ARC: Spark Lend Profit Share Proposal

Greetings Aave community!

The newly proposed Phoenix Labs, an R&D company building the Spark Protocol upon the foundations of the Aave v3 codebase to deploy Spark Lend, a liquidity protocol centered around DAI.

We want to propose a profit-sharing program with the Aave community for the FOSS work done.

Motivation

We are ardent proponents of permissionless innovation that has enabled our space to create new and exciting solutions rapidly. However, the funding of said public goods is a problem in itself that we’ve seen come into less than desirable outcomes in years past.

While Business Source License creates breathing room, it does not address the value created for the community. To that extent, we want to propose a model that further rewards FOSS by proposing a profit share program to benefit Aave for the work done on Aave v3.

We are excited about the developments in this space such as Optimism’s retroactive public goods funding and hope to support a new value-sharing model for the industry.

Implementation details

We propose a profit sharing of 10% of the gross profits (cost paid to DAI suppliers subtracted from the revenue generated by DAI borrows) from the DAI market in Spark Lend.

The profit share will take effect when the DAI market size reaches 100 million and will expire two years later.

  • DAI market = portion of the pool with underlying asset ERC-20 DAI
  • Profit = Gross Profit = Revenue - COGS
  • Revenue (B) = revenue from borrows in the DAI market
  • COGS (S) = Cost of Capital = amount paid to suppliers in the DAI market
  • Profit share = 10%
  • Minimum DAI borrows = 100 million DAI
  • Duration = 2 years after reaching minimum DAI borrows

Profit share paid to Aave = 10% * (B - S)

Estimated profit share

The main factors driving protocol revenue - and resulting profit share payable to Aave - include:

  • TVL
  • Utilization rate
  • Interest rate curve

In a bull case scenario where Spark Lend DAI market has a high TVL and utilization at optimal levels, we expect to be able to distribute over >2,000,000 DAI to Aave within a profit share period two years.

Payment Implementation

MakerDAO governance will ratify and issue payments on a quarterly basis based on the accumulated profit.

We want to present two options for the Aave community to consider. The latter splits a minority of the profit share to purchase AAVE tokens to align both organizations further.

  1. Entire profit share direct payment in DAI to the Aave Treasury at address 0x464C71f6c2F760DdA6093dCB91C24c39e5d6e18c
  2. Split between direct payment and purchase of AAVE
    1. 8% profit share direct payment to the Aave Treasury at address 0x464C71f6c2F760DdA6093dCB91C24c39e5d6e18c
    2. 2% profit share used to purchase AAVE tokens held in the Spark Protocol Treasury

Snapshot vote

Snapshot link
Start: 2023-02-24 00:00:00 UTC
Duration: 1 week
We’ll consider the snapshot vote valid once a quorum of 80k AAVE is reached.
We understand that there is no need for a subsequent AIP vote.

Next steps

We welcome feedback from the Aave community and would like to present the Snapshot vote in 2 weeks.

The launch of Spark Protocol is pending MakerDAO governance approval, which will be finalized at the end of March.

11 Likes

Hello @PhoenixLabs and kudos on your hard work on Spark Lend.

The Aave Chan Initiative is supportive of stablecoin diversity and Spark Lend is an interesting new opportunity in a long lasting synergistic relationship with the MakerDAO protocol.

While we will need a bit of time to form a definitive opinion on the product implementation and potential risks of a D3M based lending protocol (D3M DAI is by definition unbacked), we have a keen to explore this with you.

in terms of payment Implementation we lean towards option 1 to keep things simple.

4 Likes

Big morning for Aave - first this, then GHO on Tesnet…

I’ll start by giving you my congratulations on the development of Spark Lend - as a Maker delegate as well its refreshing to see folks committed to renewing the lending opportunities for Maker.

Looking at these products side by side, it is a fork and clone of Aave.

I am glad you are coming to the forum, and acknowledging the role Aave’s FFOS played in the build. Without doing such, it feels like there would have been some frustration.

There is overlap in the markets offered - but difference in the UX. Hyperdrive seems like an interesting integration to create for a cleaner, fixed-rate borrowing. For markets offered (at launch), see below:

  • DAI
  • ETH [E-Mode Enabled]
  • Lido wstETH [E-Mode Enabled]
  • WBTC
  • Savings DAI (DSR-locked DAI) [Collateral Only]

This was outlined in a post in the MakerDAO forum, here.

