Crypto: Balancer prepares to shut down after v3 fails

Balancer could gradually disappear after more than five years among DeFi’s biggest names. The crypto protocol is preparing an orderly wind-down following the failure of its restructuring launched in the spring. The problem is not technical: Balancer v3 works. The problem is economic. The new products have not generated enough revenue, the older v2 remains the main source of income, and the November 2025 exploit continues to weigh on the brand. The DAO still holds at least $9 million in treasury funds. It now prefers to return that money to BAL holders rather than spend it extending an unprofitable operation.

Crypto: Balancer prepares for a gradual exit

The wind-down plan was published on September 14 by Marcus Hardt, CEO of Balancer Labs, on the protocol’s governance forum. BAL holders must still approve it in a Snapshot vote scheduled for September 25 to 29.

Balancer joins a series of crypto projects facing a question that has become much more pressing in 2026: can a protocol actually generate enough money to fund its development? Bref Crypto recently documented the case of Router Protocol, which will shut down on September 30 after its business model failed.

For Balancer, the proposed answer is now closure.

The plan officially presented to the DAO calls for commercial development to end, a period allowing liquidity providers to exit, and the infrastructure to be gradually reduced to the strict minimum.

On October 30, 2026, pools that can be paused would switch to withdrawal-only mode. Those whose smart contracts do not support such a suspension would continue operating, but protocol fees would be reduced to zero where technically possible.

From November 1, Balancer would essentially maintain only the tools needed for withdrawals, documentation and the unlocking of veBAL positions.

This is therefore not an abrupt shutdown.

The non-custodial smart contracts will remain accessible, and users will be able to recover their assets directly from the contracts when necessary.

v3 works, but not enough people are using it

This is probably the most important sentence in the entire story.

« The product worked. It did not sell enough. »

Marcus Hardt does not describe Balancer v3 as a technological failure.

The new architecture works, and several expected products were launched: v3, Boosted Pools, AutoRange Pools, new integrations and various partnerships.

The problem lies in revenue.

Balancer had nevertheless restructured its operations in the spring around an extremely clear objective: turning the protocol into an economically sustainable crypto business.

BAL emissions were halted. Costs were reduced. The token model was simplified. A larger share of protocol revenue was directed to the DAO.

This reorganization followed the closure of Balancer Labs announced in March. At the time, the team still believed it could maintain the protocol with a much leaner structure.

Bref Crypto noted in its analysis of the major crypto narratives of 2026 that DeFi now had to prove its ability to generate real revenue, well beyond yield farming and token emissions.

Balancer has now provided a rather cold demonstration of that reality.

Despite the savings and new products, v3 did not replace v2’s revenue.

And that is almost the paradox at the heart of the entire restructuring: the older product, the one Balancer was gradually trying to move beyond, remains the protocol’s main source of income.

The technological migration succeeded.

The economic migration did not.

Revenue falls from $1.13M to almost nothing

The figures show the scale of the deterioration.

According to DefiLlama data cited by Cointelegraph, Balancer generated approximately $1.13 million in revenue in October 2025.

Then November arrived.

The protocol suffered a major exploit affecting certain Balancer v2 Composable Stable Pools. Estimates of the amount have changed depending on the sources and periods; Cointelegraph now puts the incident at around $128 million.

Monthly revenue immediately fell to $371,000 in November.

Then it continued to decline.

In August 2026, DefiLlama recorded just $56,781 in monthly revenue for Balancer.

The DAO’s own wind-down document uses an even more conservative measure: approximately $30,000 in protocol revenue in August, compared with $97,000 in June.

The difference between the two figures is notably due to the accounting methods and scopes used. The overall magnitude remains far below expenses.

Balancer estimates its current costs at approximately $150,000 per month. Even with the roughly $25,000 per month currently generated by treasury management, the calculation remains unfavorable.

$30,000 in revenue. $150,000 in expenses. It quickly becomes clear why the DAO no longer wants to wait. This is not a protocol that has suddenly emptied its coffers. It is a protocol watching its treasury decline every month without a credible scenario for reversing the trend.

The 2025 hack continues to be costly

Balancer emphasizes one point: the closure is not based solely on the November 2025 exploit. This matters. Balancer v3 uses an architecture different from that of the affected v2 pools. Closing the protocol by simply saying “a hack killed it” would therefore be reductive.

The hack’s most lasting effect occurred elsewhere: in trust. “The event has followed the Balancer name in every conversation since,” Marcus Hardt acknowledges. He says he underestimated how long the incident would continue to limit v3 adoption.

A partner considering integrating a DeFi protocol does not look only at whether the latest version of the code contains the same vulnerability. It looks at the name. The history. The risks. The liquidity. The users already there. And that cycle can become particularly difficult to reverse.

Less trust produces less liquidity. Less liquidity makes a DEX less attractive. And fewer transactions produce less fee revenue. Lower revenue then reduces the ability to fund development, security and user acquisition.

The entire DeFi sector faces this problem. Bref Crypto had already counted nearly 70 attacks in the second quarter of 2026, causing approximately $746M in losses.

Security is therefore not merely a technical expense. It can now directly affect a protocol’s commercial capacity for several months after a vulnerability has been fixed. At Balancer, the exploit did not break v3. It appears to have damaged something more difficult to repair: trust in the brand.

