AAVE Considers Token Burn; PUMP Continues Buybacks: Where Do Repurchased Tokens Go?

Written by Xiao Bing

On September 29, Aave founder Stani Kulechov posted that the team is "considering" adding a token burn mechanism to Aavenomics 3.0, prompting AAVE to rise over 11% in a single day.

On the same day, Pump.fun's automated buyback contract continued operating as usual, using 50% of the platform's daily revenue to purchase PUMP tokens on the open market and permanently burn them.

Buybacks and burns have become the most important narrative and fundamental driver of this cycle, but they vary significantly across projects.

Crypto projects say "buyback," and investors often intuitively compare it to stock buybacks: a company uses its profits to repurchase its own shares, reducing the number of outstanding shares and naturally increasing the value per share.

In the crypto world, there are at least four completely different outcomes for what happens to tokens after they are bought back, each with vastly different implications for token holders.

Enter the treasury/ecosystem reserve. Tokens disappear from the open market but remain in the total supply. They are deposited into a DAO treasury or ecosystem reserve address and can be re-spent at any time through governance voting. This is more like moving money from one pocket to another—reducing selling pressure in the short term, without truly contracting long-term supply.

Permanent burn. Tokens are sent to an address with no known private key, permanently removing them from the total supply. This is the only action truly analogous to stock cancellation. However, even then, if the project still has a large amount of unvested tokens pending release, the burn rate may not keep up with the vesting rate.

Transfer to a specific reserve or fund. Tokens are deposited into an address with a designated purpose, such as an insurance fund, liquidity pool, or staking reward pool. Key questions to clarify: Does this address have withdrawal permissions? Is there a lock-up period? Will the tokens eventually return to the market in another form?

Distributed to stakers. Repurchased tokens are distributed to stakers as rewards. This represents real income for stakers but is equivalent to dilution for non-stakers. More critically: do these rewards come from the protocol’s genuine business revenue, or from additional token issuance?

Aave's buyback story can be divided into three phases.

In April 2025, the Aave DAO approved a structured buyback program: executed by the Finance Committee, purchasing approximately $1 million of AAVE from public markets weekly, with an annual budget of about $50 million. By February 2026, the program had cumulatively spent approximately $42 million to acquire over 205,000 AAVE tokens, representing 1.28% of the total supply.

Where did these tokens go? They all went into the Aave Ecosystem Reserve. This reserve is used to fund staking incentives, developer grants, and service provider fees, meaning these tokens can be spent at any time.

In March 2026, a DAO vote reduced the annual buyback budget from $50 million to $30 million. Buybacks were paused from April to June. On June 27, Aavenomics 3.0 launched, replacing the previous committee-managed process with an on-chain automated mechanism that now purchases approximately 292 AAVE daily.

Stani now says “considering a burn,” but as of publication, there is no governance proposal, no burn ratio, no implementation timeline, and no specific parameters. As Cointribune put it: the market heard the word “burn” and buying followed—but there is an entire governance process between “considering” and “deciding.”

Pump.fun has taken a completely different path.

From token launch until April 2026, Pump.fun repurchased 100% of platform revenue in PUMP tokens—among the most aggressive buyback commitments in the crypto industry. On April 28, 2026, the team burned the entire accumulated amount of repurchased PUMP, valued at approximately $370 million, representing 36% of the circulating supply. Simultaneously, they announced a shift to a 50% revenue buyback and burn (automatically executed via an irreversible lock contract for one year), with the remaining 50% allocated to business development.

By the end of July 2026, Pump.fun had cumulatively spent approximately $414.6 million to repurchase and burn 153.73 billion PUMP tokens.

However, PUMP's price remains at rock bottom, having dropped from its issuance price of $0.004 to around $0.0013, a decline of approximately 89% from its all-time high of $0.01214.

Why don't prices rise for projects that are "burning real money"?

