Smaug: Debt Based Treasury Management


  • Proposal to partner with Sushi, UMA (and possibly Yearn) to build a debt based treasury management protocol on top of Sushi’s ecosystem.
  • Ideally leveraging MISO, Sushi’s AMM and possibly BentoBox.
  • Sushi gets a way to raise funds from locked treasury assets that they can buy back if needed.
  • Sushi then has a way to make revenue off of other protocols that need a more diverse treasury.


Over the past few months, I’ve been tinkering around with trying to solve the treasury diversification problem through the issuance of convertible-debt-bonds collateralized by treasury tokens.

I had debated though if there would be much interest in the purchasing of these styles of tokens.

It seems with the Sushi attempts at fundraising, there is not only high interest in locked token bonds, but, also a need for a controlled market of these which allow for the sale of these assets in a regulated fashion.

Ultimately, I think the Sushi raise represents a great opportunity for this to be developed as a joint venture between myself, Sushi and UMA Protocol (who have done great works in developing novel treasury tokens.

So with a bit of a pivot from the initial concept, let me present “Smaug”

Core Problems:

  1. As Lex_Node pointed out in the strategic raise thread, it is likely (from the Telegram legal case) that the direct sale of any existing, liquid, treasury tokens could indirectly taint the security status of the rest of the tokens.

  2. Treasuries sit on massive piles of their own token, but lack methods to diversify or liquify. The moments when they need cash the most are the times in which it is usually the worst to sell them.

  3. VCs usually want discounts for lockups, open markets want dilution over lockups, and treasuries just want money that doesn’t hurt their tokenomics.


Smaug is a protocol and market for selling treasury backed, convertible debt bonds.


DAOs use tokens in their treasury as collateral to issue an SBond.

SBonds contain:

  • X amount of the DAOs token
  • A declared strategy in how that token is used (perhaps via Yearn vaults?)
  • A maturity date
  • The minimum amount of USD requested for the bond.
  • An annual interest rate
  • The option of if the bond pays ‘only interest’ or ‘interest and appreciation’
  • A voting strategy if the bond has pass through voting rights.
  • An unlock range (optional)

Each ERC20 token issued from the bond contract represents $1000 USD of initial debt.

Process of an SBond:
For example, Sushi could choose to set aside 5M units of Sushi for collateral.

The 5M units of Sushi are placed into the SBond contract as collateral.

They decide they want to raise a minimum of $25M of collateral, and are willing to have an interest of 10% year + token appreciation.

The SBonds could then be auctioned off above the floor price.

This leaves Sushi with $25M of USD to run their operations.

Overtime Sushi has two options:

  1. Pay back the debt to each bond position to buy back the underlying collateral.

  2. Let the SBond expire in which case it converts into its share of the underlying $SUSHI tokens.

Unlike a traditional financial instrument this model:

  • Creates a maximum exposure for the treasury (their underlying token collateral)
  • Leaves an avenue to buy back all or a portion of the collateral if they exceed expected growth.
  • Gives funds a way to partake in allocations but forcing a lock mechanic, so there is less concern around tokens being dumped.
  • Gives funds a better way to ensure benefit from growth and not getting positions bought out from under them on a standard convertible debt denoted in tokens.
  • Ensures that the bidding is more open to get a market rate rather than a negotiated deal. If the treasuries deal doesn’t sell they can reprice and reissue.

Smaug Market:

The Smaug Market would be an after market where issuers (DAOs) and allowed buyers (regulatory dependent) can buy and trade existing market SBonds through AMM pools (provided by Sushiswap)

Since each SBond is always issued at the equivalence of $1000 of USD debt it means that each unit from issuers are comparable.

While 1 Sushi SBond might have better earning terms than 1 YFI SBond, both were worth $1000 of USD debt at the start of term.

This allows flexibility within the markets where the pricing is focused on preference and payment terms, and not units.

It also allows for creative opportunities of funds or even DAOs creating automated index pools of specific assets so they can buy broad exposure to the market.

Such a market also gives the DAOs a distinct opportunity to use their profits to buy back their own SBonds at any dipping price point to reduce their collateral exposure.

This essentially creates a liquid market both for issuing SBonds and maintaining a diverse treasury.

Eventually this market can serve as a way for wealthy DAOs to be buying investment exposure to new DAOs without 1:1 negotiations.

How would this benefit Sushi:

  • The underlying system for these auctions could run on Sushi’s MISO. It is also a natural upsell for the MISO listing partners sushi has right now, and the farming partners that sushi incentivizes.

  • The underlying AMM would be built off of Sushiswap.

  • We may be able to build this as a BentoBox app (depending on the current flexibility of that system - needs review)

Both of these meaning that Sushi’s protocol would slowly be earning long term treasury assets of all other DAO protocols.

How would this benefit UMA:

  • UMA’s oracle system would be used as the core of price tracking for these products.
  • It may also be possible to build these bonds entirely as a specific type of UMA Synthetic asset.

How would it benefit the Yearn Ecosystem:

  • Issuers have the ability to add strategies to tokens allowing them to gain underlying yield.
  • With Sushi being a core part of the Yearn Ecosystem, and Yearn being leaders in securely managing vault strategies, it would make sense for these strategies to be private Yearn vaults.

