Since January 2025, crypto protocols have spent or destroyed approximately $18.8 billion through token buyback and burn programs. The figure is large. The share reaching token holders is not. A persistent structural gap separates protocol revenue generation from actual token holder value accrual,...
"Out of 1,244 protocols, only about 20 passed more than $10 million in value to holders." — 1kx Research
Since January 2025, crypto protocols have spent or destroyed approximately $18.8 billion through token buyback and burn programs. The figure is large. The share reaching token holders is not. A persistent structural gap separates protocol revenue generation from actual token holder value accrual, and the buyback wave of 2025–2026 has made this gap measurable for the first time.
Six major protocols — Aave, Aerodrome, Hyperliquid, Pump, Sky Ecosystem, and Uniswap — collectively generated $7.42 billion in revenue in H1 2026, according to KuCoin research. Token prices for most of these protocols declined over the same period. The disconnect between protocol-level cash flow and token-level returns is the central tension in DeFi governance today.
This report examines nine protocols operating distinct value-return mechanisms — buybacks, burns, fee switches, and staking distributions — and assesses which structures actually transfer value to token holders versus those that accumulate it in protocol treasuries or corporate entities.
On-chain treasury management tooling is seeing active development. The koeppelmann/GnosisDAO_treasury repository pushes daily automated treasury data snapshots — the latest commit dated August 23, 2026 reads "Update treasury data 2026-08-23." This pattern of daily automated commits indicates ongoing programmatic tracking of DAO treasury positions, a signal that institutional-grade treasury management is becoming standard practice.
Similarly, eqty-dao/treasury refreshes snapshot data multiple times per hour (three commits on August 23 alone), suggesting real-time treasury monitoring infrastructure. The M0 Platform's ttg (Two Token Governance) repository, which implements a dual-token governance mechanism for managing communal property, holds 11 stars and 2 forks — modest adoption, but the architecture of separating governance power from economic rights is conceptually relevant to the buyback discussion.
On the AI-crypto intersection, sentient-agi/CryptoAnalystBench, a benchmark for evaluating crypto AI agents, was last updated August 15, 2026. The esskslifetech/BagsAI-Agent-Forge repo implements token buybacks and holder rewards within an AI agent framework on Solana, pointing to buyback mechanics embedding themselves in new protocol categories.
The GitHub signal is clear: treasury management and value distribution code is being actively maintained, but the repos are infrastructure-level — tooling for tracking, not innovations in distribution.
According to Tokenomist research, the breakdown across 27 tokens since January 2025:
Burns dominate. More than four-fifths of the dollar volume comes from token burns — primarily exchange tokens like BNB (quarterly Auto-Burn), OKB, and BGB. The buyback programs that generate headlines — Hyperliquid, Jupiter, Sky — account for roughly $2.8B combined.
The critical finding from Tokenomist: only three of ten major tokens with active buyback programs — AAVE, HYPE, and SKY — outperformed Bitcoin during their buyback windows. Buyback announcements generate short-term price responses, but sustained outperformance requires revenue durability, not just treasury spending.
As CryptoBriefing and DWF Labs note, a buyback funded by recurring protocol revenue can run indefinitely; one funded by a finite treasury balance depletes. The funding source is the single most important variable in assessing buyback sustainability.
Hyperliquid operates the most aggressive value-return mechanism in DeFi. The Assistance Fund channels 97% of protocol trading fees into continuous, automated HYPE purchases. Per crypto.news and AMINA Bank research:
The Assistance Fund is not a corporate entity. It is an automated on-chain mechanism with no manual team intervention. The purchases are funded entirely by trading fees — no token issuance, no treasury depletion, no external capital. This is the closest DeFi has produced to a self-funding, perpetual buyback.
Corporate structure note: Hyperliquid Labs, the development entity, has not raised external venture capital. The protocol retains minimal governance — there is no DAO. Value flows directly from traders to the buyback mechanism to token holders. The absence of a venture-backed corporate layer is structurally significant.
Aave's governance approved a permanent buyback program allocating $50M per year from protocol revenue to repurchase AAVE tokens. Per BeInCrypto and CryptoDaily:
The gap is stark. Aave generates nearly $1B in annualized fees but passes a fraction to holders. The $50M annual buyback represents roughly 5% of gross fee revenue. The remaining 95% funds the protocol treasury, safety module, and operational expenses managed by the Aave Companies (the corporate entity behind the protocol).
According to CryptoDaily's analysis, Aave exemplifies the structural gap: high protocol fees, modest token holder accrual.
Sky has deployed approximately $114.5M in buybacks as of March 2026, purchasing ~1.83B SKY tokens, according to CoinDesk and CoinMarketCap.
The buyback is funded by protocol revenue from DAI/USDS stability fees — a sustainable, recurring source. Sky's transition from MakerDAO to Sky included a structural redesign of token economics, with the buyback as a central pillar.
Jupiter's Litterbox Trust received 50% of protocol fee revenue for JUP repurchases. Per crypto.news and CoinDesk:
Jupiter's case is instructive. $70M in buybacks did not prevent an 89% price decline. The Litterbox Trust retains tokens rather than burning them, meaning supply is reduced from circulation but not from existence. A governance proposal to increase the allocation and shift toward burn reflects holder frustration with the hold-not-burn model.
