Gearbox V3 Teaser: ReDEFIning Leverage and Lending.

Mugglesect
Gearbox Protocol ⚙️🧰
15 min readMar 13, 2023

Gearbox Protocol. Too often we have been asked “Why that name?”, “Why Gearbox?”. Let’s step back and ponder the question lesser asked, “Why Protocol?”. Think of any protocol. TCP/IP, HTTP or any that pop up in your mind. What’s the common thread? They all work as a common layer between the participants in a network. While the applications and hardware might differ between participants, the protocols keep functioning the same throughout. And that’s why, Gearbox Protocol.

The goal for Gearbox has never been to be another leverage dApp on chain, the goal for Gearbox Protocol has always been to become the base layer of leverage for DeFi. While the dApps, the chains, the tokens might change, Gearbox’s success lies in having the ability to natively deliver leverage across DeFi. In the most secure and safe way possible. — © not Satoshi Nakamoto

Redefine Lending, Redefine Leverage!

While that might not have been entirely possible until today, it’s continuous improvements that’ll get us there. And today, we bring you a big one. Today, we bring you how we reDEFIne lending, how we reDEFIne leverage, how we reDEFIne ourselves… All with Gearbox Protocol 3.0

Gearbox V1 was our Proof of Concept and Safety, largely functioning in Test in Prod mode. Gearbox V2 was Proof of Ability to Scale with Safety, focusing on finding initial traces of PMF, or Product-Market Fit, and gauging how to improve safety frameworks. V2 delivered on both of those. Running smoothly on all parameters with a $100M TVL and a Safety Framework that has constantly ensured $0 bad debt.

The protocol recently took a TVL drop due to USDC drama over the March 10–13 weekend. Everything has stayed safe throughout!

Excited to hear that and want to know what all Gearbox V3 is capable of achieving? Well read on, anon. The article below will give a high level view of all things Gearbox V3 and what to expect. The detailed information on the topics will be released as and when the DAO votes happen and we approach launch. You can read the analysis that led to V3 design decisions in our V2 introspection article below.

1. Redefine Lending: Welcome Alpha pools!

V3 effectively builds upon V2 by creating capabilities to not just add pools, tokens and protocols even with lower liquidity but enabling Gearbox to add L2s to its sources of growth as well. And it does so while seamlessly making the $GEAR token a key part of the protocol functioning itself.

NOTE: V3 doesn’t just build upon V2’s performance, it literally builds on V2’s codebase itself. Retaining most of the security improvements and the battle-tested nature of the code. This enables Gearbox to scale faster. No migration would be required for passive lenders, making it super hussle-free for users ❤

We have always said that Gearbox Protocol is modular. Well, it’s time for us to show you the true power of that.

Passive lending side will get a new piece: lenders would have a choice to remain in the current non-changeable lending pool (Main Bluechip) — or opt in to earn more APY by taking a bit more risk in the Alpha pool. Since Gearbox DAO has very high security standards, the extra risks might be seen as negligible by many DeFi participants. Well, if so, easy opt-in then!

The main Bluechip pool will remain with conservative integrations and conservative asset LTs (LTVs), representing the most safe avenue and value proposition especially for large capital. Lido, Curve, Convex, Yearn, Uniswap, and so on. Somewhat boring, comfy, trusted, secure… This is how we want everyone to see it. No funny business here!

And then there will be a new Alpha passive lending pool. That will be an extra lending pool where the main pool dTokens can be deposited into, as an opt-in essentially. Re-deposit. Here, you would opt in for a bit more risk yet can make higher APYs: the choice is up to the passive lenders — up to you! So what extra risk are we talking about here and why would you do it?

  • There would be newer integrations here. Still with larger and safer DeFi / NFT / etc. protocols, but the composition will not be as conservative comparatively speaking. See it as just more exotic lending. This pool’s integrations will still have conservative enough asset LTVs as well as strict liquidation rules. Security is everything!
  • For this extra hassle, you’d be expected to make more APY as a passive lender. That happens because the integrations here would allow Ninjas (leverage borrowers) to generate more APY via more avenues and fancier strategies. As a result, they would be prepared to pay higher borrow rates. Not just 3–5%, but likely 7% or even 10+%!