How would this benefit Sushi as an SBond issuer:

  • Instead of raising funds on the open market, Sushi would ensure the funds are locked on terms they choose.
  • Sushi’s maximum exposure would be the collateral tokens they choose.
  • Sushi would have an opportunity to buy back the collateral if Sushi grew faster than expected.
  • Unlike private VC deals, the auction process would ensure that Sushi gets the best price for these assets.
  • The assets are non-dilutive.

How would this benefit Sushi users:

  • Sushi gets no dilution today.
  • Sushi MISO and AMM products get new volume and revenue streams.
  • Possibly built on BentoBox so that we drive more liquidity volume into Bento.


Why Smaug?

Smaug is the name of the great fire drake that lived in Lonely Mountain in The Hobbit. He guards a mountain of ancient treasury, and cares very much about that treasury not being in circulation.

Do SBonds have to have an unlock curve?

No an issuer would choose to have the tokens either unlock all at the date of maturity or to vest in a linear fashion between two date timestamps.

Does a DAO have to pay the interest rate by the maturity date?

No. The SBond basically tracks debt as an internal metric that accrues overtime.

Meaning that every block it gets slightly more expensive for the DAO to buy back their collateral from the SBond.

If by the maturity date, the DAO has not paid back their total debt, then users can begin to claim the proportional share of remaining collateral, in which case they are surrendering their claim to any of the USD debt.

Who gets the yield?

That is an open question. If a bond has a strategy, should that yield flow through to the seller, the buyer, or go against the sellers debt position.

Currently, I think going against the debt position is the best option as this allows the DAO to automate a manner in which the debt pays back overtime.

Can DAOs pay back from revenue?

Yes, DAOs could set up a revenue wallet that makes calls to the SBond position to automatically pay back their debt from incoming revenues.

Can DAOs buy their own SBonds?

Yes, its possible that DAOs may find that buying an SBond on the secondary market is cheaper than paying back the equivalent amount of debt.

DAOs would be able to buy back and cancel this SBond token, lowering the proportional outstanding debt.

This measure is actual crucial as it acts as a backstop to the market price of SBonds and their underlying debt drifting too far a part in the later time stage of the bonds lifecycle.

Would Smaug be a part of Sushi?

No, I am proposing it as a sperate partner protocol similar to Yearn, Pickle or Keep3r.

However, it would use Sushi’s existing tooling like MISO and the AMM, meaning Sushi would generate revenue from all Smaug related actions which is the more lucrative part of the stack anyway.

There may also be opportunities to build it out as a BentoBox DApp, which is something that still needs more review. But, the end goal is having BentoBox as a flexible smart wallet so it should be doable.

more faq to be updated as they evolve


Made an account just to say…

This is genius.

Please do this.

1 Like

This is a very creative way to better align incentives of all parties and improve Treasury Management by providing additional options to diversify assets.
Although convertible debt is a common instrument in TradFi, this looks like a first in DeFi.


RULER is ready to be that partner!

Let me first explain that RULER Bonds has been in production as a fix to the issues faced with Rari Capital after their exploit, how do we raise without selling the tokens?

Enter RULER Bonds.

This is an outline of the Bonds process however we have also made some additions that will allow decentralized private or vested style raises for new or existing projects. This means that we can perform a decentralized raise via a Dutch auction and still have the underlying tokens vested to a specific set of terms.

The model also allows for teams to raise capital for specific purposes with the intention of repayment and paying the investors a yield for the service. This could even be under collateralized raising as well.

The point is RULER has been planning to use the MISO launch pad as a way to distribute sale tokens for the raises, the idea of a Dutch auction that allows a market for find a fair price is right in line with our product ethos. There is also the ability to raise at a fixed price using the orderbooks created specifically for RULER tokens.

Below is a basic concept with the linked tweet from one of our developers, Alan, on how this could work. Given we have the infrastructure and are a fair launch project ourselves we saw it as a great way to align and grow another tool for defi rather than build out another project with another token.

RULER would love the opportunity to discuss how this can work and what is required to launch (which is not much I might add).



I appreciate the enthusiasm, but RULER is currently not anywhere near the type of functionality you’d need for this. So it wouldn’t get us there any quicker.

If however, RULER is proposing to build the token type that may be a different story.

But, the SBonds are:

  1. Vesting
  2. Interest and appreciation bearing
  3. Vote/yield passthrough
  4. Callable by the debt issuing party
  5. Convertible to the underlying collateral by default.
  6. Disputable by optimistic oracle and dispute resolution (this is a big part of why UMAs oracles would be useful)

But, if RULER wants to come to the table and talk about working together to architect that token type, I’m happy to chat.

At this point, I’ve received a lot of interest from other teams interested in the idea, so it seems likely worth pushing forward even if Sushi decided not to use it for issuing themselves.


I think you missed the part where I did mention we are ready.

Vesting - tick
Interest and appreciation bearing - tick
Vote/yield passthrough - work
Callable by debt issuing parting - tick
Convertible to underlying collateral by default - tick
The oracle is not really a massive hurdle come to oracle pricing at a particular point in the future.

