Aave Labs: $86 Million, 23% of the Token Supply, and this is their Track Record
Date: 2026-02-25
Author: Marc Zeller, founder of ACI
ACI published a full transparency report. Every dollar the DAO paid us. Every dollar of revenue we attributed. On-chain verified, independently reproducible, 3.4 cents per dollar of revenue growth. We started with ourselves because every entity spending DAO funds should answer three questions: what did you deliver, what did it cost, and what was the return.
This article answers those three questions for Aave Labs.
Labs has received $86M in total capitalization: $16.2M from the 2017 ICO, $32.5M from VC rounds, $31.93M in direct DAO payments, and ~$5.5M in unapproved swap fees. The founding team retained 23% of the LEND token supply at the 2017 ICO, which converted to AAVE at 100:1 during the 2020 migration. Current AAVE holdings are undisclosed. Not all of that is DAO money. But Labs was not building from zero when they started asking the DAO for funds. They were already capitalized with $48.7M in ICO and VC proceeds. The DAOâs $37.4M came on top. The âAave Will Winâ proposal asks for $51M more. Labs publishes monthly development activity updates. No accountability report has ever been published: no cost-per-outcome breakdown, no financial disclosure, no wallet transparency.
Hereâs the track record.
1. The Product Graveyard
Labs tried to build beyond the core protocol. Standalone products. Separate ventures. Hereâs what happened:
The Product Graveyard â Six products, all failed or unprofitable
Six products. All failed or unprofitable. The strongest one (GHO) had to be rebuilt by others after Labs delivered a version that could not hold its peg.
-
Case Study: Horizon â $500M TVL, $24 Spent Per $1 Earned
The one number that looks good on paper. On-chain query of the Horizon pool via archive node tells a different story.
The pool has 9 listed reserves but only 7 with any supply. Using the Aave oracle for price-adjusted values: ~$466M in total supply. Of that, 69% ($322M) is stablecoins (RLUSD $214M, GHO $80M, USDC $28M). Only 31% ($144M) is RWA collateral, and 94% of that is a single asset: USCC (Superstate Credit, $135M at $11.48/unit). The remaining RWA: VBILL $6.2M, JAAA $2.5M, USTB $0.4M. USYC and JTRSY: zero supply.
The RWA collateral is real and it is being used as intended: 15 addresses supply USCC as collateral and borrow stablecoins against it. Thatâs the product working. But the numbers around it do not hold up.
Scanning all 354 addresses that ever interacted with the pool (220 with active positions): the three largest positions are a single RLUSD depositor ($140M, zero borrows, pure incentive farming), the GHO DirectMinter ($79.5M aGHO, zero borrows), and one USCC whale ($56M USCC collateral, $45M in stablecoin borrows). Three positions, 59% of the pool.
Remove the incentive-farmed RLUSD and the idle GHO, and the actual RWA lending market is ~$135M USCC collateral backing ~$86M in RLUSD borrows and ~$17M in USDC borrows across 15 addresses. Thatâs a $135M single-asset, single-issuer market. Not a diversified RWA marketplace.
In February 2026, headlines reported Horizon hitting â$1 billion in RWAs.â The Horizon weekly highlights tell a different story: $164M in actual RWA supply, $335.7M in stablecoins. The $1B figure includes total deposits: RWA, stablecoins, and idle GHO. The number that matters, cumulative revenue in the collector, is still ~$216K regardless of TVL.
GHO variable debt: $0.30. Three-tenths of one GHO. 100M GHO was minted via the DirectMinter, 79.5M sits idle in the pool, and nobody is borrowing it.
Cumulative DAO revenue in the Horizon collector: ~$216K (55K aGHO + 57K aUSDC + 104K aRLUSD). Since launch on August 27, 2025, the DAO has spent ~$4.2M in direct Merkl incentives (peaking at $43K/day in November, currently $23K/day). On top of that, every GHO in circulation costs the DAO money to maintain via the sGHO savings rate (~$14.5M/yr across 527M total GHO supply). Horizonâs 79.5M idle GHO is 15.1% of total supply, bearing ~$1.05M of that cost over 5.8 months. Total cost since launch: ~$5.25M.
For every dollar of revenue, the DAO has spent twenty-four dollars.
The original proposal included a new token with only 15% allocated to the DAO (85% to Labs and VCs). Community revolt killed it in a 108-reply Temp Check. They dropped the token, agreed to a 50/50 revenue share. At the AIP stage, BGD raised governance concerns: the âOperationalâ and âExecutiveâ roles were defined backwards, with BGD concluding âthere is a pretty strong case to be made that roles are factually reversed.â Labs dismissed the concerns as âsurprisingâ and said they would âmove forward as planned.â ACI voted no. Ignas voted no, calling it âa call for reunification.â The AIP passed anyway: 583K FOR vs 378K AGAINST. On-chain analysis shows a single wallet (0xea0c12fdâŚ, the same 333K AAVE Multisig delegation identified in the COI vote cluster below) cast 57% of all FOR votes. Without it, the vote fails 250K vs 378K. The founderâs own voting power overrode community opposition on his own proposal.
Horizon could have been built as a Core market configuration. The separate instance was not a technical necessity. It was an extraction mechanism.
The Avara brand itself was retired on February 3, 2026. The same day Family was shut down. Lens operational leadership was transferred to Mask Network in January. Stani returned his full attention to Aave because everything Labs tried outside Aave failed. They could not raise for Lens. Family did not find users. The strategy now is to maximize Aave DAO treasury extraction.
This matters for the âAave Will Winâ vote. The proposal asks the DAO to fund four new products: Aave App, Aave Pro, Aave Card, and Aave Kit. Growth grants of $5M each (plus $2.5M for Kit). From a team with a 0-for-5 record on standalone products. The DAO is betting $17.5M that this time will be different. No explanation of what changed.
2. Who Actually Built Aave
I want to be precise about attribution, because attribution determines pricing.
Labs built Aave V1. Labs built Aave V2. Labs delivered the initial V3.0 codebase. These are real contributions. The early protocol exists because of that work, and I have never claimed otherwise.
But the engineers who built it all left. Ernesto Boado (CTO) and Andrey Kozlov (Head of Backend/Frontend) departed Aave Labs in late 2021. Emilio Frangella (Head of Smart Contracts) followed in 2022. The ex-COO of Aave wrote at the time: âI worked alongside them for the past 4 years and they are the Aave Protocol.â
Emilioâs commit history tells the story. He made 1,051 commits to Aaveâs GitHub through V3.0. His last substantial commits landed on March 15, 2022 â the day before V3.0 launched. After that: zero commits for three months, a handful of bugfixes through November 2022, then permanent silence. Zero commits to the aave org since. He eventually returned to Labs in a management role, VP of Engineering, then Head of Engineering. Not the same as having the engineer who wrote the code still writing code.
