Library Credit and stable value2020Design paperCorpus record

Aave Protocol white paper

Aave. Aave.

Aave's v1 protocol paper: pooled lending with stable and variable rates, and loan features such as rate switching and uncollateralised flash loans inside a single transaction. It sits in the same family as Compound, with a different rate and product surface.

Aave's paper extends pooled lending with aTokens that rebase the balance itself, and with flash loans that must be repaid inside one transaction.

The five-minute read

The pool is still the product

Depositors share a market. Borrowers post collateral. Rates move with utilisation. That much is the Compound family. Aave's paper then adds features that changed how the pool is used.

aTokens rebase

Instead of an exchange rate on a receipt, the balance in the wallet increases as interest arrives. Integrations have to expect a balance that moves without a transfer.

Flash loans are uncollateralised and atomic

A contract may borrow the whole available pool with no collateral if it repays, plus the fee, before the transaction ends. If it cannot, nothing happened. The paper is clear that this is not a term loan.

Stable and variable rates are both specified

A borrower may choose a rate the protocol tries to hold, with a rebalance clause when it drifts. The clause is the part posters forget.

The paper is not a risk-off switch

Later safety modules, isolation modes and portals are subsequent. The 2020 text should be cited for the pool, the aToken and the flash loan.

One action, walked through

A depositor supplies an asset and receives aTokens. The aToken balance increases as the liquidity index moves.
A borrower posts collateral and draws either a variable-rate or a stable-rate debt, subject to health factors.
A flash-loan caller receives the underlying in a callback, does arbitrary work, and must leave the pool repaid plus the fee.
If a borrower's health factor falls through one, a liquidator repays part of the debt and takes collateral at a bonus.
If a stable rate diverges too far from the market, the paper's rebalance lets the protocol move it. The borrower does not have a permanent fixed coupon.

The argument, unpacked

Flash loans did not invent atomicity

Ethereum transactions already revert as a unit. The paper makes the pool's idle liquidity available inside that unit. The risk is not that the pool loses the principal when the callback fails. The risk is what other protocols do when someone can suddenly be very rich for one call. Aave's own solvency and the safety of neighbours are different questions.

A rebasing balance surprises accounting

Systems that snapshot balances, or that escrow a fixed amount, mis-count aTokens. The paper's choice is friendlier in a wallet and harder in a contract. Integrators have to read the index.

Stable is a name with an exit

The rebalance rule means 'stable' is stable until the protocol decides the subsidy is too large. A borrower who needed a true fixed rate has not got one. The paper should be quoted on the rebalance, not only on the choice of rate mode.

What has to be true

The liquidity index updates, and every integration uses it. A raw balanceOf snapshot is wrong.
Flash-loan fees and repayment are checked after the callback. Any token that does not arrive in the amount sent breaks the check.
Health factors use a price feed that can be defended. Flash liquidity makes price manipulation cheaper to attempt.
Liquidators exist. A flash loan does not liquidate an underwater borrower by itself.

What happened after the paper

Aave v2 and v3 added collateral toggles, isolation, portals and a safety module. Flash loans became standard tooling for arbitrage and for attacks on other protocols. The paper's own pool accounting survived. The lesson it did not finish writing is that atomic liquidity is a public good and a public hazard.

What to check before you use the idea

Does this integration assume a constant token balance? aTokens do not have one.
Can a flash loan move a price this other contract trusts, inside one transaction?
What is the stable-rate rebalance rule?
Which version is live, and which risk module is not in the 2020 paper?

Terms

aToken: An interest-bearing balance that increases as the pool's liquidity index increases.
Flash loan: A loan with no collateral that must be repaid before the transaction ends.
Health factor: A ratio of discounted collateral to debt. Below the threshold, liquidation is allowed.
Rate rebalance: The protocol's right to move a so-called stable borrow rate when it has diverged from the market.

The problem the paper names

A single floating rate does not match every borrower, and some uses of inventory only need the asset for the length of one transaction. Aave's paper adds a choice of rate and a loan that never leaves the transaction that borrowed it.

What the design proposes

Suppliers deposit into a reserve and receive a claim.
Borrowers choose a rate mode where the protocol offers one. Collateral rules still apply to open loans.
A flash loan must be repaid, with its fee, before the transaction ends, or the transaction reverts.

How the mechanism is specified

Reserve factors and loan-to-value settings are risk parameters. They are not derived inside the paper from first principles.
Flash liquidity is real only while the pool holds the asset. It is not a credit opinion about the borrower, because there is no time in which they hold the asset unsecured.
Later Aave versions add risk admins, portals and different token mechanics. Cite the version you mean.

What this page does not treat as proven

A flash loan is not a banking licence and not a statement about borrower creditworthiness.
Stable-rate modes can still be repriced under the protocol's own rules. 'Stable' is a defined term, not English.
This page does not describe liquidity mining or safety-module staking.

Why a venture studio still reads it

Use this paper when a product pitch says 'flash loan' without the revert rule, or 'stable rate' without the clause that lets the rate move. Those clauses are the product.

This is Blockchain Lab's reading of a public design paper. It is not the paper, not a copy of it, and not an offer of tokens, equity, custody or a partnership. Later network behaviour can diverge from the text. Nothing here is investment, legal or technical advice.

Research status: Design paper. Last reviewed: 1 October 2026. This is a reading of a public paper, not investment, legal or security advice.