$9M will be redistributed to BAL holders

The most unusual part of the plan concerns the treasury. Balancer still holds at least $9 million in managed assets, according to the current estimate from its manager kpk. Other wallets, positions and assets belonging to the DAO still need to be inventoried before the distribution. Hardt therefore poses the question quite simply: should this money continue to be spent in the hope of a recovery, or should what remains be returned to BAL holders?

His proposal chooses the second option. The wind-down plan would cost a maximum of $400,000: $150,000 for the first phase through May 2027, $30,000 for the remainder, and a $220,000 reserve usable only if necessary. Any unused budget would return to the treasury being distributed.

The first distribution is scheduled for the end of May 2027. Holders would then have to burn their BAL in order to receive their proportional share of the assets actually held by the treasury.

This detail matters. Users would not receive an additional payment in BAL. The token would instead serve as a ticket enabling them to claim a fraction of the DAO’s remaining assets.

A second distribution is to follow after the end of the first claim period. It will include unused budgets, revenue received in the meantime and assets not claimed during the first round. Finally, a last “sweep” is planned six months later. The process would therefore not actually end until around July 2028.

Closing a DAO is far more complicated than closing a company

The Balancer plan also shows why shutting down a crypto project can become much more complicated than liquidating a traditional company. A conventional company can lay off its employees, sell its assets, repay its creditors and close its accounts. Balancer must deal with smart contracts deployed across several networks.

The ecosystem involved is not limited to a few liquidity pools. It also includes veBAL positions locked up, derivatives such as auraBAL, sdBAL and tetuBAL, several multisigs, sensitive administrative permissions, and DAO assets spread across different wallets. This is compounded by liquidity providers already affected by the 2025 exploit, making any restructuring or closure operation even more complex.

There is also a governance system in which token holders must still approve certain steps. Funds recovered after the attacks, for example, will not be included in the $9 million to be distributed to BAL holders. They belong to the affected liquidity providers and follow a separate process.

The code, licenses and certain deployments may also be transferred or taken over, but each transfer will require its own Snapshot vote. Then there are the permissions. The DAO plans to gradually remove or transfer its privileged access so that the protocol can ultimately reach a state in which Balancer itself no longer needs to exist for users to withdraw their funds.

It is almost a liquidation programmed into the code. Decentralization makes closure more complicated. It also allows certain contracts to continue operating after the organization that created them has disappeared.

Balancer reached $3.3 billion in TVL at its peak

The decline is all the more striking because Balancer was not a small experimental project. During the previous cycle, it was one of Ethereum’s major DeFi infrastructures. Its TVL reached approximately $3.3 billion in November 2021. In October 2025, before the exploit, it had already fallen to around $800 million. A few months later, after the hack, it represented only approximately $158 million, according to figures published during the spring restructuring.

A decline in TVL is not necessarily a death sentence. The value of deposited tokens changes with the market, and some protocols are specifically seeking to earn more with less capital locked up. Balancer also had genuine innovation.

Its automated market makers made it possible to build pools with more than two assets and customizable weightings, whereas Uniswap’s classic model historically relied on 50/50 pairs.

Boosted Pools also contributed to DeFi’s technical evolution. None of this was enough. The story recalls that of Router Protocol, but on a different scale: a product can work, have an experienced team, survive for several years and even invent mechanisms adopted elsewhere without finding a sustainable economy. It is a much less spectacular lesson than a bull market. It is probably more important.

DeFi enters the age of revenue

For a long time, the crypto industry measured success using three figures: TVL, token price and user numbers. The Balancer case requires a fourth column: how much does the protocol actually earn? Not how many BAL tokens it distributes. Not how many dollars are temporarily deposited in pools attracted by incentives.

How much revenue are its users actually willing to pay? This is where the proposed closure becomes almost a test for DeFi as a whole. Balancer is not immediately short of cash. With more than $9 million in treasury funds, it could probably continue for a long time in a reduced configuration.

Hardt is explicitly rejecting that approach. Spending several million more to buy time makes sense only if there is a realistic path toward revenue exceeding costs.

After the April restructuring, Balancer tried. Costs fell. Emissions disappeared. v3 was delivered. AutoRange was launched. New integrations were pursued. The expected revenue never arrived. “Continuing on the current trajectory spends the treasury to reach the same place later,” the proposal summarizes.

This is probably the sentence that will remain from this closure if the DAO votes yes. Crypto has long viewed treasuries as fuel intended to extend development indefinitely until the next bull market.

Balancer is proposing something else: stop before the treasury is depleted and return the remaining capital to token holders. The vote from September 25 to 29 will determine whether BAL holders accept this conclusion. If they vote yes, withdrawals-only mode will begin on October 30, and the final distribution is not expected until 2028.

If they vote no, the current model will remain in place. But the economic diagnosis will not disappear with the vote. Balancer has already succeeded at the most technically difficult part: building a DEX used across several cycles and delivering a new architecture after a major hack.

What it failed to rebuild was revenue. And in 2026, in a crypto industry that has become far less generous toward loss-making projects, that may be enough to close even a historic protocol.