Uniswap activated its fee switch on December 28, 2025, becoming the largest DeFi protocol to redirect swap fees toward token value accrual. Per CryptoBriefing and Blockworks:
Governance Proposal 100 expanded the fee switch to v4 pools across all seven networks, nearly tripling daily revenue. The fee rate is 17% of swap fees. The UNI case is significant because it transforms a pure governance token into a revenue-accruing asset — a shift that, per Talos research, pressures every other DeFi governance token to justify its economic structure.
Ethena's fee switch went live after meeting activation benchmarks: USDe supply above $6B and cumulative protocol revenue exceeding $250M. Per Cryptopolitan and Tokenomics.com:
Ethena's fee switch is funded by the protocol's basis trading revenue — the spread between spot and futures positions that back USDe. This is real, recurring revenue, but it is cyclical. Basis trade yields compress in low-volatility environments, making Ethena's token holder distributions inherently variable.
Pendle executed one of the most significant governance overhauls of 2026, replacing its multi-year vePENDLE lockup system with flexible sPENDLE staking on January 20, 2026. Per The Block and CoinDesk:
The sPENDLE migration removed the primary friction point — multi-year lockups — while preserving the 80/20 fee split favoring holders. Revenue flows automatically to stakers without requiring manual gauge voting. Existing long-term lockers received up to 4x boosted rewards as transition compensation.
Why this matters: Pendle's structure distributes 100% of protocol revenue to token holders, with zero allocation to the protocol treasury. This is the most aggressive holder-first fee model in DeFi. The tradeoff: no treasury buffer against future operational needs.
Maple, the institutional on-chain lending protocol, pivoted from staking to buybacks via governance proposal MIP-019. Per Maple Finance and VaasBlock research:
Maple's 25% revenue-to-buyback ratio is conservative relative to Hyperliquid's 97% or Pendle's 80%. However, Maple generates revenue from institutional credit — undercollateralized loans to vetted borrowers — a fundamentally different and arguably more durable revenue stream than trading fees. The corporate entity (Maple Labs) retains the remaining 75% for operations and growth.
Treehouse, a DeFi protocol with $294M in deposits, launched a TREE token buyback funded by 50% of fees from tETH (its Ethereum yield product). Per DL News and Yahoo Finance:
Treehouse is a useful comparison point because of its scale. At $294M in deposits, it demonstrates that buyback programs are not exclusive to billion-dollar protocols. The 50% revenue allocation mirrors Jupiter's initial structure, and the DAO treasury retention (rather than burn) follows the same hold-not-destroy model.
The central question: where does the money go?
| Protocol | Gross Fees (Ann.) | To Token Holders | Holder Share | Mechanism | |---|---|---|---|---| | Hyperliquid | ~$975M | ~$595M (Q2 ann.) | ~61% | Automated buyback (97% of fees) | | Pendle | Undisclosed | 80% of revenue | 80% | sPENDLE auto-distribution | | Uniswap | ~$892M | ~$119M (ann.) | ~13% | Fee switch + burn | | Ethena | ~$732M | Variable (4.5–15% on staked) | ~10–15% est. | sENA staking yield | | Aave | ~$951M | ~$50M (buyback) | ~5% | AFC buy-and-distribute | | Sky | Undisclosed | ~$114.5M cumulative | N/A | Revenue-funded buyback | | Maple | Undisclosed | 25% of revenue | 25% | SSF buyback | | Jupiter | Undisclosed | 50% of fees | 50% | Litterbox Trust (hold) | | Treehouse | Undisclosed | 50% of tETH fees | 50% | DAO treasury hold |
The data shows a clear spectrum. Hyperliquid and Pendle sit at the holder-maximalist end, directing 61–80% of value to token holders. Aave, despite generating nearly $1B in annual fees, allocates approximately 5% to holders. The corporate entities behind protocols — Aave Companies, Maple Labs, Uniswap Labs — retain the capacity to capture value through equity, service agreements, and operational budgets that sit outside the on-chain fee flow.
Per Ancilar's allocator framework, the distinction between fee capture, protocol revenue, and holder accrual is now the defining analytical framework for DeFi valuation.
DeFi's buyback era has produced $18.8B in headline activity, but the data tells a more nuanced story. The majority of that figure comes from exchange token burns, not protocol-to-holder distributions. Among DeFi protocols specifically, the total buyback volume is closer to $2.8B — material, but not the paradigm shift the aggregate number suggests.
The protocols that deliver real value to token holders share three characteristics: revenue funded by recurring fees (not treasury depletion), automated on-chain mechanisms (not committee discretion), and burn or permanent lockup (not treasury retention). Hyperliquid satisfies all three. Pendle satisfies the first two via direct distribution. Most others satisfy one or none.
The structural gap between protocol revenue and token holder cash flow remains the defining feature of DeFi tokenomics in 2026. Six protocols generated $7.42B in H1 revenue; the share reaching token holders is a fraction of that. Until governance structures force a higher pass-through rate — or token holders vote with their capital — the buyback narrative will remain more compelling on announcement day than in the quarterly data.