Every genesis protocol engineer left Labs. Emilio returned as Head of Engineering after eight months of zero commits. The team that claims credit for V1 through V3.0 is not the team asking for $51M.
V3.0 was the last major protocol version Labs delivered. Everything since was built by DAO service providers.
Who Built What â Labs vs BGD version history
The DAOâs service provider ecosystem built that $140M+ in 2025 protocol revenue. BGD wrote the protocol upgrades (V3.1 through V3.7, Governance V3, Umbrella), helped design Chainlink SVR, added multiple defense-in-depth security layers, and simplified the codebase. Chaos Labs and LlamaRisk calibrated every risk parameter on every market on every chain. Without continuous risk management there is no revenue. There is a protocol waiting for the next exploit. If V3.0 had been deployed and never touched again, Aave would not have survived the market conditions that generated $140M. It would have accumulated bad debt and lost user trust. Chaos Labs alone authored 506 forum topics and deployed 257 on-chain payloads, more governance output than any other service provider. TokenLogic managed treasury operations, rebuilt GHOâs peg stability, and ran the BD that brought Mantle, Linea, and a dozen other chains. ACI ran governance coordination and growth strategy. Every dollar of that revenue required every one of those teams. Labs contributed zero to this era.
V3 generated over $140M in protocol revenue in 2025. It is the most proven lending infrastructure in DeFi. DAO service providers built every version from 3.1 through 3.7, rewrote governance, overhauled the safety module, calibrated the risk parameters, and managed the treasury. Labs cannot claim the upside on a revenue engine they did not build.
The TokenLogic revenue data makes the split explicit. During the V3.0 era (March 2022 through mid-2024, before the DAOâs upgrades reached critical mass), Aave V3 generated $3.33M in DAO revenue. Total. After DAO service providers rebuilt it into V3.1 and beyond, the same protocol generated $179M. Thatâs 98.2% of all V3 revenue, generated on protocol versions shipped by DAO service providers, not Labs. The code was BGDâs, but the revenue required the full stack: risk parameters from Chaos Labs that enabled safe leverage, treasury management from TokenLogic that grew GHO to $12.7M/yr, growth strategies from ACI that brought chains and integrations. The DAO paid Labs $16.28M for V3.0. That version produced $3.33M before it needed to be improved by someone else. The $140M+ everyone points to? The DAOâs work.
The code quality gap is documented. Labs delivered GHO V1, and it traded below peg from day one. The DAOâs service providers spent months fixing it. BGD improved V3.0 into seven subsequent versions, Chaos Labs calibrated the risk parameters, TokenLogic rebuilt the peg stability mechanisms.
BGD Is Leaving Aave
On February 20, 2026, BGD Labs announced its departure from Aave effective April 1, 2026. The team that built the V3 revenue engine is gone.
The deterioration was visible for months. BGD publicly abandoned its Project E initiative citing unfair conditions. Co-founder Ernesto Boado publicly condemned Labsâ governance conduct as âbreaking all codes of trust with the community,â calling the rushed vote submission âdisgraceful.â But the departure announcement itself is the clearest indictment. BGDâs stated reasons read like a summary of this article:
âAave Labs believes that the whole Aave DAO and contributors should pivot in the direction they believe inâ without adequate consideration of existing contributorsâ expertise.
âEvery time we think/will think about improving v3, there will be some type of implicit/explicit artificial constraint.â
They cite Labsâ control over brand, communications, and voting power as creating âstructural imbalances difficult to overcome.â They describe V4 marketing through ânegative v3 comparisons,â V4 development âwithout broader collaboration,â opposition to V3 improvements, and aggressive V3 deprecation timelines âdespite v4 remaining unreleased.â
This is not ACIâs assessment. This is BGDâs, in their own words, in a departure letter.
The consequences are immediate. Who maintains the V3 codebase after April 1? Who reviews V4, which ships with zero contribution from the only team that understands the protocol at a foundational level? If other SPs are asked to work on V4, they would be linking their name and reputation to a codebase that has not been independently validated. The âAave Will Winâ proposal asks the DAO to ratify V4 and freeze V3 features. That means migrating from the revenue engine BGD built to an untested system Labs built alone, with the builders of the original system walking out the door.
3. Trashing the Revenue Engine to Sell the Replacement
V3 generated over $140M in protocol revenue in 2025. It secures $30B+ in deposits. Every dollar the DAO collects comes from V3. Labsâ own AWW proposal admits it: âAave V3 already generates over $100 million in annualized revenue.â
The same proposal calls V3 âapproaching its architectural limitsâ and asks the DAO to âpause any new features for V3.â V4, still on testnet with zero revenue and $13.5M already spent, is presented as the only path forward. Labs is requesting $51M to build the replacement for a protocol they are simultaneously trying to freeze.
This did not start with the AWW proposal. It built over sixteen months.
The Timeline
October 2024. Stani on BGD Phase 4 (source):
âI personally do not recommend too many changes to the existing V3 codebase unless these upgrades are necessary or involve clear security improvements.â
Measured language. Framed as caution. The first public signal that V3 development should slow.
March 2025. BGD proposes V3.4. Stani and Emilio oppose it across 16 posts.
Stani (source):
âI donât recommend to proceed with this current upgrade given they donât materially change main usage of the protocol, thus the risk reward is not there. I do think these development resources should be rather used on other, more high value initiatives instead.â
âOther initiativesâ means V4. Stani escalates (source):
âThis direction of frequent protocol updates is playing with fire.â
âI also donât see a point of altering a codebase that is going to be deprecated overtime similar to V1 and V2.â
He calls V3 âdeprecated.â The codebase that generated over $140M in 2025. Emilio backs him (source):
âOne simple mistake is enough to kiss goodbye to the whole thing.â
Then the pivot to V4:
âV4 was designed and developed to address almost all of the aspects in which V3 could be stronger. I have no doubt V4 will âcannibalizeâ V3 given that itâs simply much better.â
V3.4 eventually passes, but only after Labs extracts a concession: Stani âtemporarily pre-approvesâ and mandates that Labs co-authors an Upgradability Framework governing when V3 can be updated. Labs now co-owns the rules for touching the protocol they did not build.
September 2025. BGD publishes a V3.X long-term vision. Stani responds with a 10-point manifesto against V3 development (source):
âI will be voting against if any such proposal is put forward.â
âContinuing development of V3 is a distraction.â
âEvery dollar spent on V3 is one not spent accelerating V4 adoption.â
âContinuing to invest in V3 beyond maintenance undermines governance credibility and wastes treasury resources.â
âIf BGD introduces proposals to continue developing V3 in a way that competes with V4, I will vote against it.â
Every dollar the DAO spent on V3 development went to BGD. Every dollar redirected to V4 goes to Labs. The five quotes above are not a technical position. They are a budget reallocation strategy.