I do not believe that this will be ready for a raise SUSHI today, the proposal would be for RULER to help build out the remaining of this idea for Smaug.

Happy to chat further on the matter but most of what you are saying we actually have built. It may not be there for you to use right now but we can talk.


Got it - happy to arrange a call with the team and understand this better then.


tl;dr: I deployed Adam’s Smaug design to mainnet and minted a SUSHIsBOND using UMA contracts.

Adam and I had a call earlier today where I asked him some questions about Smaug and how it differs from some of the designs UMA has proposed (like Range tokens). Turns out, they are quite close! So close that with minimal modifications I was able to deploy a version of Smaug.

What I deployed:

  • The token is called SUSHIsBond
  • Expiry is set to July 31, 2024 (about 3 years)
  • Each SBond token is backed by 200 $SUSHI
  • At expiry, each SBond will expire to be worth $1000 of $SUSHI if $SUSHI is above $5
  • If $SUSHI is below $5 at expiry, each SBond will be worth exactly 200 $SUSHI

These numbers match the example Adam used in his proposal: if you take 5M units of Sushi, you will be able to mint 25k SUSHIsBOND which will expire to be worth $25M USD of Sushi.

Note one difference: the example I deployed is a zero-coupon bond (all interest is embedded in the price of the bond). This is different from what Adam proposed where the SBond pays interest at maturity. Both are mathematically equivalent, and this contract could be easily changed to pay interest at maturity too.

How I deployed this:
Although I have a technical background, I’m not a solidity developer. I used UMA’s example repo launch-LSP to create a LongShortPair contract with parameters that match the Smaug design. Mathematically, the Smaug design is a Range token where there is no upper range, so this was easy to construct. The only complication was adding a new financial product library to match the Smaug payout.

The UMA contract that holds the SUSHI collateral is deployed here.
The SUSHIsBOND ERC20 token is here.

Note that I haven’t had a second set of eyes sanity check the parameters, so this shouldn’t be used for sizable trades until that happens.

How this differs from a Range token:
The Smaug design is like a Yield Token without risk of liquidation. That means this token will pay out a fixed dollar amount ($1000) of SUSHI at expiry provided the contract has enough SUSHI to actually pay $1000. Since each SUSHIsBOND is set to hold 200 SUSHI, it will have $1000 to pay so long as SUSHI is above a price of $5. If the price is below $5, you just get the 200 SUSHI (whatever that’s worth). This looks like a Range token where the bottom of the Range is $5, but unlike the Range token design, the Smaug token has no upper Range (or “call option”) that pays out additional upside if Sushi rallies.

What’s next:
I wanted to build this Smaug design to showcase how UMA could be used to build Smaug using audited and proven code (and how it’s quick and easy to do). I also agree with Adam in his post that there is a big opportunity to build treasury backed, convertible debt like products: it’s like creating a bond market for DAOs!

In terms of Sushi’s strategic raise, this is a very different option compared to selling tokens (this is like selling debt). There are pros and cons for Sushi in terms of this approach, which I attempted to map out in this post. Regardless, these debt-like structures are very cool and I think the Smaug design, or Range tokens (or some other similar product) have an important role in the future of DAO financing.


I will change my approach. :grin: Probably a little too forward in my initial approach - thank Adam for the feedback - but we are definitely very keen to help in anyway we can.

How do we help support this new product? RULER has a number of features I feel could help benefit a product like this with SUSHI. How can we support the current UMA proposal to work with you guys closer?

One interesting function is the ability to create Private Sale type vesting for tokens for vested allocations. Allowing the vested schedule to also be traded on a secondary market.

What I think we could offer is an add on to the raise as above, this particular setup as an sBOND makes a lot of sense for projects that have some sort of credibility and value to their token already being able to raise.

Also I think there needs to be a focus also on security around who can launch such a product with a whitelisting process built in when raising against an already launched token.

Our model is pretty similar to what @hal2001 has already outlined as a 0 coupon bond, but again could be altered to accrue the interest rather than an upfront settlement. It just works in a slightly different fashion.

We believe that this product could cater for not only current projects wanting to expand their treasuries but new projects launching their token seeking capital by adding in a vested token model for defi.
Basically DeFi SAFTs - no paperwork, no middle person to create.

It would also add a level of flexibility to the aforementioned “yield token without liquidation model” outlined.

What can we do to be part of this cool concept?


Thanks Ted

As I said lets hop on a call. Grabbing a time on the calendar you sent me.

This should be an open and collaborative process where any team who wants to get involved can bring concepts and tech to the table and we can iterate to see what we can come up with that is best for all parties.

At the end of the day, treasury diversification is easily a $100B+ problem and a critical public good for this space, so I don’t see there being one team, one solution here. Makes more sense to welcome any stakeholder to the table to try and build a standard across the industry so it can really thrive!


Wow. This is a fascinating read. It is seriously outside of my depth of knowledge for treasury management, but I want to follow along and see where this goes.

If this makes it to a vote, you have my SUSHIPOWAH.

Brilliant initiative.