Compound Finance — Gateway

Or whatever it’s called

No argument on the substance here, but…

Thesis

Compound’s money markets are a core primitive of the decentralized finance ecosystem that’s sprung up on Ethereum over the last year. Gateway will provide for Compound’s expansion of these markets to other Layer 1 chains as well as Layer 2 sidechains. It will dramatically increase the total addressable market of the protocol. Beyond this growth, the change in network dynamics and introduction of validator incentives will introduce a new class of COMP buyers that add incremental value to the token.

Given the lack of chatter about Gateway (search for it on Twitter), and the COMP token’s relative underperformance, I am taking advantage of a material downturn in token prices to start buying COMP in Pineapple Street Allocators.

Background

Compound is a money market protocol built on Ethereum. It enables users to borrow and lend crypto tokens. Borrowers pay interest to lenders. The rate of interest varies per token pair based on supply and demand dynamics and the volatility of the token being used as collateral. Borrowers must also post collateral, such that every loan is overcollateralized. If the collateral drops in value below a certain threshold, it’s eligible for liquidation. Performing a liquidation is a profitable endeavor, so there has been a robust ecosystem of service providers (liquidation bots) built to handle these cases.

Gateway will perform largely the same actions, but will allow for cross-protocol borrowing and lending. In the Ethereum-only version of Compound that’s currently available, you can lend and borrow 10 different tokens, all built on Ethereum. Gateway will open up the possibility to lend and borrow any number of tokens across any number of peer chains. So, for example, you could post collateral in Serum (or any arbitrary Solana asset) to borrow Wrapped Ether. Importantly, the current iteration of Compound pays and charges interest in kind, meaning that if you borrow REP, your interest payable accrues in REP. Gateway will transition to a native unit of account called CASH. Lenders will earn CASH and borrowers will pay it. CASH will initially be pegged to $1, but this setting is arbitrary and can potentially be changed to an index or basket in the future.

This normalization of the protocol’s unit of account simplifies end-user risk and return calculations, and the cross-chain functionality greatly expands the opportunity set with which a user can use the Compound protocol to interact. To analyze the potential impact of these changes to the value of the COMP token, it is important to understand the mechanics of how Gateway bridges the gap between protocols. High-level overview incoming:

Gateway is a standalone blockchain built for the sole purpose of providing money markets for an arbitrary number of crypto assets on an arbitrary number of Layer 1 and Layer 2 blockchains. As with all blockchains, it is important that there is a canonical record of the chain’s transaction state; which transactions were executed in what order and by whom? Gateway uses a method called Delegated Proof of Authority to manage this record keeping.

In this system, a handful of Validator Nodes monitor the network for transactions. A Validator Node is a specialized piece of software run by a dedicated team that is paid to do the following:

• When a node sees a transaction that they believe to be valid, they’ll broadcast that transaction, along with a condensed version of all transactions ever, to all the other nodes.
• Those nodes will confirm the transaction as valid (or reject it if not) and quickly go back to searching for the next valid transaction.
• In the valid case, the network’s transaction history will be updated to include the new transaction.

Validator Nodes also listen outside of the Gateway network to manage the uploading and downloading of peer chain assets to and from Gateway. The peer chains will all have specialized smart contracts called Starports whose sole purpose is to handle moving that peer chain's native assets to and from Gateway. So for example, if on Ethereum I upload 10 COMP to the Ethereum Starport, a validator node will see the event emitted by that Starport and update my Gateway account with that collateral balance. If I then use that collateral to borrow 10 SOL which I want to move back to Solana, I submit an extraction request to the Compound Chain, where it is queued and processed by off-chain workers. Once the request has been processed, it will be signed by a validator to form an extraction notice, which can be taken back to the Solana Starport and redeemed for 10 SOL.

To recap — specialized contracts on the peer chains control the deposit and redemption of chain-specific assets to Gateway. These contracts emit events that are monitored by the Gateway validators. The interplay of the Gateway validators and the peer chain starports maintains Gateway peer chain asset-specific account balances for individual users. CASH balances, credited to lenders and debited from borrowers, are maintained by the Validators.

What does this mean for COMP? To answer this question, it’s helpful to define COMP’s current ownership and use case, and to think about how that use case may change with the advent of Gateway.

COMP tokens allow their holders to vote in actions to change the parameters of the protocol. 53ish percent of the COMP tokens that will ever exist already do, and they were allocated to early supporters of the protocol (users, team, investors). However, the protocol mints continually mints new tokens for distribution to users (borrowers and lenders). This distribution will continue for another three years in an effort to decentralize the governance to interested parties.

The parameters that can be changed by governance include but are likely not limited to adding support for new assets to the protocol, spending money from the protocol’s treasury, changing the parameters of the risk engine, etc. The full list of historical governance proposals and their status can be found here. These changes can have far-reaching financial implications for users of the protocol.

With the release of Gateway, Compound as a protocol will be accessible from many chains. However, governance will remain on Ethereum, and the current voting mechanisms will remain unchanged. Holders of the COMP ERC20 token will be responsible for governing the protocol.