In the same thread, Emilio attacks Marcâs credibility and dismisses every improvement BGD shipped (source):
âPretty much all of the improvements BGD applied are if anything symptoms of an aging protocol.â
He takes credit for V3 in the same post where he dismisses it. âAs lead architect of the first iteration of V3â begins the paragraph that reduces every subsequent version to maintenance on âan aging protocol.â Section 2 documented that 98.2% of all V3 revenue ($179M) was generated on versions shipped by DAO service providers, not Labs.
October 2025. Stani on BGD Phase 6 (source):
âAs Aave V3 and its ecosystem mature, the work will naturally shift more toward full maintenance rather than continuous feature overhauls.â
He frames BGDâs future as diminishing maintenance work on a declining asset. Then suggests they should âbranch into product development (ramping down servicing the Aave Protocol).â Translation: the team that built the revenue engine should go find something else to do.
January 2026. Staniâs âHow AAVE will winâ blog post (source):
âThe level of innovation has not increased over the past couple of years, and this means lost opportunities for the protocol.â
âThe level of innovation has not increased.â During the period Stani describes as stagnant, BGD shipped Liquid eMode (V3.2), which unlocked the LRT/LST loop generating $37M/yr in protocol revenue. They shipped new eModes (V3.6), which enabled every correlated-asset market the protocol now runs. They helped design Chainlink SVR, added multiple defense-in-depth security layers, and simplified the codebase radically. V3 revenue grew from $3.33M to over $140M in 2025. Stani never consulted BGD on the current state of V3 before making these claims. The âlost opportunitiesâ framing erases $137M in annual revenue growth to set up a $51M ask.
February 2026. The AWW proposal lands (source):
âAave V3 has served the protocol well, but it is approaching its architectural limits.â
âIt also makes sense to pause any new features for V3 if this framework is passed.â
V4 ratification and V3 feature freeze bundled with $51M funding. One vote.
The Contradiction
The âlindyâ argument is the clearest tell. Staniâs position against V3.4 (source):
âEach time the codebase is changed, the lindy effect restarts since new code is introduced. This is not only the perception but also the reality.â
If every upgrade resets lindy, then migrating to an entirely new codebase resets it to zero. V3 has four years of battle-testing with $30B+ secured. V4 has zero. The argument Labs uses to freeze V3 is the strongest argument against rushing to V4.
The argument is dishonest on its own terms. Lindyness is architectural stability and battle-proven core logic, not line counts. V3âs architecture has not changed. The improvements are additive. Stani knows this. The CTO of the protocolâs founding entity publicly arguing that additive improvements reset security is not a technical position. It is a market signal designed to erode confidence in the $30B system generating $140M/yr, to create urgency for its replacement.
The DAO voted for V4 development in June 2024. Labs cites this vote repeatedly to justify blocking V3 work: âThe DAO already paid for V4.â But the DAO voted for V4 development, not V3 deprecation. No proposal has ever asked the community to freeze V3 features. Labs treats a development contract as a mandate to shut down the current flagship product and revenue engine.
The Revenue They Dismissed
Labs does not just argue V3 should stop improving. They actively minimize the improvements that generate the most revenue.
Emilio on BGDâs V3 work (source):
âThe only changes that improved something that was directly introduced in V3 were liquid emodes (which improved the original eMode design, that was anyway functional and what made Aave what it is today in the first place) and LTV0 dynamics.â
âAnyway functional.â The original eMode was a per-asset system with increased parameters. BGD rebuilt it from scratch into generic groups of assets that achieve both capital efficiency and risk isolation through granular parameter control. The two systems share a name. They share nothing else architecturally. That redesign enabled the LRT/LST loop mechanic: supply weETH, borrow WETH, convert, repeat. The loop that took WETH borrows from $1.1B to $5.87B. The loop that generates $37M/yr in protocol revenue, 28% of every dollar the DAO collects. 97.6% of all WETH borrowing demand traces back to the stack that Liquid eMode made possible.
Stani takes it further. On ACIâs transparency report (source):
âAave Labs [âŚ] invented eMode (the architecture that enables the LRT and Ethena strategies referenced in your report).â
Labs âinvented eMode.â BGD rebuilt it into Liquid eMode. The revenue comes from the rebuilt version. Labs claims credit for the revenue while dismissing the rebuild as incremental.
Two people from the same entity. The same pattern: take credit for the foundation, dismiss the improvements, minimize the revenue impact. And it is not just BGDâs code they dismiss. The risk parameter work from Chaos Labs that keeps $30B+ in user deposits safe across every market. Without it there is no protocol, just an unmanaged pool waiting for the next bad debt event. The treasury management from TokenLogic that grew GHO from a broken stablecoin to $12.7M/yr in revenue. The BD from ACI and TokenLogic that brought Mantle, Linea, and a dozen chains onto the protocol. All dismissed as incremental. All generating the revenue Labs now claims credit for. Then argue the protocol should stop receiving improvements and the DAO should pay for a replacement.
The Financial Incentive
Labs built V3.0. DAO service providers improved it into the revenue engine. If V3 keeps improving under BGDâs stewardship, V4 becomes less urgent and Labsâ funding justification weakens. If V3 is framed as âapproaching limitsâ and development is frozen, V4 becomes the only path forward. Labs controls V4 development. V4 funding is Labs funding.
The entity requesting the funding is the same entity that has spent sixteen months publicly arguing that the revenue engine should stop receiving improvements. A funding strategy dressed as a technical disagreement.
BGD confirmed this in their departure announcement: âevery time we think/will think about improving v3, there will be some type of implicit/explicit artificial constraint.â The team that built V3 left because Labs made it impossible to keep improving it. The strategy worked. The builders are gone. V4 is the only option now.
The Endgame: Make V3 Unusable
The verbal campaign has a planned conclusion. From the AWW proposal itself (source):
âOnce V4 is mature, V3 parameters should be gradually adjusted to encourage migration, following the same approach used in past version transitions.â
âGradually adjusted to encourage migration.â In plain language: make V3 worse so users leave. Raise reserve factors. Steepen interest rate curves. Reduce capital efficiency. The same playbook used on V2.
Except the V2 precedent does not say what Labs claims it says.
V2 deprecation started in November 2022 and accelerated through 2025: frozen markets, disabled borrows, reserve factors pushed to 99%, slope2 set to 300%. But the conditions were entirely different:
- V3 was live and proven when V2 deprecation began. V3 had been on mainnet for over eighteen months and was already generating real revenue. V2 was a legacy system with $954K in accumulated bad debt.
