A fork in the road.

My previous post was a long-form exploration of the future of Synthetix. This post is going to take a different approach. It will be a simple 🤷‍♂️ and practical guide to help the community reason about the experiments we must run over the next few months. All of the proposed experiments below would not be possible without the new more flexible V3 architecture, it has taken time to get this right, but we are now well-positioned to rapidly test and iterate.

The practical considerations are listed below:

• How do we roll out V3 and Perps V3?

• How do we scale collateral?

• What is the limiting factor in scaling collateral?

• Do we enable non-SNX collateral?

• What is the optimal strategy for maximizing value to SNX holders?

• How do we convert fees into value for SNX holders?

• Are there untapped traders on other chains?

• How do we efficiently access these traders?

The current state of the protocol:

• Synthetix V2x is live on Ethereum.

• Synthetix V2x is live on Optimism.

• Synthetix V3 is deployed to Optimism, but so far unutilized.

• Synthetix V3 is deployed to Ethereum, but so far unutilized.

• Perps V3 and Synthetix V3 are deployed to Base testnet.

• The majority of liquidity is in the V2x system, on Optimism.

• There are other EVM networks we can deploy to; Arbitrum, Avalanche, Polygon etc

• We will introduce a new SNX staking pool in V3.

• We do not currently have a viable strategy for true cross-chain liquidity.

This is a small aside, but an aspect of the project that creates confusion is the naming conventions. We have Synthetix V2x and V3 as well as Perps V2 and Perps V3; this is hard to follow for those outside the community. I propose we create names to help distinguish these releases. We can lean into our prior use of astronomical names by choosing galaxies for these releases. We can combine them rather than having distinct releases; so Perps V3 on Synthetix V3 can be called Andromeda. Even if we have deployed some of V3, it will not be the Andromeda release until we have all of the included functionality.

Experiment One - Deploy Andromeda to Base

This deployment contains both Perps V3 and Synthetix V3 plus non-SNX collateral. We need to decide which non-SNX collateral to include as well as the fee split between SNX and non-SNX collateral. This will help us determine whether there is interest in supplying collateral on Base. I believe we should enable both USDC and ETH (we should hold off on LSDs initially) as collateral. My reasoning is that this is a distinct deployment, and we will be able to determine the balance of demand between ETH and USDC as collateral. The other reason to deploy only ETH and USDC is that it will mean that we do not put any pressure on Optimism liquidity. If we were to enable SNX collateral on Base, it is possible a large percentage of LPs might migrate, which would reduce liquidity on Optimism. Andromeda will enable users to LP ETH or USDC and borrow sUSD; this sUSD will be used by traders on Base and will not be fungible with OP/ETH sUSD. We will also need to incentivize a USDC/sUSD curve pool on Base. Additionally, we should enable a USDC/sUSD wrapper so that traders with USDC can easily get sUSD without needing to swap into sUSD.

I believe we should start with a fee split of 50% to LPs and 50% to SNX stakers. In terms of paying those fees to SNX holders, I believe we should buy back and burn SNX on Base. This will require creating a wrapped SNX token on Base bridged from Ethereum, luckily there is a SIP for this, unluckily I already voted against it 🤦‍♂️.

One of the critical factors will be formalizing the split of fees between SNX Holders, LPs, and Integrators. There are a number of proposals being discussed by the community, but there appears to be some consensus towards 20% for integrators. This would result in a split of 20/40/40 between integrators, LPs, and SNX holders.

What are we measuring?
We are unfortunately measuring a number of things simultaneously; this is not ideal, but we must start somewhere.

1. Demand to LP ETH on Base

2. Demand to LP USDC on Base

3. Sensitivity to fee share

4. Sensitivity to liquidity incentives

5. Demand to trade Perps on Base

6. Demand for traders to wrap USDC to sUSD

What does success look like?
$10m+ in LP liquidity for ETH and/or USDC @ 50% fee split$20m+ daily volume or 25% of OP volume

What if it fails?
Thankfully, points 1-4 are fairly distinct from point 5; as long as we have sufficient liquidity, trading should not be impacted. If we have no demand from LPs, then trading will not be possible, and we will need to rerun the experiment with SNX collateral or with higher incentives. If we have low LP demand, it will allow us to tweak the fee split or LP incentives to hopefully increase demand. If this fails, there are several possible conclusions that we will need to test in follow-up experiments. One is that there is minimal demand for LPing on Base; we can test this by running the experiment on another network like Arbitrum.

What if it works?
If this experiment is successful and we have high demand from both Traders and LPs, we must then decide on our next experiment. There are two options: replicate this experiment on Arbitrum or replicate it on Optimism.

Experiment Two - Deploy Andromeda to Optimism

If experiment one works, I prefer to deploy Andromeda to Optimism; to be clear, this is Perps V3 + V3 + USDC and ETH collateral. This is a large shift in our approach, as it would mean that SNX would not be used as collateral. However, legacy perps would still be operating, so we would get a reasonable test of demand for SNX-backed perps vs. ETH/USDC-backed perps. Because while the upgrades to Perps are valuable, they are not necessarily going to drive incremental volume. This new deployment should have incentives for both LPs and traders in the form of OP/SNX.

