According to Woofun AI, Babylon Labs submitted a paradigm-shifting ARFC proposal to Aave V4, which aims to completely break the custodial trust dilemma that has plagued the industry for a long time by allowing Bitcoin to remain in the owner's wallet during the mortgage lending process through the BaB zero-knowledge proof system.
For a long time, Bitcoin, as the largest collateral among crypto assets, faced structural contradictions where control must be handed over in the DeFi lending sector. Of the 2010 million bitcoins that have been mined, only 1.7% exist in the form of tokens on other blockchains, and almost every bitcoin behind it is held by a custodian or group of signers.
However, there is still strong demand for such loans from coin holders. Coinbase (COIN.US) connected its Bitcoin collateral lending business to a marketplace on the Morpho lending protocol. In the year ending September 30, the USDC borrowing scale in this market grew 74.9% to reach $1.5 billion. On October 14, Babylon Labs officially submitted an ARFC proposal requesting Aave V4 to support native bitcoins left in the owner's Bitcoin Output (Output) as collateral. The core of this mechanism is that only an honest challenger can block malicious withdrawals, and Bitcoin holders themselves can take on that role. Aave voters passed the temperature check vote with 299,743 AAVE votes in favor and 0 against, marking that this innovative mechanism is about to enter the Snapshot voting phase.
Judging from the comparison of trust assumptions, Babylon Treasury lowered the threshold of trust required for borrowers to a single honest side. In contrast, WBTC's custodian uses 2/3 multi-signatures to manage its reserves. tBTC requires 51 of 100 signature seats to transfer tokens, while LBTC relies on 10 of the 14 consortium members. However, Babylon's vault requires only a single honest challenger, and this role can be played entirely by the coin holder himself.
This simplification of the trust model has fundamentally eliminated reliance on centralized custodians or multi-signed alliances, and enabled asset control to be returned to users.
Technological innovation is the key to the implementation of this proposal. Under the BaB system, dispute resolution costs on the Bitcoin chain are only $37.82, compared to $14211 for BitVM2. The proof system developed by Babylon Labs will be presented at the ACM CCS 2026 conference, and it will also drastically reduce the initial setup data for each vault from 40.5 GiB to 22.2 MiB.
This breakthrough was jointly completed by researchers from Babylon Labs, the University of California at Berkeley, and Stanford University. The BaB system, built on Witness Encryption (Witness Encryption), not only kept the dispute cost at an ultra-low level of about $38, but also drastically reduced the amount of data by 1,868 times, solving the huge storage and cost bottlenecks of previous technical solutions.
Looking back at the historical background, the reason why Bitcoin previously had to rely on escrow is rooted in the limitations of its scripting language. Bitcoin scripts can verify signatures, hashes, and time locks, but it can't directly verify zero-knowledge proofs or read the state of other blockchains. Writing a proof validator in the Bitcoin script requires each proof to take up approximately 900 MB of block space, which is equivalent to the capacity of 225 full blocks. Because of this, Bitcoin in DeFi has always required some kind of entity to hold it on its behalf: a custodian, a group of signers, or an alliance that signs a release transaction after the loan has been repaid.
BitVM, which was released in October 2023, moved the verification process off-chain, making certain claims stand by default unless someone proves false. BitVM2, launched in August 2024, reduces the trust required for cross-chain bridges to a single honest party, even though a single dispute costs $14211 in processing fees and relies on non-standard transactions that the vast majority of Bitcoin nodes won't forward. BitVM3 uses a Garbled Circuit to reduce the dispute cost to $37.65, but each vault still needs to store 40.5 GiB of initial setup data. The advent of BaB has finally achieved a double breakthrough in cost and efficiency in this series of technological evolutions.
According to data compiled by Woofun AI, in terms of operating mechanism, Babylon Treasury has built this proof system around a single owner's token. Before a Bitcoin is locked, all participants pre-sign all possible flows of that Bitcoin: refunds, withdrawals, settlement of claims, and challenges. Subsequently, the Bitcoin network only accepted these pre-signed operation paths. The Bitcoin claim must include valid proof and must go through a challenge window lasting several days. During this window, treasury providers, app keepers (Keepers), general challengers, or Bitcoin holders themselves can take action to prevent malicious claims.