- V2 deprecation targeted frozen markets with bad debt, not active revenue-generating pools. The goal was risk reduction on a system users had already voluntarily left.
- V2 deprecation was proposed by independent risk providers (Gauntlet, Chaos Labs) based on risk analysis. Not by the builder of the replacement system.
The V3 situation is the inverse on every dimension. V4 is on testnet. Zero mainnet deployment. Zero revenue. Two audits. V3 generated over $140M in 2025 with $30B+ secured. It is not a legacy system with bad debt. It is the entire revenue engine.
Labs is proposing to degrade the parameters of the protocol that generates all DAO income to push users toward a protocol that does not yet exist. The entity building the replacement is the entity proposing to break the original. Labs is using governance power to manufacture urgency for a product that cannot yet compete on merit.
I talk to every major LP and partner in this ecosystem. It is part of the job. Multiple institutional players have independently reported that Labs leadership has privately described the plan for V3 in terms far blunter than âgradually adjusted.â The word they heard was âunusable.â When institutional capital hears that the entity building V4 intends to make V3 unusable, they do not hear a migration strategy. They hear a threat to their existing positions. Several reached out directly to express concern. A protocol should never intentionally hurt its own users unless there is a critical security issue that forces the decision. There is no security issue here. There is a funding proposal.
The community has pushed back. One delegate proposed that V3 parameter adjustments should only begin â12-18 months at the very leastâ after V4 launch, âOR/AND when V4 reaches a milestone such as $15B deposits and a stress test is passed.â That is a reasonable, milestone-gated approach. Labsâ proposal contains no milestones. No conditions. No minimum V4 adoption threshold. Just âonce V4 is mature,â determined by Labs.
The pattern across sixteen months:
- Talk down V3: âdeprecated,â âaging protocol,â âapproaching its architectural limitsâ
- Block V3 development: oppose V3.4, co-author Upgradability Framework, threaten NAY votes
- Freeze V3 features: âpause any new features for V3â
- Degrade V3 parameters: âgradually adjusted to encourage migrationâ
Each step makes V4 more necessary. Each step benefits the entity that builds V4.
4. The Frontend That Breaks Launches
Labs has one product responsibility everyone can see: the aave.com frontend. Itâs the Face of the protocol, the interface every user touches, regular users judge the entiere Aave ecosystem by it.
It breaks on almost every major launch.
eModes 3.6: MegaETH rendered unusable. When BGD shipped the 3.6 upgrade introducing new eMode behavior, the frontend did not support it correctly. Assets that were non-borrowable outside an eMode could not be borrowed inside it on the interface, even though the protocol handled it fine. This made the entire MegaETH deployment unusable at launch. Same bug delayed Mantleâs frontend. EzR3aL flagged it publicly in February 2026: âThis instance in its current form is basically unusable. No looping, leverage, etc. possible right now.â The protocol worked. The interface didnât.
syrupUSDC: launched with configuration errors. When syrupUSDC was listed on the Aave Base instance in January 2026, the launch required a follow-up governance proposal to fix errors. Delegates reported the frontend also shipped without the token logo and without v3.6 LTV0 e-mode display. Basic launch requirements that were not met.
Base rate display: 2.5% shown as 0%. In December 2025, a delegate reported to the frontend team that the base variable borrow rate of 2.5% was displayed as 0% on the Aave frontend. A risk parameter visible to every user, displayed incorrectly. The protocol had the correct rate. The interface did not.
PT token logos: never updated on time. Every Pendle PT token batch that gets listed through governance has the same problem: the logos are not ready when the listing goes live. The same follow-up proposal that fixed syrupUSDC also had to correct âPTs for srUSDe e-mode label.â It signals a team that does not ship in sync with the protocol it supposedly maintains.
The pattern is the same every time. BGD builds the protocol upgrade. TokenLogic and risk teams configure the parameters. ACI shepherds the governance proposal and AIP creation. The asset gets listed, the chain gets deployed, the feature goes live. And then the frontend breaks it. The one link in the chain that Labs controls is the one that fails.
BGD builds it. TokenLogic configures it. ACI promotes it. Labs breaks the display.
Who Was Actually Maintaining the Frontend
What makes the broken launches inexcusable: ACI was doing Labsâ frontend work for free.
As I stated publicly during the CowSwap investigation: âACI has contributed heavily to the @AaveLabs frontend via our engineers @MartinGbz and @Nandy.eth, making dozens and dozens of PRs because we firmly believed that contributing to this frontend was contributing to the best interest of the Aave DAO. We agreed to bend the frontiers of our paid scope to support a company that was focused on ventures other than a lending protocol at the time.â
The GitHub record is public. NandyBa alone filed 52+ pull requests to the aave/interface repository. Token logos, incentive campaign configurations, bug fixes, reward token updates. Work that Labs was being paid to do. MartinGbz filed PRs for features Labs ignored for weeks, including LST native yield APY display (PR #2092, filed June 2024). Every token listing, every incentive campaign, every logo update that actually shipped on time during 2024-2025 went through ACI engineers submitting PRs to a codebase that Labs maintained.
When ACI stopped subsidizing the frontend, the quality collapsed. The eModes 3.6 bug that made MegaETH âunusable.â The syrupUSDC launch without a logo. The base rate showing 0%. These all happened after ACI withdrew frontend support.
The timeline matters. While ACI engineers were doing free QA on Labsâ frontend, Labs was building the CoW Swap integration that diverted swap fee revenue from the DAO to a Labs-controlled address (see Section 7). The frontend team was not fixing broken logos or supporting new eModes. They were building a private monetization layer on top of the interface ACI was keeping functional. When we stopped, the product broke. The only thing that improved was the revenue extraction mechanism.
5. The Business Development Track Record
Labs claims business development as a core competency. Institutional relationships, chain partnerships, protocol integrations. This is what they say they do better than anyone.
The record says otherwise.
MegaETH Deserved Better
MegaETH was a strong partner. Their team was competent and proactive. Their co-founder namik publicly chose the DAO governance process. They deserved the same quality of service Mantle received: proper incentive management, growth coordination, a working frontend. Instead, Labs took over and delivered none of it.
ACI ran the MegaETH relationship for nine months. We brokered a 5,000 ETH deposit from their token sale proceeds. We published a TEMP CHECK that passed in 10 days with 816.5K votes, then an ARFC with risk assessments from LlamaRisk and ChaosLabs. The process paused because MegaETH was on testnet. Thatâs the elected process working as designed.
