On-Chain Capital Budgeting

Author: Prathik Desai, Translated by: Block unicorn

In 1965, after Warren Buffett took over Berkshire Hathaway, he began writing letters to the shareholders of a struggling textile mill. In these letters, he detailed how the company’s cash was being reinvested and the reasons behind it.

Capital allocation truly became a matter of transparency and openness. Although shareholders are the owners of the company, they usually trust management to operate as trustees. Buffett’s public letters made shareholders feel that their opinions could be heard during annual discussions about the company’s future.

Most publicly traded companies conduct capital budgeting behind closed doors. Corporate law requires the disclosure of audited figures, not the reasons behind decisions. But Berkshire Hathaway has always transparently disclosed the rationale behind its decisions. For sixty years, the company’s letters read like a continuous dialogue about the next steps in capital investment.

A form of this practice is now appearing on the blockchain.

DeFi protocols also have their own “board of directors”: informal channels, founder conference calls, and representative meetings. However, voting results, proposal texts, objections, amendments, and final proposals are all recorded.

On April 12, Aave token holders voted on a proposal to allocate $25 million. The proposal was overwhelmingly approved, granting Aave token holders the right to vote on governance, protocol upgrades, and risk parameters.

Besides Aave, this move also helps shift the debate in cryptocurrency from whether tokens have claims on protocol cash flows to allowing token holders to decide which projects to fund next and how much to allocate.

In today’s analysis, I will tell you how.

What was the voting outcome?

Aave’s vote is crucial because it plays a significant role in the lending and borrowing market. Currently, Aave has over $26 billion in total value locked (TVL), generating about $8B in fees last year. Most of this flows to suppliers earning deposit interest, leaving the protocol itself with roughly $907M in annual revenue. This revenue ultimately goes into the DAO treasury.

As of December 31, 2025, Aave’s active loans on the books exceeded $21 billion.

This is larger than Robinhood’s margin account size of $16.8 billion at the end of 2025, nearly twice LendingClub’s total assets ($11.8 billion). LendingClub is a listed online lending platform on the New York Stock Exchange, with a business model most similar to Aave.

In 2025, LendingClub, with a business model similar to Aave’s, achieved a net profit of $136 million on revenue of $1 billion. This 20-year-old company’s total assets reached $11.8 billion at year-end 2025. It is competing with some well-known traditional fintech firms.

In 2025, LendingClub’s net income was $136 million on a revenue base of $1 billion. Aave’s $125 million revenue comes from the DAO, and its loan scale is nearly twice that of LendingClub, meaning that for every dollar lent, LendingClub earns about twice as much as Aave.

This is because Aave is just a protocol. Since it does not hold deposits, it cannot charge higher interest rates without disrupting its own market. Raising reserve requirements would push suppliers toward more profitable competing protocols. LendingClub can charge higher interest because it owns loans, bears credit risk, and has built a complete underwriting system and default reserves.

Aave’s proposal is to allocate all income from all Aave-branded products (including App, Card, and Pro institutional version) into the Aave DAO treasury. In return, the DAO will pay Aave Labs $25 million plus 75,000 AAVE tokens (distributed over four years) to support ongoing development.

Aave Labs is the development entity behind the Aave protocol, which previously retained part of the product revenue for its own operations. The integration of Cowswap on Aave.com charges a fee of 15-25 basis points for token swaps, and about $10 million annually (roughly 10% of the DAO’s potential revenue) is funneled into Labs’ private address. Marc Zeller of ACI calls this the “covert privatization of DAO revenue.” Under the new structure, revenue will be paid by token holders, with payment terms set by holders themselves, and payments can be paused based on milestone nodes.

Token holders did not approve the “temporary approval” plan proposed in early March. They split the proposal, which included V4, branding, funding, and growth grants, into multiple separate votes, extending the vesting period from two to four years. They also linked fund disbursement to milestones rather than paying it all at once.

Not everyone supported the original version. Some agreed to negotiated versions, while others chose to exit after long-term cooperation.

Over the past two months, three major contributing institutions have terminated their cooperation with the protocol. BGD Labs withdrew in February. After contributing 61% of governance actions and 48% of protocol revenue over three years, Aave Chan Initiative (ACI) announced its exit in March. Chaos Labs, which helped increase Aave’s TVL from $5.2 billion to over $26 billion without major bad debts, stepped down from risk management on April 6.

A bigger shift is that large protocols with product-market fit and surplus now enable token holders to do what was traditionally only possible for the board of directors.

Uniswap is heading in the same direction from a different angle. Its unification proposal relies on protocol fees to burn UNI tokens, disable Labs-level interfaces and wallet fees, and allocate funds from the treasury to Labs.

In 2025, Uniswap’s fee revenue reached $1 billion, nearly matching Aave’s. But historically, almost all of these fees flowed to liquidity providers. Its current annual protocol revenue is only $42 million. UNI holders get nothing.

In March alone, Uniswap’s decentralized exchange (DEX) traded $42 billion. Meanwhile, Robinhood’s app recorded $34 billion in nominal cryptocurrency trading volume in Q4 2025. Uniswap’s monthly trading volume slightly exceeded that figure.

