Stablecoins and payments at Aave

Context

Aave is the leading crypto lending protocol. I was a Senior Product Manager leading Newt: an internal experiment lab building on top of the Aave protocol. We built experiments for RWA pools, undercollateralised lending, identity primitives, analytics tooling, etc. GHO was the most significant experiment I incubated, where I iterated with engineering on architecture, as well as led an early proof-of-concept exploring GHO as a rollup-native stablecoin.

My ownership spanned from experiment selection to rapid prototyping, assessment, and (if assessment passed) handoff to Aave Core engineering for full build.

GHO Stablecoin

Problem

As a lending protocol, Aave was one of the largest venues for stablecoin borrowing, but the interest paid by borrowers flowed to USDC and DAI suppliers, not to Aave (outside of a small fee cut). Every dollar of stable borrowing demand Aave generated was, in effect, revenue for Circle and MakerDAO.

Solution

GHO was designed to move Aave more "down the stack." At its core, GHO is an over-collateralised stablecoin (backed by assets worth more than what's minted) issued directly by Aave. stkAAVE holders - users who stake AAVE tokens in Aave's Safety Module - borrow at a discount, creating a flywheel between AAVE staking, governance participation, and stablecoin demand.

The core architecture was the Facilitator model: governance-approved contracts, each with a bucket cap limiting how much GHO a given venue can mint. Rather than a monolithic minting contract, GHO could be issued by any venue the DAO (decentralized autonomous organization - Aave's on-chain governance, where AAVE token holders vote on protocol decisions) approved and sized (e.g., the Aave V3 Pool, a Flash Minter, a Stability Module, cross-chain Facilitators).

GTM

For a DAO-governed stablecoin, go-to-market is a governance vote. The bucket cap design meant the DAO could size exposure deliberately - the work was writing an AIP (Aave Improvement Proposal) that gave governance enough signal to set caps responsibly, and sequencing Facilitators so the primary venue proved the model before more complex ones were introduced.

The governance proposal passed with 99% approval in July 2023. GHO reached $564M in circulation within its first year at 245% average collateralisation, generating over $2.3M for the DAO treasury in the process.

Stablechain PoC

Being a stablecoin, we also wanted to use GHO for payments, for Aave's enterprise partners as well as the Aave app. We also wanted to experiment with GHO as native gas on emerging chains. A dedicated chain for GHO (i.e., a "stablechain") would make all of this possible.

If a chain settles to Ethereum (i.e., it posts its transaction proofs to Ethereum mainnet) and if Aave also holds collateral on that same chain, then GHO could be minted and used natively on this chain, as well as backed by onchain collateral with no bridging risk.

To build a dedicated chain for GHO, I needed to evaluate several "rollup kits" (what are now called Polygon CDK, OP Superchain, Arbitrum Orbit, and ZK Stack before those names existed). In theory, the dev tooling was there, but the question was whether they were mature enough to ship anything production-ready. Sadly, they weren't, and the PoC remained early-stage and unpublished research.

During this experiment, I realised that even if the rollup stacks had been ready, GHO would still have needed a reliable way to move between chains. The same rollup kit making it easy to launch a stablecoin chain was going to make it easy to launch a payments chain, an NFT minting chain, a gaming chain, etc... each one would be optimised for its own use case and in their own isolated island.

But the rollup kit thesis required a cross-chain primitive that didn't exist yet: something that could connect new chains automatically. This insight was the beginning of Catalyst.

Reflections

The rollup stacks weren't production ready in 2022. What we were building required mature sequencer economics and a governance framework for Facilitator management across chains - none of which existed.

The stablechain thesis proved out two years later with the launches of Stable, Tempo, Circle's Arc, and Plasma now exist. Goes to show that a strategic insight is only as good as its timing.