In December 2025, Labs published a competing proposal. No Skywards process. No coordination with ACI, BGD, or TokenLogic. ACI learned about it when it appeared on the forum. MegaETHâs own co-founder namik chose the DAO process: âWe are currently working with Aave DAO service providers to shape an optimal proposal, aligned with the governance framework.â
I withdrew ACI BD resources. Labs owned it end-to-end from that point. No incentive management. No growth coordination. A partner that chose the right process got the wrong outcome.
Both MegaETH and Mantle deployed within 24 hours in February 2026. The MegaETH frontend could not handle eModes, rendering the instance âbasically unusableâ at launch. Today: MegaETH $19.4M in total deposits, $33K borrows (0.2% utilization, ~$0 in protocol revenue). Mantle $459M in total deposits, $150M borrows ($410K/yr in annualized protocol revenue). Same protocol. Same week. The variable was who managed the launch.
Staniâs response on X: âLet me clarify this misinformation. First of all MegaETH team was underserved, ghosted multiple times and asked for help so we jumped in, and good that we did as we got $10M guaranteed revenue for the Aave DAO. a thank you would be more appropriate here.â
âUnderserved, ghosted multiple timesâ: the forum shows a TEMP CHECK that passed in 10 days and an ARFC with two risk assessments. The process paused because MegaETH was on testnet. âWe jumped inâ: Labs bypassed Skywards with no coordination, and MegaETHâs own co-founder sided with the DAO process. â$10M guaranteed revenueâ: the deployment launched with a broken frontend and sits at $19.4M in total deposits with ~$0 in protocol revenue. Mantle, managed by the DAOâs service providers, hit $459M generating $410K/yr the same week. The MegaETH team did everything right. Labs failed them.
Coinbase: Labs Failed to Compete
Coinbase is the biggest institutional on-ramp in crypto. Labs owned this relationship. They failed to close it.
Coinbaseâs crypto-backed lending product runs on Morpho. Not Aave. Labs had the incumbent advantage, the brand recognition, and the relationship. They could not close. Labs had every advantage and still lost.
The consequences of that failure keep compounding. In February 2026, Coinbaseâs lending product processed its largest liquidation wave since launch. Approximately $170M in seven days. All executed via Morpho. That was Aaveâs liquidation revenue to lose.
Apollo Global Management, $938B AUM, made the biggest TradFi-into-DeFi investment of the cycle. They signed a cooperation agreement to acquire up to 90M MORPHO tokens over four years, roughly 9% of total supply. Labs managed that institutional BD relationship. The outcome went to the competitor. Two of the biggest names in institutional finance chose a protocol with a fraction of Aaveâs TVL because Labs could not close.
World Liberty Financial: âThe Art of the Dealâ
ACI sourced and won a governance-approved whitelabel deal with World Liberty Financial (WLFI) in October 2024. The terms: 7% of WLFI tokens plus 20% of protocol fees to the Aave DAO. It passed both Aave governance (892.8K votes) and WLFI governance. ACI supported it publicly. The deal was real.
Labs took over the relationship. In August 2025, when WLFI publicly denied the 7% allocation as âfake news,â Stani called it âthe art of the dealâ on X, quote-tweeting a WLFI post that has since been deleted, and claimed the deal was worth â$2.5 billionâ to the Aave DAO. AAVE dumped 8% on the news.
Fourteen months after governance approval, WLFI never launched the Aave instance. In January 2026, they launched âWorld Liberty Marketsâ on Dolomite, a competitor protocol. Dolomiteâs founder, Corey Caplan, is WLFIâs own CTO. The entity that Stani described as â$2.5 billion in valueâ to the DAO chose its CTOâs competing protocol instead. No governance vote on the switch. No explanation. The DAO received zero tokens, zero revenue, zero acknowledgment.
A community member asked the obvious question: âWhat happened with the World Liberty deal? Why didnât we win it? I remember Stani bragging on X about âthe art of the deal.â Where did that go?â
It went to Dolomite. Labs took over a closed deal from ACI, failed to make the terms enforceable, and the counterparty walked. âArt of the deal.â
The Pattern
Six major BD relationships. Zero clean wins for Labs.
BD Scoreboard â Six deals, zero clean wins
Now compare that to what the DAOâs own service providers won without Labsâ involvement. TokenLogic led the BD relationships for Linea (Consensys), X Layer (OKX), Bybit, Mantle, the Rabby wallet integration, and the MetaMask integration. These were not handed down from Labs. TokenLogic sourced them, ran them, and closed them. The Mantle numbers are in the table above. The rest are live, generating revenue, and required little Labs involvement beyond social media amplification.
TokenLogicâs total compensation from the DAO: approximately $3.9M over two and a half years. Labs is asking for $51M in Year 1 alone. The team getting paid 13x less is the one closing deals.
The pattern is consistent. Labs takes over BD relationships, sometimes from pipelines ACI originated, and the results degrade. When ACI or Tokenlogic manage the same type of deal, the results are measurably better. This is not a matter of opinion. It is a matter of TVL and revenue.
The TradFi pipeline is where this hurts most. Labs does not just do institutional BD. They claim it as their defining advantage. From the AWW proposal (source): âAave Labs invests heavily in partnering with the largest institutional and fintech names to integrate with Aave.â From Staniâs âHow AAVE will winâ post (source): âThis insight comes from hundreds of hours of conversations with institutions.â From Stani on the BGD V3.X thread, where he name-drops NASDAQ tickers to bolster credentials (source): âThere are also large institutional players like Galaxy Digital (NASDAQ: GLXY) and BTCS (NASDAQ: BTCS) that have been major borrowers with contributions from Aave Labs⌠where I had to be personally involved.â
Hundreds of hours of conversations. Personal involvement. The largest institutional names. This is the pitch. Hereâs the scorecard: Coinbase chose Morpho. Apollo chose Morpho. Both relationships were Labs-managed.
Labs had the relationships, the brand, the incumbent advantage, and eight years of market presence. They lost every major institutional deal to a team of twenty-somethings with no prior protocol deployment and no track record in institutional finance. Coinbase chose the competitor. Apollo chose the competitor. The protocol with 10x the TVL and the most recognized name in DeFi lending could not close. The DAO pays twice: once for the failed execution, once for the lost opportunity cost of blocking other service providers from those same pipelines.
6. Governance: Dead Last in Both Columns, 100% of Obstruction
Two measures of governance participation, applied equally to every entity. Forum topics: all topics authored on the Aave governance forum, excluding monthly development updates and product marketing (applied consistently; only Labs publishes these). On-chain payloads: all proposals deployed from each entityâs confirmed address(es). Immutable, on-chain, verifiable.
Governance Participation â Dead last in both columns
| Entity | Forum Topics | On-chain Payloads |
|---|---|---|
| ChaosLabs | 506 | 257 |
| ACI | 431 | 575 |
| TokenLogic | 269 | 84 |
| BGD Labs | 129 | 186 |
| Aave Labs | 47 | 43 |
Labs is dead last in both columns. The DAOâs largest funding recipient is its least active governance participant.