There are at least two factors introduced by the release of Gateway that will increase the demand for COMP tokens. The first is just the expansion of the total addressable market. Moving from 10 supported ERC20 tokens to an arbitrary number of tokens on an arbitrary number of chains increases COMP’s exposure to the growth of the crypto ecosystem as a whole. As more activity occurs on Gateway, it will be more valuable for users to have a say in how Gateway runs. This is a directional, instinctual valuation. It’s difficult to put a value on, because the governance changes that increase economic value for users do so over a long time horizon for a disparate group of users.

Anecdotally, however, it proves out. The proposal to add Tether (USDT) support was the first to pass Compound governance in May 2020. It passed overwhelmingly, though there were some notable dissenters.

At the time, USDT was an $8B asset, split between Tether issued on Omni and Tether issued on Ethereum. USDT is currently a $60b asset, with the majority issued on Tron. Adding USDT as an asset was a mildly contentious first test of Compound governance when it had a market cap of $6b. The inevitable vote to add Tron USDT as a supported asset on Gateway will have at least 10x the (slightly diffuse) economic incentives at play.

The second factor that will increase the demand for COMP is the direct economic incentives provided by governance over Gateway’s validators. These nodes will be profitable enterprises, run by professional firms with large sunk costs and a keen motivation to keep both their position as a validator and the specifics of that position as beneficial to them as possible. Governance will be responsible for deciding the total number of validators, who those validators are, the transactions which are eligible to be charged, and the amount charged per transaction. In sum, the totality of a validator’s business model can be tweaked by COMP governance votes. Where the short-term economic effects of a current COMP governance vote are hard to forecast, and the long-term effects can be predicted directionally at best, the short- and long-term effects of a vote that changes the validator settings on the Gateway network will be obvious and instant.

While the total revenue generated lending and borrowing activity will scale only linearly with the expanded addressable market, the incentives to participate in governance will certainly grow faster.

Risks

There’s the obvious existential risks for crypto and defi.

• Smart contracts are hard; code can be audited a million times and still be exploitable
• The eye of Sauron is honing in; regulators have a keen eye on what’s happening in the ecosystem and there will likely be some effort to reign in risk to retail

There are also several Gateway-specific risks:

1). Execution. Is this team going to be able to launch a completely new, purpose-built blockchain? As a peon, this is a difficult but not impossible thing to underwrite. I obviously don’t have a direct line to the team, but they are highly responsive to thoughtful questions in the community Discord. There are daily progress updates, and many interactions between the team and those teams that are running validator nodes on testnet. I’ve asked questions and had them answered by community members, engineers, and the Compound Labs CEO.

You can also look at history; Compound launched the yield farming craze in June 2019 after successfully decentralizing their governance structure. **(note: it’s been pointed out to me that Synthetix invented yield farming, which is true and honestly I’d forgotten. I stand by the idea that Compound was the match that lit the fire last summer.)** This was not a straightforward process, as there were many institutional shareholders involved, as well as a larger crypto community who had never seen a process like this before. Even with that being the case, they managed those disparate stakeholders smoothly and got the product out the door to launch DeFi summer 2020.

2). Competition. Aave and Maker, the two major competitors on Ethereum, are pursuing slightly different avenues of expansion. Aave is growing tremendously on L2, with their Polygon integration attracting significant attention and TVL. However, the experience of lending there actually hammered home the importance of cross-chain interoperability for me; it’s currently not possible to lend Polygon assets and borrow Ethereum assets. Gateway will launch with this functionality out of the box (a Polygon Starport is under development).

Maker does have a plan for cross-chain interoperability, and a track record of delivering. DAI is without a doubt the decentralized stable coin with the most adoption. However, as detailed in that post, blockchain bridges to move DAI around between chains will be bilateral. This extra level of complexity impedes both UI/UX and the management of risk parameters; if those parameters are enforced at the bridge level, how do you implement governance changes across N bridges? With gateway, those changes will be executable at the chain level and applicable to all Starports. It just seems like a much cleaner solution.

It’s also worth considering the fact that the vast majority of lending and borrowing in DeFi is done through protocols that require borrowers to be over collateralized. It’s reasonable to think, given the pace of progress in the last 18 months, that someone is going to crack uncollateralized lending. In fact, just yesterday Goldfinch raised $11m from A16z to work on this problem. It’s a tricky one though, and if they do it, it’s not going to be instantaneous. I’ll be keeping an eye on Goldfinch, and TrueFi, and others, but my instinct is that until the use of onchain identity is pervasive, these undercollateralized protocols will be cordoned off for large, known borrowers. And the size (and fees) comes with scale.

Metrics

There have been many attempts to value protocols based on their usage in terms of the value of tokens that are locked in as collateral. This a useful metric, but I don’t really have an instinct as to what the correct number is (because I don’t think that exists). Rather, I plan to watch a set of metrics as the life of my investment progresses and use changes in the trends (combined with market sentiment and good old meat space logic-ing) to re-evaluate my investment thesis on a regular cadence. Reserving the right to add and delete metrics, I’m currently watching:

• 7-day total revenue (interest paid from borrowers to lenders)
• 30-day total revenue
• Annualized revenue
• Total value locked
• Annualized borrowing volume
• Price/Sales
• Total number of users