FIP-27: Revamp FEI Bonding Curves

Motivation + Context

Bonding curves are the main way that new FEI enters circulation. They are also one of the main ways that the Fei PCV increases in size (the others being asset appreciation and yield farming). Bonding curves are a foundational mechanism that allow the protocol to grow.

The way that Fei bonding curves work is that each bonding curve uses a specific asset (e.g. ETH, DAI, or RAI). Users can acquire FEI off of a bonding curve using that asset. Bonding curves sell FEI at either a discount or a premium depending on whether the amount of FEI issued by that bonding curve is below or above a scale parameter. Each bonding curve has its own scale, discount, and premium parameters. Each bonding curve also has a cap parameter on the total amount of FEI it can issue.

This table gives a summary of the current Fei bonding curves:

Scale Cap Discount Premium Cap Reached?
ETH 100M 0.0% 1.0% No
DAI 50M 50M 0.3% 1.0% Yes
RAI 6M 10M 0.3% 2.0% Yes
DPI 10M 10M 1.0% 2.0% Yes

At present, all of the bonding curves except ETH have reached their cap. The caps should be raised for them to continue issuing FEI and growing the PCV.

Proposed Changes

We propose a series of changes to the bonding curves to help increase the amount of circulating FEI and grow the PCV:

  • Modify ETH bonding curve: premium 0.5%
  • Modify DAI bonding curve: scale 100M, cap 150M, discount 0.0%
  • Modify DPI bonding curve: cap to 50M, discount 0.0%
  • Modify RAI bonding curve: cap to 20M

Final Notes

A summary of previous bonding curve activity can also be found in this dashboard: Dune Analytics

Please share any thoughts on the above changes.


Can you please elaborate a bit about the rationale on the various discount/premiums? If the bonding curve applies a premium, does this mean that Fei Protocol is offering at higher market than price? Is any of these arbitrage opportunities absorbed as income for the protocol?

Otherwise raising the cap is agreeable. I also suppose this bonding curve is only available to PCV asset exchanges and not generally available to the public?

Finally, any planned bonding curves with Frax or Alchemix? You are in the D4 pool together.

Interesting proposal, some comments:

What will be the incentive to investors to use this bonding curve as the discount is 0%? Saving trading fees? When ETH price move quickly it can open an arb opportunity to use this?

Why not reducing all the premium to 0.5% to simplify?

At a first glance, 200M of DAI on PCV seems too high (more than 20% of PCV). DAI is currently 53% backed by USDC, thats a risk. If regulators want to put some pressure on stablecoins, DAI seems to be more vulnerable. Dont know the probability of it, but , anyway, it weakens our differential of Fei being a more decentralized stablecoin.

yes. bonding curve intake will be much slower when there is a premium compared to a discount. but they will still get used by arbitrageurs when prices on DEX’s and elsewhere randomly get desync’d enough

all bonding curves are open to everyone. I think the only GUI that uses them is etherscan but they are open smart contracts with no access controls on who can buy FEI off of them

I think LUSD has a pretty unique value prop here by being both governance-free and entirely collateralized by ETH. to me LUSD seems the most appropriate next asset for the PCV but Im open to other suggestions. definitely down to explore our stable friends. FRAX and alUSD are great innovative projects and I’m sure we will continue collab’ing with them either through PCV or other joint defi ventures

arb opportunities are probably the biggest one. each asset and each individual DEX pool will shift in price relative to one another over time. when they shift far enough out of sync an arb opportunity opens up. also when someone wants to shift from one asset to another, the bonding curve is like a 0-slippage constant fee dex from the other asset into FEI

none of these assets other than ETH have a chainlink oracle with 0.5% precision. they’re all >=1% precision. choosing a premium below the oracle precision would basically be subsidizing the purchases

this is a good point. it’s difficult to quantitatively reason about DAI’s USDC risk though. what do you think would be a more appropriate cap? keep in mind that with a 0% discount the bonding curve intake rate will be lower than before

1 Like

I agree, its a good move to add LUSD and we could wait a little bit more before including the other stables.

I agree that is difficult to quantify the probabilities. As an individual investor, I feel more safe to hold USDC than DAI, because I feel worried if they could freeze MakerDAO holdings in an extreme scenario.

Anyway, I think that if Fei rely too much on DAI it loses its differential of being more resilient to regulatory attacks. I see that each project in DeFi has its strength. One important thing that make Fei unique is the decentralization aspect. That’s why I would prefer having more LUSD, RAI, FLOAT.

For the DAI potential allocation through bonding curve, considering these concerns, I would suggest a max of ~10% of PCV. If something goes wrong with DAI in this extreme scenario, we are still in a region minted TRIBE could help.