Anyways - to comment on your proposal, we too would opt for option #1:

  1. Entire profit share direct payment in DAI to the Aave Treasury at address 0x464C71f6c2F760DdA6093dCB91C24c39e5d6e18c
2 Likes

Just a quick point of information. All DAI is backed, including any and all D3M DAI.

While the proposal is ongoing on Maker as well, I also think that launch of this can introduce several new collaboration points between Maker and AAVE. Will be exciting to see how it progresses

1 Like

Thank you for coming to Aave’s Governance Forum.
My request is as follows.
・Profit sharing should be reconsidered instead of expiring after 2 years.
・AAVE tokens to be purchased should have a lock period.
I look forward to Spark Lend!

1 Like

Hi everyone!

We have put up the Snapshot vote for your consideration and we have also updated the information in the ARC, including a minimum quorum for the snapshot vote of 80k AAVE.

2 Likes

Thanks @PhoenixLabs

Congrats on the development and we are in favour of the partnership!

Given there is no info on how the purchased Aave will be used under option 2, we are leaning towards option 1 alongside @MarcZeller and @fig.

While this might make things a little complicated, I wouldn’t be opposed to Aave receiving sDAI instead of DAI.

1 Like

I voted for all of it to be given to Aave Treasury. Absent of a clear governance structure for spark protocol and knowing who will control this, I don’t think it makes sense to have it accumulate Aave.

1 Like

At the moment it is clear that the latest movements of MakerDAO are deceptive regarding profit sharing while using it as some kind of fake “good partner” exposure.

Some arguments:

  • Due to the nature of their DAI reserve (allowing for borrowing DAI that gets minted on demand by the D3M mechanism), the main supplier (not only, but by design majority) of DAI is MakerDAO itself, meaning the whole yield goes to them as “profit”. This is in the order of 3% on notional, currently ~150-200m DAI, so ~6m.

  • MakerDAO chose to configure the Reserve Factor parameter of their Aave v3 instance (SparkLend) to 0% on DAI, which means that profit is 0 from it, and consequently, profit sharing is 0.

  • By organizational design, it is fairly clear that SparkLend is simply a subsidiary of MakerDAO to expand to different DeFi models, something completely legitimate. But in this case, they are using an accounting ruse to make a virtue of how much they contribute to partners and FOSS, while being factually false.
    At best, this is total marketing deception, at worse, this is just an orchestrated behavior from MakerDAO to trick the Aave DAO participants.

  • Let’s put things into perspective:

    • SparkLend is an instance of the Aave v3 software, using v3-core, v3-periphery and v3-deploy.
    • SparkLend benefits from all the FOSS (MIT by general rule) tools that the Aave DAO and its contributors continuously create in public, for example, via BGD.
    • SparkLend has also forked the Aave UI of Aave Companies.
    • Aave contributors (me included) have been always helpful when contacted by people of SparkLend, regarding technicalities of Aave, same as any other alternative instance.
    • The Aave DAO supported DAI from day 0, creating the biggest secondary market for DAI.
    • When D3M appeared, the Aave DAO implicitly supported it.
    • SparkLend presents arguments of “technicalities” making more complicated profit sharing. My argument: no substance, I know the Aave protocol a bit better than them, and the problem is lack of will.
  • The situation by itself is so ridiculous and “cheap” that shows the appreciation (lack of) that MakerDAO has for Aave: we are talking about maybe not even $1m yearly in profit sharing, in exchange for using the whole software of SparkLend + the growth to MakerDAO because of it. All of it is from completely new revenue to MakerDAO.

  • MakerDAO/SparkLend had no obligation to propose any profit sharing, compensation, or anything of the sort to Aave: they could do whatever they wanted with Aave’s software.
    However, they decided to do some commonly called “cloud-chasing”, associating themselves as some kind of “partners” of the Aave DAO doing profit sharing.
    It is simply disgusting and to be ashamed of.


Personally, I have interacted in the past with contributors to MakerDAO and they have always been exemplary. Also, I think MakerDAO has done a lot for the space with primitives surrounding SAI first, DAI now.
But this is not really the way to go.

