Aave: Flash loans!

Let's take a look at flash loans on Aave and how they are used.

Top Tokens Borrowed

note that the discrepancy in USD volume of between the bottom 2 graphs and the above 2 graphs can be attributed to the fact that we are only looking at tokens with >1M USD in volume.

ETH has consistently been the highest token borrowed for flashloans with the stablecoins DAI and USDC with USDT trailing very far behind. This, in part, could be due to the much greater supply of USDC, WETH, and DAI in the liquidity pools. This graph only looks at the coins with >1M in volume over that given week. What's interesting is that LINK has also seen a fair share of weeks over the 1M threshold. LINK had a similar volume to BTC which is surprising to me because BTC is much more widely known and owned/supplied in general. LINK joins elite company in

Flash Loans and Use Cases

Flash loans are instant loans that must be repaid within the same transaction. There are two cases

Arbitrage

In the case that there are price discrepancies between pools on different protocols, you can execute instant trades and make a pretty penny.

Self Liquidations

Imagine this: using Compound, you were able to get a DAI loan using ETH as your collateral. Unfortunately, ETH's price is dropping and you will soon be liquidated and you don't want to collateralize more ETH. Utilizing Aave's flash loans, you can take a flash loan to pay off the DAI you owe, withdraw your ETH and then swap ETH to DAI to pay off the flash loan + premium while keeping the rest of your ETH. You're saved!

Note on the data

aave.flashloans was first queried grabbing flash loan events (and their tx_id) which were then referenced and used in ethereum.events_emitted to find the protocols used. I focused on finding these 2 use cases here - with Swap, TokenExchange, and Repay being the main events I am looking for.

Entries were grouped weekly, and top borrowed tokens are defined as having >1M in volume. protocols with no labels were grouped into an unknown category. All currency values are in USD.

Protocol Interaction

The data here kind of explains what we were expecting. With the majority of interactions being with the UNISWAP, SUSHISWAP, and COMPOUND protocols. Notice that the week of May 16th when the crypto crash happened, volume to compound surpassed sushi's volume while the majority of the past 18 weeks, volume to compound mostly sits under 10M.

Personally, I didn't know about the other smaller platforms shown here such as balancer and DMM. Within the last few weeks, there has even been the emergence of SakeSwap, an extension of SakePerp.