Since its launch on Ethereum on March 30, Aave V4 has divided lending into hubs (hubs) that hold liquidity and spokes (spokes) that set exclusive rules for various types of borrowers. Two new spokes were added to the ARFC proposal. Among them, Babylon Spoke, written by Aave Labs, is a derivative of Aave's standard Spoke, which uses native Bitcoin as its sole collateral. Aave has previously launched LBTC Spoke, and Babylon's proposal will be the first spoke to target unwrapped native Bitcoin.
Stani Kulechov called this exclusive Bitcoin hub 'a use case created for the V4 architecture'. The asset owner locked BTC in the Bitcoin network's Taproot output, then the adapter on Ethereum recorded it as an unmanaged, non-transferable, non-tradable, and unupgradeable credential token. The owner then borrows stablecoins from Hub and pays Aave's benchmark interest rate plus a Babylon interest fee (a rate not yet set in the ARFC proposal). The clearing mechanism gave Aave a second stream of interest income.
If a position's Health Factor (Health Factor, Aave's measure of the ratio of collateral to debt) falls below 1, the liquidator will pay off the debt and seize the entire treasury. Swap Spoke bought the vault at a discount below the Oracle price and paid it to the liquidator using WBTC borrowed from the Hub; the authorized arbitrator then repaid the WBTC and redeemed the Bitcoin from the Bitcoin network within a few days. Since borrowers only lent 2.10% of the 34,545 WBTC supplied on Aave V3 Ethereum on September 30 (compared to 14.6% in May 2024), this lending model makes full use of reserve pools with abundant surplus liquidity.
For institutional investors, this mechanism provides the dual value of balance sheet preservation and interest rate advantage. Institutional investors have long used similar methods to finance assets. In a tri-party repo (tri-party repo) transaction, collateral remains with a custodian trusted by both parties, and the lender holds an enforceable claim; the Babylon treasury maintains this structure, with the difference that a Bitcoin script acts as a custodian agent, and the custodian is the asset owner himself. Described by Fisher Yu as “Bitcoin that does not leave the owner's possession, balance sheet, and national borders,” HashKey Cloud plans to allow Asian institutional clients to borrow using native bitcoins as collateral through these vaults and invest the borrowed stablecoins into various income strategies.
The pricing of loans has also ushered in a revolution. The annual interest rate for loans announced by Bitcoin lender Ledn ranges from 11.49% for loans under $250,000 to 9.25% for larger loans of $2 million and above. In the 30 days up to September 30, the average annualized interest rate for USDC loans on Aave V3 was only 4.64%, and a flat rate was applied regardless of the size of the loan. Babylon expects the annualized cost of treasury borrowing to be between 3% and 5%, and plans to charge 0% treasury agreement fees in the early stages of the launch.
In terms of risk aversion, institutions' strong demand for decentralized self-escrow loans stems from painful historical losses fresh in the memory of coin holders. Previously, the lending platform BlockFi lent clients' crypto assets to institutional borrowers, while Genesis lent crypto assets from the Gemini (GEMI.US) Earn project. Both platforms declared bankruptcy before January 2023. When the custodian BitGo (BTGO.US) transferred WBTC custody to a joint venture with BitGlobal in August 2024, stablecoin issuer Sky immediately voted to reduce its WBTC loan limit to zero.
As of October 20, WBTC reserves held 116,512 BTC, down 25% from 154,707 on the day of the escrow change announcement. With Babylon Vault, asset allocators are no longer concerned with the risk of the custodian's weak balance sheet, but a transparent on-chain exit mechanism. After a challenge window of several days, Bitcoin can be returned back as it was.
Treasury contracts on Ethereum can be suspended immediately by two multi-signature mechanisms: one is Interim Governance (Interim Governance) for 4/7 seats, and the other is Emergency Guardian (Emergency Guardian) for 3/5 seats, which includes Aave Labs and LLAMarisk seats. Emergency Guardians can be independently unsuspended, while contract upgrades have a 5-day time lock. Aave's risk service provider will set borrowing limits and pledge rates (LTV) before on-chain voting.