The extra APY in the Alpha pool would be coming in an unusual way: it would be at minimum the base APY of the Bluechip pool + base APY of the Alpha pool + the extra fees taken from the quotas (see section 2.4 below). As such, Alpha pool will earn the minimum of the Main pool + extra.

In the future, there can be Pool N or other pools like the Alpha pool… modularity doesn’t have to stop. This modularity overall helps granularize risks while allowing for growth. Such a pool setup doesn’t fractionalize liquidity which is also quite important for lending-like protocols.

But, modularity alone doesn’t solve all issues. With that said, meet:

2. Redefine Scale: Asset Limits and Quotas

Gearbox Protocol will now introduce overall exposure limits for assets in the Allowed List. The way this works is that for each Allowed List asset there is a threshold (called total limit) denominated in the underlying asset of the corresponding lending pool. The sum value of exposure in this asset (which is the borrowed debt swapped to the asset and the user collateral or quotas — as we will call it) across all Credit Accounts borrowing from that pool cannot exceed this total limit. Thus, limiting risk.

Time to move the Gear to Alpha!

The actual amount of asset used as collateral on a particular CA is determined by a quota, which is also denominated in the underlying asset. Each borrower can set any Quota value they want for each of their mid-tail assets. Which means the value of Collateral they intend on putting up and the overall Borrow value they intend on taking(although the larger the quota, the larger the interest that they will pay — see below), as long as the total sum of all quotas is within the total limit.

It’s very simple to understand: just a maximum number per asset that the protocol is potentially exposed to.

To give an example, suppose that we have a USDC Credit Manager connected to the USDC pool, and a total limit on MIM3CRV in that pool is 10M USDC. Each user that deposits into MIM3CRV would then set a quota on the maximal value of MIM3CRV they will have exposure to on their account. So if you’ve set 1M USDC as a quota for your MIM3CRV, then at most 1M USDC worth of your MIM3CRV (according to the oracle) will be counted as exposure. If your MIM3CRV is worth more than 1M USDC, then anything above 1M will not be counted. If you have less than 1M USDC worth, then the actual value will be used — i.e., 500k USDC worth of MIM3CRV with a 1M quota is still 500k. The sum of all such quotas set by users cannot exceed 10M USDC.

As a result, the exposure of the pool to an asset is effectively capped — since borrowers can only have up to Quota’s limit worth of asset as exposure, they can only borrow total limit * asset LT against it.

This mechanic solves many problems:

  1. Addition of Long Tail and Medium Tail assets: It creates an upper bound on the amount of an asset that needs to be liquidated at any given time, which allows the addition of long-tail and medium-tail (i.e., lower liquidity) assets without increasing protocol risks. Currently, due to how Gearbox functions internally (and that is the issue of Gearbox copies) adding any riskier asset almost fully exposes the protocol to it. Nothing stops borrowers from collateralizing $10M of an asset that only has $5M worth of liquidity on the market — this leads to huge price impacts during liquidations, which leads to liquidations not being profitable, which leads to possible bad debt. Having a limit on the value of the asset being collateralized naturally fixes that.
  2. Deploying on L2s: As a corollary, cross-chain deployments also become possible. L2s and side chains suffer from lower liquidity, and as the multichain grows (or not) this might be a growing problem. As a result, adding even Curve farms on L2s is quite risky as they are at max $10M per pool. Hard to add such assets or anything else when it can be so cheaply-manipulated. Of course, there are other ways to prevent manipulations, but making integrations for such tiny sizes isn’t really fun is it?
  3. New cool protocols and asset classes (which usually have low liquidity and high risk, but also large yields) can also be added, capped at $1–2M limits. Negligible size for the current lending pools, yet can serve as a great way to keep leadership and give a boost to the newer DeFi entrants. Yay!

The total limits would of course be finalised basis risk committee recommendations and governance.

And to top this all off — let’s make some $$?