8Blocks' analysis cuts to the core: Pump.fun's buyback amount covers only about 2% of daily trading volume. More critically, PUMP lacks essential use cases within the product. Users can create meme coins, trade, and use PumpSwap on Pump.fun without ever needing to hold or consume PUMP. There is no mandatory link between platform growth and token demand.

Additionally, tokens belonging to the team and early investors are still unlocking according to schedule, so the rate of burn cannot keep up with the rate of new releases. The net circulating supply has not genuinely decreased.

Pump.fun taught the entire industry a lesson: buyback and burn mechanisms can achieve textbook-level transparency and irreversibility, but without intrinsic demand for the token, no amount of supply-side manipulation can create lasting price support.

For comparison, Hyperliquid's HYPE offers a nearly entirely different sample.

Hyperliquid's Assistance Fund uses 97% to 99% of the protocol's trading fees to automatically purchase HYPE on the open market. In December 2025, validators officially classified the HYPE held by the Assistance Fund as "permanently burned," removing it from both the circulating and total supply.

As of September 2026, approximately 48.42 million HYPE tokens have been burned, representing 4.84% of the total supply of 1 billion tokens, with an acquisition cost of approximately $1.321 billion and a current value of approximately $4.366 billion. The annualized buyback intensity is about 7% of the market cap, or 4 to 5 times the Ethereum EIP-1559 burn rate.

HYPE is currently trading near its all-time high, with a market capitalization of approximately $14 billion. Hyperliquid's revenue from the beginning of 2026 to date is approximately $429 million.

Why did HYPE rise while PUMP did not, despite both undergoing buybacks and burns? There are at least two structural differences.

HYPE has real functional use cases within the Hyperliquid ecosystem: staking, governance, and gas fee payments. There is a direct transmission pathway between platform growth and token demand. Moreover, Hyperliquid’s trading volume is sufficiently high ($493 billion in settlements in Q1 2026), generating enough fees to make buyback volumes materially impact supply.

However, HYPE is not without risk. The buyback mechanism is a protocol strategy, not a contractual obligation, and can theoretically be modified or paused at the governance level. Additionally, only about 22% of HYPE is currently in circulation, with the remaining tokens unlocking over several years.

According to律动, from January to August 2026, crypto projects collectively spent approximately $638 million on token buybacks, a 17% year-over-year increase.

Nearly 90% of them come from two sources: Hyperliquid and Pump.fun.

The remaining buybacks are distributed across the Sky ecosystem (approximately $26 million in SKY buybacks), Uniswap (UNI burns beginning in December 2025 using protocol fees), and other projects.

But of the $638 million spent, how much actually reduced token supply? How much was merely moved between the treasury and the market? How much was completely offset by unlocked releases? Each of these questions should be asked under every "buyback" headline.

When a project announces a "buyback," token holders should verify at least five things:

Where does the funding for buybacks come from? If it comes from genuine protocol revenue (fees, interest, licensing fees), that’s healthy. If it comes from token reserves in the treasury or additional minting, it’s just moving money from one pocket to another.

Where did the purchased tokens go? Permanent destruction (sent to a verifiable zero address) is the only true way to reduce supply. Sending tokens to the treasury, ecosystem reserves, or discretionary funds still leaves open the possibility of them being re-released.

Can the buyback rate outpace the vesting rate? If a project still has a large amount of team, investor, or ecosystem incentive tokens pending release—destroying 1,000 tokens per week but unlocking 100,000 tokens per month—the net effect is inflation, not deflation.

Is the mechanism irreversible or adjustable? Pump.fun’s locked contract is superior to Aave’s previous committee model. However, even Hyperliquid’s Assistance Fund could theoretically have its parameters modified through governance.

Is there inherent demand for the token itself? This is the most easily overlooked but most fundamental question. Without use cases, buybacks merely hold off selling pressure—they don’t create buying pressure. When the buyback budget runs out or the unlocking peak arrives, the price will still fall.

A single tweet from Stani triggered an 11% price increase, revealing how strongly the market craves the "burn" narrative. But the gap between desire and reality is often the difference between profit and loss.