In terms of the revenue distribution model and token economy, the ARFC proposal finely allocates eligible income: revenue sources include Hub Liquidity Fee (Liquidity Fee) withdrawn from interest paid by Babylon Spoke borrowers, and fees arising from liquidation. The remaining interest paid by the borrower is distributed directly to Hub's liquidity provider. Aave exclusively earns the first 500,000 in revenue each year.
The excess was divided according to the tiered range. Babylon's share ratio increased from 18% to 31.5%, and another 4% to 7% of the revenue was used to repurchase AAVE and BABY tokens on the open market (85% of which are AAVE). The repurchased tokens will be split equally between Aave DAO and the Babylon Foundation. For BABY token holders, these vaults open the first direct channel from borrower interest to token value capture.
When the qualifying annual revenue reaches $10 million, the share used for repurchases will buy $8.4 million of BABY each year, while Babylon's own agreement revenue share is more impressive; Babylon plans to launch a new plan in January to reduce inflation and guide token emissions towards treasury use. Since December 1, 2025, BABY has accumulated a decline of 37.1%, compared with a 9.2% drop in BTC and a 14.6% drop in AAVE. Currently, BABY corresponds to a market value of about $60.4 million, in stark contrast to the 40,843 BTC pledged under the Babylon agreement.
In terms of implementation path, revenue deduction, and future prospects, since the proposal was released on October 14, Stani Kulechov and other community members have given firm support to the ARFC, and Babylon has also responded quickly within 10 hours to answer core questions about the vault's suspension of authority. Since this year, Aave voters have passed 73 out of 74 non-type Snapshot proposals.
According to the Aave governance framework, the main network still needs to go through the following four standardization steps before implementation: risk and technology review, Llamarisk publishes a risk assessment report in the forum, and Aave Labs publishes a technical assessment report. This is a hard precondition for Snapshot voting on new assets in the framework adopted by Aave in August; Snapshot voting, which establishes a 1-day delay period and begins a 3-day community vote, which must meet the statutory requirement of 320,000 AAVes Threshold; the on-chain AIP proposal, which formally establishes parameters such as the loan limit and pledge rate (LTV), and then enters a 1-day delay period, 3-day on-chain voting, and a 1-day time lock execution period; re-audit and phased launch. Coinspect, Sherlock, Zellic, ABDK, ZK Security and Runtime Verification have previously conducted multiple rounds of audits of the treasury code. Sherlock has reported 8 high-risk and 55 medium-risk issues, all of which have already been Resolve or confirm knowledge completely.
Coinspect, ABDK, and Zellic will re-review the final code changes before the mainnet goes live. The main network will be launched in stages with a conservative initial quota limit; the previous December plan had set the target of launching the main network in April. Assuming that the treasury loan scale reaches Coinbase's volume as of September 30 (that is, borrowing $1.5 billion and interest rate is 4%, which falls within the 3% to 5% range expected by Babylon), if calculated at a 10% to 20% liquidity rate, it will generate $6 million to $12 million in eligible revenue per year.
In Aave V4, its Core Hub charges a 10% liquidity fee for USDC, 15% for Plus Hub, and 20% for USDG; the ARFC proposal leaves Babylon Spoke's decision on which hub to connect to subsequent AIP proposals. With an annual qualifying revenue of $6 million, Aave DAO can keep about $4.5 million a year, Babylon split about $1.3 million, and another $280,000 to buy back AAVE and BABY.
If annual revenue reaches $12 million, these figures will jump to $8.2 million, $3.2 million, and $700,000, respectively. Under the 50% collateral ratio (LTV) assumption, the loan size of $1.5 billion would require approximately 36,800 BTC collateral at the October 20 price, which is comparable to Babylon's current 40,843 BTC staking size. However, pledge interest and treasury collateral loans are two different application scenarios for tokens. Following the exploration of the BitVM series of technologies, this is a critical breakthrough in Bitcoin's native DeFi in trust minimization and cost efficiency, which is expected to reshape the institutional Bitcoin liquidity infrastructure pattern.