3. Redefine Revenue: Quota interest

The new design will also introduce a new kind of interest that is derived from quotas. This extra interest paid by borrowers is essentially a %APR on their quota size for each asset, instead of their borrowed amount (i.e., if the asset’s quota interest is 5% and a borrower has a quota of 1M USDC in this asset, they will pay 50k USDC per year of this extra interest, for that asset). The quota interest is paid to the Alpha pool, which will receive base Bluechip + Alpha APYs + all quota interest collected. The DAO charges its usual interest fee on top of that.

The APR on quotas is configured through a GEAR voting system (see 2.6 below) and consists of two parts:

  • The minimal risk premium paid to the LP for the risk of being exposed to a particular asset. The APR cannot go below this value.
  • Everything above that (i.e., actual APR — minimal premium) can be considered a profit share between Ninjas and the LPs.

See it the following way: there are conservative stablecoin farms yielding 3% APY, some medium-grade farms in the 5–10% range, and then some new low-liquidity degen slot machines that can yield 20% even before leverage. All of these assets will be leveraged with funds taken from the same borrowing pool.

It’s not possible to reflect this diverse set of asset classes and rates within a simple interest model based on a utilization curve:

  1. You can set your utilization curve low to accommodate all assets, but then Ninjas would just be able to print money on high risk / high reward degen farms, without properly compensating the LPs for the extra risk they take;
  2. You can set your utilization curve higher, but that would just exclude low risk / low yield asset classes from that particular pool. You would have to set up a multitude of pools graded by risk, which is just unwieldy and breaks composability in a major way.

To compensate LPs fairly while preserving the flexibility of Gearbox, you have to price discriminate based on the borrower’s portfolio somehow — and the quota interest is an elegant extension of the quota mechanism that allows to do that.

Splitting the pie would not be as hard as it might sound at first. Just see it as an extra buffer on positions that are making too much compared to the nominal utilization rate. The extra interest does not purely take away from Ninjas either — higher APYs will bring more LPs, so the base rates will not spike as much due to large movements in the pool. On the whole, Ninja strategies should become much more predictable over time due to that.

Configuring this APR is a careful balancing act — governance and LPs need to be not too greedy as to not eat into Ninja profits too much and turn them away. But also not too lenient, in order to still make it an attractive endeavour for the passive lenders to engage in. Balance! As APYs change and Ninjas make less or more, weekly fees can reflect that.

But how exactly would GEAR holders configure this? *Cue Final Boss music*

4. Redefine Tokenomics: MVT via Staking

The parameters above, like any changes or improvements in Gearbox Protocol, can only be altered after a GEAR DAO voting process. The same goes for quota interest APRs. However, this sort of risk premium / revenue share tweaking needs to be constant, to reflect changing APYs from Ninja strategies and changing risk associated with certain assets. As such, Snapshot governance is not really enough for this, and a more continuous on-chain system is needed. It would work as follows:

  • APY Range setting: For each asset, there will be two parameters: the minimal risk premium mentioned above, and the max APR. The MRP is set by the risk committee, while the max APR is estimated by the economics committee. Both are periodically revised.
  • GEAR holders decide split: Once the parameters are set, GEAR holders can vote into one of two buckets — the LP bucket or the Ninja bucket. The proportion of GEAR in each bucket determines where the actual APR value will be in the [MRP; max APR] range. More GEAR in the LP bucket moves the needle closer to max APR, while the Ninja bucket moves it closer to MRP. All of this is per collateral asset.
  • Eventual Predictability: Voting is done in weekly epochs, and the rates are only recalculated based on votes at the beginning of the new epoch. This should make quota APRs more predictable for Ninjas and make the system more resistant to manipulation. On epoch start, the voter’s existing votes are automatically rolled into the same buckets and assets where they were, unless the voter explicitly changes their votes.
  • Avoiding Exploits: GEAR tokens that are voted with are staked, and there will be a small unstaking period (probably 1 to 4 weeks, TBD by DAO). This is a security measure that prevents flash-voting to skew capital into certain tokens and pools. GEAR will be staked for each asset separately, so voters will need to think on how to allocate their GEAR across all assets.

