October 2026)
Phase 3 — Infrastructure Wind-Down (November 2026 onward)
Phase 4 — Treasury Distribution (2027-2028)
The protocol's smart contracts remain open-source. Anyone can fork the code. But the DAO, treasury, team, and brand cease to exist as coordinated entities.
Balancer's v2 vault architecture was among the most forked codebases in DeFi. According to DeFiLlama, 27 protocols were built on Balancer v2 code, with a combined TVL of $78 million at the time of the exploit.
The November 2025 attack did not just hit Balancer. It propagated across the fork ecosystem:
This cascade illustrates a systemic risk in DeFi's open-source model. Code reuse accelerates development but also replicates vulnerabilities. When a foundational codebase is compromised, the blast radius extends far beyond the original protocol.
Not all $128 million was permanently lost. Several partial recovery efforts yielded results:
Total confirmed recoveries: approximately $36.4 million, or roughly 28% of the $128 million stolen. The remaining $91.6 million appears unrecovered. The suspected attacker's address continued spawning new contracts and minting custom tokens post-exploit, according to on-chain analysis, suggesting an ongoing laundering operation rather than a one-time theft.
Balancer's shutdown carries broader implications for DeFi's economic sustainability.
Revenue concentration risk. Balancer's revenue was overwhelmingly concentrated in v2 pools. When those pools were compromised, the entire revenue base collapsed. There was no diversified income stream to absorb the shock. This mirrors a pattern seen across DeFi: most protocols derive the majority of revenue from a small number of pools or products.
The audit gap. Four reputable audit firms reviewed Balancer's code. The exploit happened anyway. This is not an indictment of any individual auditor, but it underscores that security audits are probabilistic, not deterministic. They reduce risk; they do not eliminate it.
Brand mortality in DeFi. Traditional financial institutions survive security incidents routinely — banks recover from data breaches, exchanges reopen after outages. DeFi protocols, lacking the institutional trust infrastructure and regulatory backstops of traditional finance, face a steeper recovery curve. Balancer's v3 was technically sound but commercially dead because users associated the name with the v2 exploit.
The 2026 attrition wave. Balancer joins a growing list. According to RootData, 99 crypto projects shut down in H1 2026. More than 40 DeFi protocols specifically closed their doors, with DeFi accounting for over half of all shutdowns. Sector-wide TVL fell 39% from approximately $115 billion in January to $70 billion by late June 2026, per CryptoTimes reporting. Balancer's closure is the highest-profile DeFi wind-down in this cycle, but it is part of a broader pattern of unsustainable protocol economics meeting market reality.
Balancer's proposed wind-down is the orderly death of a protocol that, at its peak, represented one of DeFi's core building blocks. Founded in March 2020 by Fernando Martinelli and Mike McDonald, it pioneered weighted liquidity pools and programmable AMM design that influenced dozens of subsequent protocols.
The $128 million exploit did not destroy Balancer's technology — v3 functioned as designed. It destroyed Balancer's market position. Users left and did not return. Revenue evaporated. The corporate entity closed. And now, if BAL holders vote in favor between September 25-29, the DAO will begin a methodical two-year process of returning $9 million to token holders and shutting down infrastructure.
The data suggests a structural vulnerability in DeFi protocol economics: protocols that depend on user trust and liquidity depth are fragile to security incidents in ways that their technology alone cannot remedy. Multiple audits, a new codebase, and a reduced cost structure were insufficient to reverse the capital outflow.
What remains is open-source code, 27 forks of varying health, and $91.6 million in unrecovered funds. The product worked. The market moved on.