Of Labsâ 47 forum topics, 18 are self-serving: 8 events/sponsorship budget requests, 4 direct funding proposals, 3 Horizon proposals, 1 branding proposal, 1 roadmap, 1 AWW Framework. The remaining 29 include GHO-related work spread across multiple temp checks and ARFCs for the same feature.
Labsâ 43 on-chain payloads are almost entirely GHO-specific or self-serving: GHO cross-chain launches and maintenance (~15), GHO stewards/stability/incidents (~8), events budgets (~5), their own SP proposals (~3), Horizon (~3). Zero protocol infrastructure payloads. Compare BGDâs 186: V3 upgrades, chain deployments, Umbrella, governance payloads for the entire ecosystem.
BGD, a pure engineering team with no governance mandate, produced 2.7x Labsâ forum output and 4.3x its on-chain output. ChaosLabs produced 10.8x. The pattern: show up to ask for money, disappear until the next ask.
The Committee Ghost
Proposals are one measure. Operational governance is another. The DAO runs on committees: the Aave Liquidity Committee (ALC), the GHO Stewards, and Aave Finance. These committees execute hundreds of multisig transactions per year managing GHO liquidity, borrow rates, incentive deployments, and treasury operations.
The ALC has operated two Safes since its creation in late 2023. The original Safe executed 438 transactions. Hereâs who actually created transactions:
| Proposer | Transactions Created | % of Total |
|---|---|---|
| Matthew Graham (TokenLogic) | 232 | 53.0% |
| Sisyphos (karpatkey) | 92 | 21.0% |
| TokenBrice (DeFi Collective, unpaid) | 44 | 10.0% |
| Marc Zeller (ACI) | 35 | 8.0% |
| Figue (Paladin) | 25 | 5.7% |
| Emilio Frangella (Aave Labs) | 0 | 0.0% |
Labsâ Head of Engineering never proposed a single transaction. Not once across 438 executions. The second Safe: 347 transactions, zero proposed by Labs.
His passive contribution was not much better. He confirmed 63 of 438 transactions. 14.4%. The next-lowest active signer was at 55%. His last signature was September 16, 2024. He was fully removed from both ALC Safes on July 18, 2025 and his seat was given to LlamaRisk. The ALC update proposal noted that âthe 4 of 6 requirement has proven onerous due to various contributors tending to other Aave DAO related commitments.â One signer was consistently unavailable. The threshold was lowered to 3 of 5.
TokenBrice was an unpaid volunteer from the DeFi Collective. He proposed 44 transactions and signed 74 in roughly three months before departing. No compensation. No mandate. He showed up because he cared about GHO liquidity. An unpaid volunteer did more operational work in three months than Labsâ Head of Engineering did in two years.
GHO Stewards is where this gets indefensible. The committee manages GHOâs borrow rate, borrow cap, GSM exposure, and fee strategy. GHO is Labsâ own product. The Stewards are ACI, ChaosLabs, karpatkey+TokenLogic, and previously Gauntlet. Labs is not represented at all. The entity that built GHO does not participate in its operational governance.
Aave Finance, the committee managing the DAOâs treasury operations, shows the same pattern. Labs has no representative participation.
The Culture
Numbers tell you what happened. Culture tells you why it keeps happening.
Emilio Frangella, Labsâ Head of Engineering, has 237 forum posts over five years. A consistent pattern: condescension toward anyone questioning Labs, personal attacks on competence, dismissal of governance scrutiny as uninformed noise.
On BGDâs V3.x vision thread, September 2025:
âIt seems that Zellerâs favorite strategy is to throw uninformed shades at other peopleâs work hoping to gain cheap popularity.â [âŚ] âwhere he clearly lacks any basic knowledge.â [âŚ] âI will try to educate him appropriately (again, as i was the first one teaching him basic smart contract development many years ago) and debunk this embarrassment.â
(source)
On the CowSwap integration thread, December 2025:
âIt would be great if for once ACI could validate their claims onchain instead of creating pointless sensationalism.â [âŚ] âhopefully this trend of having to perpetually educate ACI members will stop at some point.â
(source)
On the Visual Identity thread, May 2024:
âI wont even comment on the alleged âbrand colonialismâ, i dont even know what that means, must be a french thing to always turn discussions over to colonialism or dominance.â [âŚ] âIt is, quite frankly, mediocre looking to say the least.â
(source)
The âI will educate youâ framing shows up over and over. Labs positions itself as the authority and everyone else as students who have not done the reading. This is not how a service provider addresses the DAO that funds it.
Forum Activity Patterns
A data-driven analysis of the governance forum reveals a pattern around Labs employee engagement.
The Aave governance forum has a public Discourse group called Aave-Labs-SP with 33 members. The groupâs creation timeline is telling: 23 of 33 members (70%) were batch-added on December 17-18, 2025, exactly one week into the CowSwap revenue scandal thread. Three more were added February 6-12, 2026, immediately before the AWW proposal launch. The timing correlates perfectly with the two moments when Labs faced the most governance scrutiny.
Of the 33 members, 16 (48%) have zero forum posts. They have never written a single comment. Their only activity is reading and distributing likes. All 16 were added in the December 17-18 batch. In a governance forum where most posts receive 1-20 likes, a pool of 16 like-only accounts provides a meaningful engagement boost. Discourse likes are public. Anyone can click on the like count of any Labs post on a contested thread and see who liked it. The pattern is verifiable. This is coordinated engagement: manufacturing organic-looking support on contested threads. These are only the accounts we could link to Labs through forensic analysis of the public Discourse group. None of them disclosed their affiliation voluntarily. That is exactly what the COI framework should enforce. Without mandatory disclosure, the only way to identify affiliated accounts is forensics, and forensics only catches the ones that leave a trace. The actual number is likely higher.
The COI vote thread saw 6 distinct Labs group members arguing against disclosure rules. The BGD V3.X vision thread had Labs members authoring 36% of all posts.
Nobody is arguing that employees should not participate in governance. The question is whether Labs discloses these relationships. The COI proposal would have required exactly this kind of disclosure. Labsâ governance cluster voted it down with 663K VP.
Multiple delegates have independently described receiving private messages from Labs employees accusing them of conspiring with competitors after posting critical governance comments.
The Communications Monopoly
Labs controls the primary communications channels.
Discord. Labs controls the official Aave Discord server. Governance participants have reported being banned or timed out for asking questions about Labsâ operations. In at least one documented case, a delegate was banned for approximately seven days after posting governance criticism. No public review process. No appeal. One Labs employee made the call.