A FEI-FLOAT bonding curve would be great - and possibly a TRIBE-BANK swap. If there is enough appetite for this integration I can talk to the team at Float.

I’m totally supportive of increasing the bonding curve caps to improve the diversification of the PCV and allow for more FEI supply growth in other assets.

Fei Protocol is not an individual investor but rather a collective, and would be at least equally at risk of frozen funds as MakerDAO. In that case, DAI feels like a much less risky option.

At this point I’d like to maintain a pretty hard line on not directly holding censorable assets in PCV, as that dependency can quickly grow as we’ve seen in other protocols. The exposure through 3Crv and other stable-swap liquidity is important as an on-and off ramp.

I think we should do at least 150m in DAI as an upper bound just to keep room if there are supply shocks in FEI like last month. Don’t want it all to come in ETH at such a time.

Definitely aligned on this. As these projects scale we can increase the amount allocated to them. I’d support and encourage raising the RAI bonding curve cap to 20m in this proposal.

FLOAT has too low of a market cap at $5m. In general treasury swaps are a pretty big deal and should be reserved for long term strategic partnerships. Doesn’t immediately invalidate them and I’m a huge fan of the project and team. At this stage I don’t think it makes sense.

Also wrt LUSD, they aren’t currently on many major lending platforms. We could deploy the LUSD into Curve but I think that is counterproductive as it furthers the dependency on 3Crv tokens. It would make most sense to create a stability pool PCV deposit to directly provide liquidity for liquidations on their platform and earn fees

Based on the above discussion I have edited the original post:
RAI cap → 20M
DAI cap → 150M

Moving this proposal to last call, will go up for snapshot in two days


Just to clarify, I agree with you here, I was not defending to hold USDC in PCV. Actually, my point was just having less DAI. I agree that being decentralized is one of the highest value within Fei and limiting DAI exposure is important to keep this. I see the need to balance the regulatory risk by one side and the peg risk by the other. Currently, we are in a good position with the Collateralization Ratio (and v2 is coming) that’s why I thought we could have a lower exposure to DAI. The decentralization aspect of Fei helps to differentiate it from the others.

That’s a good point. It is an interesting analysis to see which bonding curve people are using and how the ups and downs in crypto are affecting it.

1 Like

To add further to LUSD discussion, I think we should wait until we have the adapter ready to stake in their stability pool as that is the most useful (and uncorrelated) deployment strategy for LUSD besides FEI-LUSD liquidity which probably isn’t very useful at this stage.

There are enough changes in this proposal with the bonding curve caps that we won’t want to block until code can be developed, so we should move forward with remaining proposal items

Ok. I removed the bonding curve creation part of the proposal and left the rest unchanged

1 Like

Code here: FIP-26 + 27 by Joeysantoro · Pull Request #183 · fei-protocol/fei-protocol-core · GitHub

Note it does not change DPI discount to 0% because we aren’t raising Scale, it would have no effect


1 Like

I think this is great to increase circulation of Fei, and the caps seem reasonable. Will support this!

1 Like

Regarding the LUSD, it would be good if we could return with the proposal to create its bonding curve.

What do you all think about adding a sUSD bonding curve?

After we made these changes, the market moved a lot, but our DAI bonding curve was not used. Shall we change the discount parameter? I would support increasing DAI to 200M. They are reducing the USDC concentration and PCV increased in value since the time we discussed.

So I changed my mind here, as now DAI is less than 10% of PCV.


I am very in favor of an LUSD bonding curve, we are working on the deposit location now (probably native staking in stability pools), but a bonding curve by itself can be easily spun up for now.


Hi @storm

I am wondering if there is any interest to increase the discount on DPI from 0 basis points to say something like 25 to 50 basis points ?

ETH has recently outperformed DeFi to the point where a DPI token was worth as little as 0.075 ETH. The spot price is currently 0.084 ETH. Around these relative valuations it might be a good time to average into a larger position. A small discount should suffice. I don’t have any math to support this, but I think a 25 to 50 basis points discount would enough. This is between a quarter and half the discount offered on the original $10M purchase. 50 basis points on $40M is $200K.

The DPI allocation within the PCV has fallen to 0.91%. With the 0.0% discount in place, looking at the dune dashboard, there has been little DPI finding its way into the PCV. I think we would need to tweak the discount to drive a change from the statuesque.

1 Like

yea I think it would be good to increase the DPI allocation

not totally sure the best way to do it. Im curious to see how the FEI-LUSD balancer auction goes if that ends up happening. should similar mechanisms be used to acquire more DPI instead of using bonding curves? or would balancer investor pool reweighting be best? or just bonding curves?

1 Like