Unlocking Ajna

This week we took a look at Ajna, a promising primitive focused on permissionless, over-collateralized, oracle-free lending!

My biggest takeaway was that it’s a hard model to wrap your head around, so my mission in today’s newsletter is to do a better job of introducing the concept than Ajna’s ELI5 which still had me confused after two reads ??

Here goes…

Why Ajna?

No Oracles, No Problem

Ajna does not rely on oracles (price feeds) to value borrow tokens or collateral.

This is big because oracles are one of the major sources of risk in DeFi lending protocols, also consuming resources to run, maintain, or migrate.

No Governance, No Drama

In addition, Ajna does not rely on governance to evaluate and manage risk parameters, another significant source of risk in other lending platforms. Collateralization thresholds are an output of market participation, along with interest rates, which are a function of supply and demand. Permissionless market creation allows any user to launch any market.

The Ajna team even announced prior to launch that they would not continue to work on the protocol after deploying, and delegated the frontend development to third parties like MOM and Summer.fi.

How Ajna Works

Borrowers

For the most part, borrowers just deposit collateral, borrow, and may manage their positions similar to other lending protocols. They pay a variable interest rate depending on supply and demand. Borrowers may be liquidated if lender risk tolerance drops, a concept we’ll expand on shortly.

Lenders

This is where the complexity comes in, so let’s walk through it slowwwly…

Each pool is a market for a collateral-quote token pair. The quote token is the asset being lent and borrowed. The pool is broken up into collateral token “price buckets” (spaced 0.5% apart), which lenders choose when depositing the quote token.

For this example, let’s imagine the following ETH-DAI pool, with the price buckets on the y-axis and the amount deposited to each bucket on the x-axis.

Next, let’s imagine a bunch of borrowers come and borrow from this pool. The six yellow bars below show the aggregate amount borrowed, about half of the total pool deposits. The green and grey price buckets signify unutilized (not yet borrowed) quote token.

Notice, however, that there are two green price buckets earning interest, but still unutilized.

Let’s take a look at what it means as a lender to choose one of these price buckets.

Choosing a Price Bucket

As a lender, choosing a price bucket can be thought of as a tradeoff between having your deposited quote token converted to the collateral token at the price of that price bucket, and having a higher chance of earning interest.

The highest price buckets are most likely to always earn interest, but also are the first price buckets to be used during liquidations.

For example, if you were to lend 1000 DAI in the 2,000 price bucket, you are saying “I’m willing to have my DAI converted to ETH at a price of 2,000 DAI-ETH”, and you would end up with 0.5 ETH if users are liquidated, or if the price of ETH drops below 2,000 (any user may trade collateral tokens for quote tokens at the price of a price bucket).

LUP and HTP

Let’s introduce two new concepts.

The LUP is the Lowest Utilized Price. This is the price of the lowest yellow price bucket, found by aggregating all utilized (borrowed) quote token from top down. Borrowers who are not sufficiently collateralized to meet the LUP become eligible for liquidation, for example if the LUP was 1,500 and a user borrowed 4,000 against 2 ETH (4,000 / 2 = 2,000, which is greater than the LUP).

The HTP is the Highest Threshold Price. A Threshold Price (TP) is simply a loan’s debt divided by its collateral. The HTP, the highest TP, is a meaningful metric because it is a measure of the riskiest borrower. A TP of 1 would imply that the collateral now has the same value as the borrowed amount, and the user should be liquidated.

All lender deposits above the LUP or HTP earn interest, and can be thought of as being within a “fair” price bucket range as determined by borrower demand. Lenders in price buckets below the TP of the highest risk borrower do not earn interest.

Withdrawing or Moving Deposits Between Buckets

Lenders may withdraw unutilized deposits, but only when the withdrawal would not bump the LUP below the HTP. To withdraw in this case, the lender must “kick” (liquidate) the borrowers between the HTP and new LUP.

Liquidations

“Kickers”, or liquidators, must post a bond to start a liquidation auction, at which point a dutch auction is initiated (in a dutch auction, price starts high and drops, in Ajna’s case exponentially, until someone bids).

If the auction settles at a price higher than the neutral price (see the whitepaper, but for now just imagine “fair market price”), the kicker loses some or all of their bond. If the auction settles at a price below the neutral price, the kicker earns a liquidation reward.

Remember, there are no oracles, so this is necessary complexity!

Interest

Borrowers pay interest-earning lenders a variable interest rate determined by supply and demand.

More

There is quite a bit more complexity in the protocol design, but the goal of this was to attempt an ELI5, so check out the whitepaper for more.

Traction & Token

Ajna benefited from substantial hype when developing and upon launch. Shortly thereafter, a vulnerability was identified and users were advised to pay back loans and withdraw funds. Ajna went back to audits and then relaunched.

Notable today is the drop in both TVL and token price, which now sits below a $20m market cap. The TVL can be seen below (the second mountain being the re-launched v2).

Borrowed amount seems to be tracking the drop in TVL, so perhaps this is just a result of early incentive farmers drying up.

Notable as well is that Ajna launched its permissionless protocol with the intention of not relying on continued support from the initial dev team. This I’m sure also has an effect on future valuation, and I would have liked to see a steward appointed.

Regardless, I think it’s a promising primitive and hope that developer activity around it increases! Whether it takes off or not soon, having a permissionless, oracle-less lending option is a valuable asset for the industry.

*And credit to whoever made the Ajna price bucket graphic from the whitepaper that I copied.

And a big thanks to our sponsor Size, building a DeFi lending marketplace with unified liquidity across maturities. Coming soon!

This newsletter is for informational purposes only and does not constitute investment, financial, or professional advice. Perform your own research and consult a professional before investing. We're not liable for any direct or indirect financial losses resulting from the use of information contained in this newsletter. Past performance is not indicative of future results, and investing in financial markets carries risks, including the potential loss of principal.