This is minimum viable because it is the simplest model that creates organic use cases for GEAR. But, there is no profit sharing to GEAR holders yet. The system is new. Profit sharing for either stakers or a safety module or any other primitive — can be done at later stages. What is important is that there is an extra revenue source now! For example, the DAO could:

  • Transition governance to being fully on-chain in the next few months
  • Make a staking module where you can vote for these quotas
  • Stakers would also share some % of the extra revenue
  • OR stakers could also be required to stake into a BPT 20/80 Safety module (stkAAVE style) and then share
  • OR some other idea, whatever design made that is approved

So effectively, the devs create the capability of token utility via staking while the DAO decides how they want to actually execute the value accrual.

So a lot of new possible combinations, but with more volatile assets, you require a much more sophisticated UX to operate efficiently…

5. Redefine UX: Automation on Chain

While we have talked about scale, more volatile exposure, governance around that and more. How do you safely execute a position with these assets? How do you create the precision of entry, exits and more? We bring to you Gearbots.

Gearbots are open source, immutable automation contracts that essentially allow CA owners to delegate certain aspects of active account management to a third party that functions in a predictable and neutral way. Bringing to you multiple use cases like:

a. Automated portfolio management features

Gearbots (coupled with the inherent flexibility of Credit Accounts) enable many aspects of CEX user experience on-chain, while still retaining trustlessness, self-custody and composability of Gearbox. The features include:

  1. Submit stop loss / take profit orders: The user can create an order that only executes after reaching a certain price point and submit it to a bot, which will then automatically be executed upon fulfilling all conditions by a keeper. Slippage protection ensures that no manipulation is possible on the keeper’s part.
  2. Automated HF maintenance: A user can register with a bot that would track their health factor and automatically convert part of the collateral into underlying if the HF falls below a certain threshold. This ensures that the user will not be liquidated with high probability when they are unable to manage the position themselves. Slippage protection and other restrictions apply, same as stop loss orders.
  3. Automatic strategy management: Strategy bots would allow the user to automatically rebalance their funds between different assets and leverage levels, based on yield, risk and other factors. This may include rebalancing into the most profitable farm out of a predetermined set, maintaining delta-neutrality, and many other use cases. In fact, this type of automation can be fully extended into trustless fund management.

This is by no means an exhaustive list of what can be done with Gearbots. As the Gearbots ecosystem matures, we expect developers to create all kinds of novel QoL services and strategies.

b. Trustless on-chain delegated fund management

We have been through a year of centralized crypto venture blow ups. Fund management in itself has had the “trust us” frailties exposed and creditors have also been left at the mercy of the funds that borrowed from them.

While there are genuine fund managers who build strategies that outperform, there are very few who do this on chain and trustlessly. With Gearbots, it’s now possible to create an on-chain fund that’s easy to manage with automation while the depositors won’t have to give the custody of their funds to a third party. You’d still have to trust the competence of the fund manager but you can rest assured that the funds itself aren’t in their custody. Additionally, the choice of trades and farms they make is relatively safer owing to our AllowedList, as well as additional restrictions that can be applied on the bot level. While the risk is managed, per the trust model chosen and the proof of reserves can be confirmed on chain any time.

As for creditors, you don’t need to worry about the “Yoo, uhh, hmm”s or being lied to about the leverage and other factors at play. By using Gearbots you can lend money knowing what the trust model is and verify all of the fund’s claims on chain.

You can read more about Gearbots in the article we have already published below.

Wait, Wait, Wait…

Balancer and Aura? The integrations are a part of the code audit but do you really expect us to spill all the alpha in a single post, anon?

Redefine series will be back and revamped every week or 2 till V3 is actually deployed, more alpha and more details around announced bits will thus become a weekly activity for the next 3 months. So stick around to keep up with all the alpha as we redefine leverage and lending in our push to create the base layer of DeFi leverage.

ETA: end of Q2 with all the audits being back.

Thank you for reading this, the article was written by Ivangbi and has been posted on his behalf. If you would like to join — just get involved on Discord. Discuss, research, lead and share. Call contributors out on their bullshit and collaborate on making things better. Here is how you can follow developments:

JOIN DISCORD

--

--