Aave, the largest decentralized lending protocol by total value locked at $14.4 billion, filed a governance proposal on July 29, 2026 to deprecate 75 asset reserves and exit six blockchain networks entirely. The affected positions total $98.1 million in supplied assets and $15.6 million in outsta...
"After a comprehensive review, Aave is deprecating 50 low adoption asset reserves across multiple deployments. In addition, Aave is orderly winding down deployments on Sonic, Scroll, zkSync, Metis, Soneium, and Aptos, covering another 25 asset reserves." — Stani Kulechov, Aave Founder, via X on July 29, 2026
Aave, the largest decentralized lending protocol by total value locked at $14.4 billion, filed a governance proposal on July 29, 2026 to deprecate 75 asset reserves and exit six blockchain networks entirely. The affected positions total $98.1 million in supplied assets and $15.6 million in outstanding debt. Three of the six departing chains generate under $1,000 in quarterly protocol revenue.
The proposal, prepared by risk provider LlamaRisk under Aave's newly adopted Risk Framework, marks the first systematic application of cost-benefit analysis to a major DeFi protocol's multichain footprint. It arrives as the broader DeFi sector contracts: over 40 protocols have shut down in 2026, DeFi-specific venture funding has fallen to its lowest quarterly deal count since 2020, and the number of protocols generating over $10 million in monthly fees has halved year-over-year.
For a protocol generating $907 million in 2025 revenue and $333 million year-to-date in 2026, the decision to exit chains producing less than $5,000 per quarter is arithmetically straightforward. The operational significance lies in what it signals: the multichain expansion era in DeFi is yielding to consolidation economics.
The governance proposal, posted as an Aave Request for Comments (ARFC) on the Aave governance forum, breaks into two components:
Component 1 — Individual Reserve Removals: 50 low-adoption asset reserves and 21 matured Pendle Principal Tokens across 11 Aave V3 deployments. These account for $85.3 million in supplied assets and $11.5 million in outstanding debt.
Component 2 — Full Deployment Wind-Downs: Complete exit from Sonic, Scroll, zkSync, Metis, Soneium, and Aptos, affecting 25 additional reserves with $12.8 million in supplied assets and $4.1 million in debt.
Combined, the cleanup touches 96 reserves (75 unique assets plus 21 matured Pendle PTs) across 17 deployments, representing approximately 0.68% of Aave's total $14.4 billion TVL.
The proposal requires an on-chain governance vote before any reserve is formally retired.
The six chains targeted for full wind-down share a common pattern: deposits peaked shortly after launch, then collapsed. According to the LlamaRisk review, every one of these deployments now generates under $5,000 in quarterly protocol revenue, with the three smallest producing under $1,000. The deposit declines over six months are severe:
| Chain | Peak Deposits | Current Deposits | Decline | Quarterly Revenue | |-------|--------------|-----------------|---------|-------------------| | Sonic | $28.9M | $7.6M | 74% | <$5,000 | | Scroll | $16.1M | $2.2M | 86% | <$5,000 | | zkSync | $7.2M | $844K | 88% | <$5,000 | | Metis | $1.4M | $297K | 79% | <$1,000 | | Soneium | $3.2M | $173K | 95% | <$1,000 | | Aptos | $18.0M | $1.0M | 94% | <$1,000 |
Aptos is notable: it was Aave's first non-EVM deployment, launched in January 2026 as part of an explicit multichain strategy. Within six months, 94% of its liquidity evaporated and deposits fell from $18 million to $1 million. Aave launched on Aptos to great fanfare; the wind-down was announced six months later.
Soneium, Sony's blockchain initiative, saw the steepest decline at 95%, falling from $3.2 million to $173,000. zkSync deposits dropped 88%, from $7.2 million to $844,000 — a figure that barely exceeds a single large user's position.
LlamaRisk's assessment is blunt: the oracle maintenance, risk parameter monitoring, and liquidation-path upkeep required for each deployment represent fixed costs that these revenue levels cannot justify.
Beyond the chain exits, 50 individual assets are being removed from surviving deployments. The largest positions affected:
Ethereum Core (BTC wrappers dominate):
Arbitrum:
Polygon:
Pendle Principal Tokens (Plasma):
The Pendle PT deprecation warrants context. These are yield-bearing tokens with fixed maturities; once expired, they serve no economic function. Their removal is housekeeping rather than a strategic judgment.
The more telling deprecations are the BTC wrappers on Ethereum. FBTC holds $11.1 million in supply but only $63,000 in borrows — a utilization rate below 0.6%. eBTC has $5.3 million supplied with zero borrowing demand. These assets exist on Aave's balance sheet as risk surface with negligible fee generation.