Events. Labs calls itself âthe events unit for the DAOâ and has pressured other SPs to rename community events. Governance participants have been told their events were âunofficialâ because Labs had not sanctioned them. Community organizing requires Labsâ permission. Thatâs not a DAO. Thatâs a franchise.
Social media. Labs controls the @aave X account and other official channels. The communications monopoly does not translate to operational wins. It translates to narrative control.
The double-speak. On February 18, the AWW proposal was published with explicit language about degrading V3: âV3 parameters should be gradually adjusted to encourage migration.â On February 20, BGD announced they were leaving, citing Labsâ artificial constraints on V3 development. Community panic about V3âs future spiked. On February 22, Labsâ Marketing Director posted on X that V3 would run âfor the foreseeable futureâ with âno rushâ to deprecate, and that V4 rollout would be âslow and responsible.â The proposal says one thing. The comms team says the opposite when sentiment turns negative. The audience is different: the proposal targets governance voters who will approve funding; the X thread targets the broader market that is selling the token. Same entity, opposite messages, four days apart.
When the entity asking for $51M treats scrutiny as insubordination, the DAO doesnât have a personnel problem. It has a governance problem.
7. $86 Million and This Is the Track Record
Thatâs the operational record. Now the money.
Lifetime Balance Sheet
Labs (previously Aave Companies, previously Avara) has received cash from four sources. Every figure below is from public filings, on-chain data, or governance proposals.
Lifetime Balance Sheet â $86M from ICO, VC, DAO payments, and unapproved fees
$86 million in cash. 23% of the original token supply, converted through migration, current holdings undisclosed. The team was well capitalized before they ever asked the DAO for a dollar. As Labsâ own proposal states: âThe LEND token sale in 2017 raised $16M to build a decentralized lending protocol.â What it omits: the team also kept 23% of the token supply, which appreciated alongside that protocol.
DAO Payments, Itemized
| # | Date | Proposal | Amount | Source |
|---|---|---|---|---|
| 1 | Sep 2022 | V3 Retroactive Funding | $16.28M ($10M stables/crypto + $6.28M AAVE) | AIP-98 |
| 2 | Oct 2023 | Events 2023 | $550K GHO | Snapshot |
| 3 | Jun 2024 | V4 Service Provider | $12M GHO ($3M up + $9M stream) | On-chain |
| 4 | Jun 2024 | Events 2024 | $850K GHO | Snapshot |
| 5 | May 2025 | Events 2025 | $750K GHO | Snapshot |
| 6 | Oct 2025 | V4 Security | Up to $1.5M | On-chain |
| Total confirmed | $31.93M |
Note on the V3 retroactive: BGDâs on-chain analysis found that the AAVE lock-up described in the forum post (50% staked for 1 year, 50% for 2 years) was never enforced on-chain. The proposal âsimply transfers AAVE to the Aave genesis team account.â The ~73-78K AAVE tokens were sent to a Gnosis Safe with no programmatic lock.
The CoW Swap Diversion: ~$5.5M Without a Vote
From 2022 through mid-2025, ParaSwap referral surplus from swaps on aave.com flowed to the DAO treasury. Established revenue. No user fee charged.
In mid-2025, Labs quietly switched to CoW Swap integration and began capturing 15-25 basis points in partner fees to a Labs-controlled address. No governance vote. No forum post. No announcement.
On-chain verification (Reth archive node, trace_transaction): 933 ETH on mainnet across 18 weekly distributions, July 2025 through February 2026. Rate accelerating from ~8 ETH/week to 204 ETH/week ($510K/week). Plus ~$2.3M from L2 chains. TokenLogic dashboard confirms ~$5.5M total.
Current rate: $510K/week and accelerating, diverted from the DAO without authorization.
The âAave Will Winâ proposal offers to send â100% of Aave Labsâ product revenueâ to the DAO. The Block, Unchained, and crypto.news all covered this story. This is not a gift. It is an offer to stop taking what was already flowing to the DAO before Labs diverted it.
Return on Investment
Hereâs what the money bought.
ROI Table â DAO cost vs revenue for every Labs product
The pattern: Labs delivers the foundation, someone else makes it profitable. V3.0âs $3.33M in revenue became $179M after DAO service providers rebuilt it. GHOâs depeg became $12.7M/yr after TokenLogic and ACI fixed it. The one product Labs still owns end-to-end (Horizon) has cost the DAO $24 for every $1 it earned. V4 has consumed $13.5M and produced nothing yet. The frontend generates $510K/week in swap revenue that goes to Labs, not the DAO.
SP Cost Comparison
SP Cost Comparison â Labs vs all other service providers
Labs is asking for $51M in Year 1. All other service providers combined cost $16.5M per year. Labs alone is 3x the entire rest of the DAOâs service provider budget. If the primary grant renews for three years, the total ask exceeds $100M ($75M primary + $17.5M growth + $8.85M AAVE). For a team that has not shipped production protocol code since V3.0, that lost every major BD deal, and that has never published an accountability report.
ACI cost the DAO $4.625M over three years and generated $142.9M in attributed annualized revenue. 3.4 cents per dollar. Labs has received $31.93M and has never published a comparable metric. They publish monthly activity updates listing what the team worked on, but only since preparing for their next funding round. They have never reported what the money produced. The DAO is being asked to increase Labsâ lifetime capitalization to $130M+ with zero accountability baseline.
8. The Governance Machine
This section draws from on-chain forensic analysis published during the COI vote. Every address is on-chain and independently verifiable.
663K Votes Against Transparency
The Mandatory Disclosures and Conflict of Interest proposal asked delegates to disclose their voting addresses and abstain when conflicted. A basic governance hygiene measure that most DAOs consider table stakes.
It did not pass. NAY 688,619 vs YAE 603,607.
On-chain forensic analysis traced 663,015 of those NAY votes, 96.3% of all opposition, to a single cluster connected to Aave founding infrastructure:
NAY Cluster â 663K votes traced to founding infrastructure
Without this cluster, the proposal passes by 578K votes.
The same 333K delegation (0xea0c12fdâŚ) was the swing vote on the Horizon AIP. It cast 57% of all FOR votes on a proposal where ACI, Ignas, and BGD raised explicit opposition. Without it, Horizon fails. The same governance power that blocked the transparency framework also forced through the founderâs own product launch.
The same cluster voted as a block on the AAVE Token Alignment proposal in December 2025. Same addresses, same delegation structures. The signer network spans 52 Gnosis Safes. Fifty-two. Not one wallet voting transparently, but a web of multisigs that requires forensic analysis to trace. Thatâs not how you hold governance power in good faith. Thatâs how you hide it.