17 Likes

Hello all, speaking as the crafter of this proposal perhaps I can fill in some details of what we were aiming to do and what’s changed since the proposal went out.

First let’s address the technicality of this line:

@eboado correctly identifies this as the “Reserve Factor” parameter which at the time of writing the proposal we were planning to use the reserve factor as revenue to Spark. You can see evidence of this in the DaiInterestRateStrategy that we deployed which contains borrowSpread to add a markup to the borrow side beyond the rate being paid out to suppliers. The problem we ran into is that only fixed % reserve factors are supported by the code. You can see this here and please correct me if there is a workaround we missed.

One alternative we investigated at the time was to have the reserve factor at 100% and burn the excess liabilities after the fact. You can see evidence for these changes here:

Closed PR to burn the excess liabilities (Mar 29-May 9): Add DaiCollector by hexonaut · Pull Request #2 · marsfoundation/sparklend · GitHub
Mainnet Spell setting RF to 0% after it had originally been configured to 100%: https://github.com/makerdao/spells-mainnet/blob/master/archive/2023-04-28-DssSpell/DssSpell.sol#L127

In the end it was realized it was simpler to forward all revenue to Maker and do quarterly off-chain settlement. Any revenue paid to Spark will also send 10% of that amount to AaveDAO as agreed. This is the intention of the original agreement, and it’s still very much intact.

As a representative of Phoenix Labs, we have no direct connection to MakerDAO. We have observed the rules voted in by the DAO called “Endgame” which allows for SubDAOs to receive credit lines from Maker at the “Base Rate” provided the rules are followed. As part of these rules Spark SubDAO is obligated to pay the full Base Rate to MakerDAO, and there is no authority to give revenue out that is not owned by the SubDAO.

The feeling is mutual, and I hope we can continue.

1 Like

Hi all,

I’ve made a short post discussing the profit sharing proposal on the Maker forum.

I personally favor the payment to be reevaluated to ensure proper alignment and fair recognition for Aave community’s work on developing the v3 codebase. I listed a few ideas for how this could be accomplished, such as basing payment on Maker D3M revenue instead of Spark reserve factor, adding in a minimum payout amount, or granting tokens to AaveDAO or AAVE token holders. This all depends on Maker governance decision of course, but I personally feel confident we’ll arrive at a solution that maintains a positive Aave<>Maker relationship and builds a solid foundation for continuing collaboration in the future.

Posting solely in a personal capacity, not as a representative of Maker or BA Labs. I own both AAVE/stkAAVE and MKR tokens and have a vested interest in the success of both protocols.

8 Likes

Here is what i think, pretty openly and transparently

I think you were well aware that the reserve factor (lets not fix on the particular feature here, and call it spark DAO revenue) would have been 0 for quite a while longer than it would have needed to reach 100M TVL. Even if the reserve factor could have been set correctly (and i still think it can be, but i am not gonna disclose how obviously) i have no doubt it would still be zero, why you may ask? because you need to be as competitive as possible to be successful in what is basically a vampire attack on a protocol that not only provided all the technology you are using, but also incentivized DAI with hundreds of millions and greatly helped its adoption by being the first one accelerating the multichain reality we currently live in. You claim you have no authority over Maker decisions (which i find hilarious) but i think it’s very simple to amend this proposal on the spark side to begin the payment when the Spark DAO revenue will be greater than zero. I am willing to wait personally.

3 Likes

I understand you put some effort into trying to create a mechanism on the logic, fully respectable.
But sounds pretty fragile to use software and understand only post-deployment that some terms agreed are maybe problematic to implement (together with a bit scary security-wise).
Definitely sounds like this uncertainty should have been clarified in this post in advance.


I would say it is simpler to keep the ceremonious language and formalisms apart. Phoenix Labs presented a goodwill proposal for the Aave DAO seeking acceptance, on behalf of SparkLend, which exists solely because MakerDAO allows.
Then I’m sure there is a framework on who has authority over what, in order to get access to D3M, revenue flows (back and forth), etc.
But there is a clear profit de-facto, which has been darkened.

So:

  • If the profit of SparkLend comes from other sources (e.g. future airdrop), could you clarify to the community if 10% of that until the 2025 date is included in the agreement?
  • Which motivation MakerDAO has to actually now set a meaningful RF (or equivalent distribution)?
6 Likes

Update from Rune

2 Likes