The two most prominent DeFi protocols are moving toward the same structural model, though mechanisms differ. Both aim to route revenue to the token layer and then publicly decide how to redeploy those earnings.

The new financial strategy changes the decision-making power over how funds in the treasury are redistributed. Should these funds be used to buy back and burn tokens? Or for operational budgets and growth grants? Or perhaps to develop new product lines or for long-term development? All these are capital allocation questions—and Buffett has been answering them in his letters for decades.

For years, the DeFi space has debated whether tokens should have claims on protocol cash flows. We’ve seen Hyperliquid and pump.fun address this with large-scale buyback programs.

What Public Capital Budgeting Requires

Buffett’s letters are voluntary. What makes his actions more than just goodwill is the audited financial statements behind the letters. He bears fiduciary responsibility, which can be enforced by law. Berkshire shareholders have rights that do not depend on Buffett’s goodwill; they can sue the company, vote to dismiss the board, and review financial records.

Regulatory agencies require companies to disclose basic information, no matter how appealing Buffett’s letters sound.

In DeFi, transparency is critical. On-chain architecture mandates that all transactions, fund flows, and voting are open and transparent. But this alone does not make crypto protocols fully transparent to the public. Token holders need a shared understanding of the business to make capital budgeting decisions.

Because of standardized accounting principles, financial statements and documents of listed companies are public and transparent. Nvidia’s “total revenue” is roughly the same as Apple’s “total revenue.” If you can understand Nvidia’s balance sheet, you’re likely able to interpret Apple’s financial statements as well.

But this standardization of financial terminology is largely absent in crypto protocols. Terms like revenue and expenses are often used interchangeably, and their meanings can be vastly different.

Token holders can still see Aave’s assets and liabilities down to the penny. But unless there are standardized definitions, they cannot distinguish which income belongs to “Aave brand product revenue.” They still need to know what happens if Aave Labs and the DAO disagree on that figure.

This is not hypothetical. One of the biggest disputes in the Aave proposal revolves around what “100% revenue” actually means. Is it gross revenue or net income after operational costs? Who verifies this? How is it enforced?

Even Marc Zeller, the founder of the now-exited ACI, admits these concerns in his X posts after the Aave DAO wins.

As a token holder, you cannot allocate what you cannot define.

Checks and Balances

Although the voting results mark a transfer of power to token holders, those who have left Aave believe the transfer is incomplete.

The two teams most closely related to Aave’s governance— the ACI team responsible for coordinating 61% of governance actions, and BGD Labs, which develops the core codebase—have both complained about similar issues. Meanwhile, the Aave Labs team applying for funds is also the one voting to approve the grants. Clearly, there is a conflict of interest.

The budget recipient has no voting rights over the funds they receive.

This is a structural problem common to all on-chain capital budgeting, especially when founders or core teams hold large token positions. While, in theory, surplus funds should be decided by token holders, in practice, the largest token holders often draft the proposals.

You might think that the higher your shareholding, the greater your influence over company operations. But if DeFi protocols want to emulate traditional corporate capital budgeting, they must ensure similar checks and balances. For example, legal frameworks can protect minority shareholders from majority abuse.

DeFi also needs similar checks and balances. Revenue data must be independently verified, including off-chain data. For instance, when Venice AI announced it would burn VVV tokens using Stripe subscription revenue, DefiLlama’s 0xngmi demanded read-only access to Stripe to audit the off-chain data supporting on-chain claims.

On-chain milestones must be established to trigger phased fund releases, and releases should be revoked if milestones are not met. Protocols must specify how voting rights are disclosed for entities receiving DAO funds. These are all feasible; it’s just that those who must agree to constraints are often the ones writing the proposals, making them less popular.

One team explicitly requested safeguards before supporting the Aave proposal, such as tracking on-chain milestones and self-voting restrictions. But these suggestions were ignored, and the proposal was ultimately approved.

What does Aave’s “victory” mean?

Voting itself is not an achievement. It’s easy to pass a $25 million funding proposal on-chain when the largest token holders support it. But there are lessons worth learning here.

Token holders rejected the first version of the proposal. They split a large proposal into several smaller decisions. They doubled the lock-up period. They forced founder-led entities to negotiate terms openly, amend terms publicly, and accept restrictions they did not initially propose. This is a step toward open capital budgeting.

If Aave’s proposal becomes a template, then the mechanism allowing token holders to modify proposals before approval must be adopted by all protocols. Just as in democratic systems, if a policy cannot withstand all objections and disagreements, it is not robust. In DAOs, opposition is not dysfunction; it makes the DAO more transparent and resilient over time.

Uniswap’s UNIfication proposal faces similar challenges. It must address questions like “Who decides the treasury’s fund allocation?” and “What if issues arise—how to remedy?” These are questions many blue-chip DeFi protocols will face. They must understand that if policy decisions lack rationality and basis, the community will not remember any buybacks or token burns.

Buffett and Berkshire did not earn respect immediately after their first letter. Shareholders took years of reading his letters to gradually trust that this entrepreneur could wisely manage their funds for profit.