75,000 AAVE = Governance Power
The âAave Will Winâ proposal asks the DAO to transfer 75,000 AAVE tokens to Labs. At $118, thatâs $8.85M. It also represents 13.6% of the DAOâs AAVE holdings (~550K tokens).
AAVE tokens carry voting power. This would transfer governance power to an entity whose holdings remain undisclosed. The same entity whose wallet cluster defeated the proposal that would have required disclosure. The DAO should know what governance power the recipient already holds before transferring more.
Who Controls the Voting Infrastructure
Labs holds the ENS names that control where votes happen.
The original Aave Snapshot space ran under aave.eth, owned by the genesis team. In August 2023, I confirmed publicly that âthe Snapshot space is the property of the ENS Aave.eth owner, this is @AaveLabs as this ENS name is extremely valuable and strategic.â
That same month, TokenLogic proposed migrating the Controller role from aave.eth to a contract governed by the DAOâs Short Executor. AaveLabs publicly supported the transfer: âWe support handing over the responsibility of the Snapshot Space to the Aave DAO.â It never happened.
In late 2024, the space migrated to aavedao.eth under the Dolce Vita Extension. When aavedao.eth expired in December 2025, Labs re-acquired it for 0.7 ETH, regaining ultimate controller access. Labs now holds both aave.eth and aavedao.eth.
The structural result is the same: the entity asking for $51M controls the infrastructure where the vote on that $51M happens. The 2023 proposal to put Snapshot under actual DAO smart contract governance was promised, supported, and never executed. Three years later, the voting infrastructure remains under centralized control.
The Bundled Vote
The FAQ section of âAave Will Winâ states: âVoting on these items separately could result in a fragmented and unworkable plan.â
Four proposals in a trenchcoat:
- Revenue alignment (broad community support)
- V4 ratification (on testnet, two audits, zero revenue, no independent review from BGD)
- Foundation creation (good idea, independence unverifiable under current conditions)
- $51M funding (the controversial part)
Bundling means accepting everything or rejecting everything. In a governance environment where one cluster controls 663K VP (enough to decide any contested vote), bundling reduces the communityâs ability to express nuanced positions. If these components stand on their own merits, they can be voted on separately.
The community has been clear. Delegates and tokenholders have asked Labs to unbundle the vote, disclose wallet holdings, reduce the budget, publish a transparency report, and establish the foundation before funding begins. Labsâ latest response (February 18, 2026) answers every one of these demands the same way: later. Revenue caps? âWe suggest setting these caps on a product-by-product basis in future proposals.â Foundation? âUnlikely to be live during the timeline between now and the final vote.â Independent verification? âWe are investigating.â KPIs? âWe are open to hearing additional KPIs.â And then the closer: âwe intend to keep the proposal moving through this phase.â Vote YES on the Temp Check now. Details at ARFC. Trust me bro.
This is the pattern. Every community demand becomes a post-vote commitment. Once the Temp Check passes, the direction is set and the leverage is gone. The DAO is being asked to approve $51M in funding before the foundation exists, before revenue caps are defined, before wallet holdings are disclosed, and before a single accountability report has been published. The community is not asking for unreasonable things. They are asking for preconditions. Labs is offering post-conditions.
Why Now
On February 20, BGD announced they are leaving. Effective April 1. Their stated reason: Labs made it impossible to keep building.
Before Christmas, the situation was painful internally but manageable externally. The protocol printed over $140M in 2025. BGD was shipping upgrades. ACI and TokenLogic were closing deals and growing the treasury. I kept the peace because the protocol was performing and the DAO model was working.
Thatâs over. Labs pushed a governance sequence that wiped over $1.4B off Aaveâs fully diluted valuation since AAVEâs December high of $205, while direct competitors appreciated. Sixteen months of trashing V3 to sell V4. Undisclosed voting power to force the framework through. And now the engineers who actually made V3 what it is walked out the door. When your best builders leave and cite your largest funding recipient as the reason, the DAO should listen. I wrote in December that talent would exit quietly at first, then all at once. BGD left on February 20.
Thereâs no status quo left to protect. The builders are gone. The token is underperforming competitors by 74 points. Labs is asking for $51M more while refusing every precondition the community has set. Staying quiet does not keep the peace. It makes me complicit in what comes next.
Closing
Three questions for any entity spending DAO funds: what did you deliver, what did it cost, and what was the return.
ACIâs answers: $4.625M over three years. $142.9M in attributed annualized revenue. 3.4 cents per dollar of revenue growth. Full report published. On-chain verified.
Labs will not answer, so we applied the same framework: $86M in cash from ICO, VC rounds, DAO payments, and unapproved swap fees. 23% of the original token supply, current holdings undisclosed. Zero major BD wins. A frontend that breaks launches. Six products, all failed or unprofitable. Governance participation on 2.5% of proposals. Zero transactions proposed across 785 ALC multisig executions. Not a member of the GHO Stewards committee that governs their own product. An engineering contribution to the current revenue engine that has been BGDâs work since V3.1. Monthly activity updates, no accountability report ever published.
BGD wrote the code. ACI designed the growth strategies. TokenLogic managed the treasury, built the revenue analytics, and ran GHO stewardship. Chaos Labs and LlamaRisk handled risk. The DAO built Aaveâs revenue engine. The engineers who built it just walked out the door.
The DAO is being asked to approve $51M for the entity that drove away its engineers, lost every major BD deal, broke the frontend on every launch, shipped six products that all failed or remain unprofitable, diverted $5.5M in swap revenue without a vote, and used 663K in undisclosed voting power to block the transparency framework that would have held them accountable.
The Slow Motion Coup Tax
Since the ownership proposal went live, the damage isnât just in the treasury. Itâs in your portfolio.
The Slow Motion Coup Tax â AAVE vs MORPHO vs SKY since Dec 19
Aave generates over $140M in annual protocol revenue. Morpho charges zero protocol fees. Aave has a live $50M annual buyback and is gaining market share. The token should be outperforming its direct competitor, not trailing it by 74 points. The market is saying the quiet part out loud.
Every number in this article is on-chain or from public sources. Judge the evidence.
Disclaimer: I am the founder of the Aave Chan Initiative, a delegate platform and service provider to the Aave DAO. ACI authored the COI analysis referenced in this article. Every address is on-chain and independently verifiable. The methodology is published. Anyone can reproduce the analysis. Every factual claim in this article is sourced to public forum posts, on-chain data, or governance proposals.
SPECIAL REPORT: Aaveâs Governance Crisis, Bleeding Capital, and Aave Labsâ Historic Mistake
The Data Doesnât Lie: Capital is Fleeing Aave
Aave Labsâ Cheap Politics and Suicidal Strategy
The Final Ultimatum: Buyout, Fork, or a Garbage Name