What are we measuring?

• Demand for LPing ETH on Optimism

• Demand for LPing USDC on Optimism

What does success look like?
20% of liquidity of V2x
50% of the volume of V2x

What if it fails?
If there is minimal demand from LPs or traders relative to the legacy system, this will provide valuable information. If LPs are scarce for ETH or USDC, we can tweak the fee split and see whether it is an issue of incentives. If there is still minimal demand, we can conclude that it is likely that SNX collateral will be required to scale the protocol. If LPs are plentiful but trading volume is low relative to the legacy system, it will be less clear what the issue is, but this seems like a fairly unlikely scenario. In either case, the next step, after tweaking the fee split, should be to reintroduce SNX collateral. This new three-token collateral system (SNX|ETH|USDC) should ideally get a new name (Barnard) to distinguish it from what is deployed on Base.

What if it works?
If it works, the next step is to wind down the legacy perps markets and lean into USDC/ETH collateral on Optimism. That means more incentives for traders, LPs, and Integrators. I should point out, that each of these experiments will require a SIP and will also be modifiable via SCCP, so this will be a continuous rather than a discreet process as we try to dial in the optimal parameters.

Experiment Three - Deploy Carina to Ethereum

If non-SNX collateral is successful on both Optimism and Base, the next experiment is to deploy an optimized version of perps for mainnet called Carina. This will be purely ETH-backed (ETH as well as LSDs) and will be designed to support stablecoin designs via delta-neutral issuance; projects like Ethena are beginning to deploy this tech, and we need a perps market on Ethereum to support it. This experiment is fairly independent of the others, so whether it is successful or not does not meaningfully impact the roadmap.

Experiment Four - Deploy Andromeda to Arbitrum

The rationale for waiting to deploy to Arbitrum until we have more data is that it is one of the most popular networks we have yet to deploy to. There is a large and active user base of blueberry farmers on Arbitrum, and we want to maximise the impact of our rollout there. It is worth pointing out again, the reasoning behind not using SNX collateral across multiple networks is to ensure we don’t fragment existing SNX liquidity, which would impact the trading experience across all networks.

Experiment five - Deploy SNX chain on the OP stack (Project Draco)

If the first two experiments are successful and stable, the next step is to deploy the Synthetix Chain; this chain will serve several purposes. The first is a place for Synthetix governance; the second is to allow SNX holders to borrow against their SNX, resulting in issued sUSD. This sUSD will be able to be used to trade on any supported network. The third purpose of this network is to have a single place where fees are distributed. However, this may be unnecessary if burning SNX on each network is the preferred method of fee distribution. It is also possible that sUSD issued on the SNX Chain could be used as collateral on other networks. The debt from the V2x system would be migrated into a V3 deployment on the SNX Chain as part of Project Draco.

Synthetix 2024

If all of these experiments are successful, the state of the protocol at the start of 2024 will be Synthetix Andromeda running on three Ethereum L2s: Optimism, Base, and Arbitrum. Andromeda contains the latest version of Perps using both ETH and USDC collateral, with USDC wrappers for easy access to sUSD. Synthetix Carina will also be running on Ethereum, a latency-optimized version of perps using ETH (and LSDs) as collateral. Project Draco will be under development, with a plan to migrate all governance functions to the Synthetix Appchain in 2024. sUSD borrowing via SNX would also be enabled as part of Project Draco. It is possible that sUSD issued on the Synthetix Chain could be introduced as a form of collateral as a future experiment alongside ETH/USDC and other forms of collateral. This would allow SNX holders to borrow against their SNX and boost collateral and liquidity on other networks.

I want to finish with some thoughts on this major shift in strategy. Removing the reliance on SNX collateral through these experiments is a major change for Synthetix. Many people thought that using the SNX token as collateral was insane back in 2018, in fact, many still do. Some people also thought it was a great idea to take this concept and tweak it a little, RIP LUNA. But regardless of where you stand on SNX collateral, it is unquestionable that it has worked until now, and it likely could continue to work far into the future. SNX collateral allowed the project to survive and thrive prior to DeFi summer, but it is time to test whether SNX collateral has been an impediment to growth in recent times.

It is possible that removing SNX collateral will simply not work, that SNX is too critical to the functioning of the network. That will be very valuable information upon which to plan our rollout to multiple networks in 2024. However, I have already noticed a significant shift in the conversations I have been having about Synthetix, and I think many of us who have clung to the purity of SNX collateral in the community may have underestimated how much it has impacted participation in the protocol. Regardless of how certain we have been historically that SNX collateral was the right approach, there have been very vocal skeptics in the DeFi community. I’m beginning to realize that for every one of those vocal skeptics, there have been many more quiet skeptics who were not willing to openly criticize Synthetix but have simply chosen not to engage at all. One of the biggest outcomes of these experiments will be to see just how much additional enthusiasm we might generate with the introduction of non-SNX collateral.

Decision-making under uncertainty is never easy; it requires a willingness to take risks and be open to making mistakes. This openness and epistemological humility have been a key aspect in the success of the Synthetix community for many years, but it is time to lean into this even further as we challenge some of our core beliefs.