The proposal is the first output of a four-layer governance standard developed by LlamaRisk, Aave's risk service provider, following the $292 million KelpDAO bridge exploit in April 2026. That exploit — in which stolen rsETH was used as collateral on Aave to borrow $190 million in WETH against assets backed by nothing — demonstrated how listed-but-illiquid reserves create attack vectors disproportionate to their economic value.
According to LlamaRisk's methodology, each listed reserve carries a fixed operational load regardless of its size: an oracle to maintain, risk parameters to monitor, and a liquidation path that must function reliably. The framework evaluates reserves on a standardized cost-benefit basis and applies systematic reviews quarterly.
The framework does not target specific chains or assets by name. It establishes criteria — minimum utilization thresholds, revenue-to-cost ratios, liquidity depth requirements — and allows deprecation decisions to emerge from the data. The six chain exits and 50 reserve removals are presented as outputs of this process, not discretionary choices.
This approach mirrors traditional finance practices. Banks routinely exit unprofitable product lines, close branches in low-activity regions, and sunset instruments that no longer justify compliance overhead. The difference is that DeFi protocols have historically expanded without equivalent discipline, deploying to new chains based on grant incentives and ecosystem partnerships rather than projected fee revenue.
The scale mismatch between Aave's core business and its peripheral deployments is stark.
Aave generated $907 million in protocol revenue during 2025. Through mid-June 2026, year-to-date revenue reached $333 million, implying an annualized run-rate between $650 million and $800 million depending on market conditions. Standard Chartered initiated analyst coverage of the protocol in 2026 based on these figures.
In April 2026, Aave governance passed the "Aave Will Win" proposal with approximately 75% support, directing all product revenue flows into the DAO treasury. Every fee, interest spread, and income stream now flows to a single entity controlled collectively by AAVE token holders.
Against this revenue base, the six departing chains collectively generate under $20,000 per quarter — or roughly $80,000 per year. That figure represents 0.01% of Aave's annualized revenue. The oracle and monitoring costs for maintaining these deployments almost certainly exceed their contribution. According to LlamaRisk, the operational overhead per deployment — covering oracle feeds, liquidation bot infrastructure, risk monitoring dashboards, and parameter governance — runs in the tens of thousands of dollars annually at minimum.
The protocol's Ethereum deployment alone accounts for 82.7% of total TVL at approximately $11.9 billion. The concentration is not accidental; it reflects where lending demand, borrowing activity, and liquidation infrastructure actually exist at scale.
Aave's cleanup occurs within a wider contraction. According to data aggregated through mid-2026:
Other protocols have made similar exits. Magic Eden terminated support for Bitcoin and EVM networks in April 2026. DEX aggregator Odos shut down permanently on July 30, 2026. The pattern is consistent: protocols that expanded aggressively during the 2024-2025 multichain push are retreating to the networks where unit economics work.
The multichain thesis — that deploying everywhere maximizes user access and fee capture — has collided with the reality of fragmented liquidity. A lending protocol needs borrowers and lenders on the same chain, at sufficient scale, to generate sustainable interest rate spreads. A $297,000 deployment on Metis cannot produce this dynamic.
The proposal does not mandate immediate shutdowns. Instead, it prescribes a phased transition:
This approach differs from emergency shutdowns seen after exploits. It is designed to minimize user disruption and allow orderly position migration. Users on affected chains can withdraw their assets and redeploy to Aave's surviving, higher-activity deployments.
Aave's proposal to exit six chains and deprecate 75 reserves is the clearest empirical signal yet that multichain expansion in DeFi has reached its limit of economic justification. The data is unambiguous: chains generating under $1,000 per quarter cannot sustain the fixed costs of oracle maintenance, risk monitoring, and liquidation infrastructure.
The broader context matters. DeFi grew by deploying everywhere and incentivizing liquidity with token emissions. That model produced impressive TVL figures but fragile economics. As incentives expire and liquidity migrates to established venues, the gap between headline deployment counts and revenue-generating activity has become impossible to ignore.
Aave is the first major protocol to formalize this reckoning through a systematic framework rather than ad-hoc decisions. Whether other protocols follow — and whether chains like Scroll, zkSync, and Sonic can rebuild their DeFi ecosystems without Aave — will determine whether the multichain era ends in consolidation or collapse.
The arithmetic is simple. A protocol earning $907 million per year does not need six chains contributing $80,000 combined. The operational risk of maintaining them — demonstrated by the KelpDAO exploit's use of illiquid reserves as attack vectors — exceeds any marginal revenue benefit. Aave's cleanup is not a retreat. It is a balance sheet decision, applied to a balance sheet that finally has one.