The below represents the opinion & viewpoint of Synapse Labs, not of Synapse DAO.
Overview & STIP Process
The Synapse DAO applied to the STIP program with high-level goals to improve the user experience for bridging into the Arbitrum Ecosystem. The grant’s primary aim was to introduce a new bridging paradigm that enabled sub-30 second bridging on large dollar sums for users, slippage-free.
The grant allocated 1.5m ARB towards “supporting faster and cheaper bridging”, as L2 ecosystems without reliable, fast, and cheap bridging dramatically suffer. Another 500k ARB was allocated to support existing Synapse supported Arbitrum tokens in the form of gas & fee rebates.
At the time of the STIP application, the objective of achieving incredibly low slippage bridging was imagined to be done through on-chain constant sum pools, incentivizing LPs. Contracts were being developed here, and were internally audited. The ARB grant was planned to be used to incentivize LPs to create on-chain concentrated bridge liquidity.
However, this initial plan was created over June - September 30, 2023.
The first round of STIP voting ended October 13th, and Synapse DAO’s proposal was not initially included in round 1 of STIP. Synapse DAO was included in the STIP Backfund proposal on December 6th.
Throughout this time, Synapse’s roadmap pivoted from on-chain concentrated liquidity pools to an intent-based RFQ architecture. The RFQ system was developed to achieve cheap and fast bridging, replacing 10-20 minute bridge times. It relies on off-chain liquidity vs AMM liquidity, so incentivizing and growing it required a new approach.
Thus, after the STIP Backfund was approved, the Synapse DAO(in SIP-32) specified in detail the grant allocation, particularly regarding the 1.5M ARB towards supporting the objectives of faster and cheaper bridging.
Given the shortened timeline, Synapse DAO had to be adaptable in how to best achieve the STIP goals set forth. The vote had two options with regards to allocation amounts, and Synapse DAO voted to use 750k ARB to incentivize RFQ relayers. Synapse Labs, its employees or affiliates did not vote in this proposal. Synapse DAO contributors posted regular updates regarding the speed and user benefits of the newly launched RFQ model in the bi-weekly update thread.
At the end of the grant period (ending March 29th), Synapse DAO, executed on SIP-32 1.24m ARB directly to users in bridge rebates (The remaining ~3k ARB was returned here). Once total relayer volume was calculated, the remaining 750k ARB was distributed to relayers.
Bridges & Introduction of Intents
Bridges play a crucial role in the ecosystem of each blockchain, enabling on & off-boarding seamlessly as compared to native bridges. Optimistic rollups in particular benefit the most from deep external bridge support, as users desire to know they can quickly on or off-board without being subject to a 7 day withdrawal period.
To solve the tradeoff between chain finality and bridge time, bridge protocols started searching for solutions to solve the time constraint in an effort to create a cohesive ecosystem between L2s and Ethereum, resulting in intent based systems popularized by UniswapX.
Synapse RFQ aimed to provide maximum speed and minimal costs for cross-chain orders, ensuring reliable liquidity for onboarding and offboarding to Arbitrum. For these systems to thrive, initial incentives for early relayers were a compelling idea to mitigate fixed costs and build trust, helping to bootstrap an eventually robust relayer ecosystem.
Synapse RFQ
Synapse RFQ is an intent-based bridging architecture, with several benefits:
- Fastest: average bridge time is less than 20 seconds
- Cheapest: Synapse is the cheapest bridge on 75% of routes across the Ethereum L2 ecosystem
- Most Capitally Efficient: RFQ bridging doesn’t rely on stale AMM pool liquidity
The architecture relies on off-chain “relayers” that compete with each other to offer the best quotes and speed to users.
While the benefits of the user are clear – incredibly fast cross-chain transactions, the benefits for the relayer are to earn a marginal fee spread.
However, these relayers take on a variety of risks and compete for the spread between the accepted quote and the requested bridge amount. This includes:
- Re-org & Finality risk: Relayers only earn a spread if they are the first one to complete the destination transaction. This incentives relayers to take the maximum finality risk on, otherwise they do not earn any spread. If an origin blockchain the relayer supported reorgs, they would lose their own funds that they relayed for that period. In the case of L2s, this often means trusting a centralized sequencer and taking on L1 reorg risk.
- Protocol Risk: Relayers are using a brand new set of smart contracts to complete transactions
- Custody: Participating as a relayer requires large sums of money to be managed securely on chain, but also being able to automatically interact with different blockchains.
Relayers take on the responsibility of these risks, as well as the associated engineering costs.
Incentivizing early relayers was proposed as a way to skip the chicken & egg issue, decrease initial fixed costs, build trust with relayers, and bootstrap a healthy competitive relayer ecosystem. Given this, relayer diversity is a key goal of Synapse RFQ.
Synapse DAO engaged several funds, market makers, and developers that would be suitable relayers, but the process took longer than expected. Protocol contributors encountered several issues, from off-chain engineering problems, to building trust with operators to run new software. In some cases, providers were happy to provide capital but did not have any engineering resources to manage the implementation.
We were overly optimistic about integration timelines compared to the STIP timeline, and had underestimated the complexity and resources required to onboard new relayers. Not accomplishing our goal was a failure of Synapse DAO contributors.
Intent Based Protocols & Off-Chain Diversity
Given the nascent state of intent based protocols, Synapse DAO assumed that liquidity incentives, like 2020-era AMM incentives, could act as a catalyst for bootstrapping relayer ecosystems.
However, from the data, it’s clear that intent based protocols are an order of magnitude more work for market participants to provide liquidity to. Throughout the STIP campaign period, a number of research papers on intent based bridges came out, which I believe the Synapse DAO could have learned from in its approach.
Li.fi published a report on March 15th on trade-offs on intent-based bridges, particularly focused on which agents are responsible for what percentage of order flow. The conclusion was that categorically, there’s a strong principal agent problem for intent based bridges.
For example, “In Across, a single agent often wins the majority of orders over $500”, and for DLN, “ [on] orders above $25,000, [a single] agent captures over 95% of the order flow”. Intent based bridges face significant hurdles to building diverse ecosystems, and structurally have economic dynamics that create somewhat concentrated orderflow.
ARB Grant & RFQ Relayers
Synapse Labs built a reference implementation of a RFQ relayer and is open-source on GitHub here. This implementation supports a number of different tokens, as well as many different rebalancing methods, such as using CCTP for USDC, or native bridges where it makes sense.
This reference implementation was built for market participants to either directly use, or reference for their own implementation. While working with a number of potential relayers, we realized that their timelines were months out, and that to solve the market bootstrapping issue, someone would need to be the initial large relayer. Synapse Labs stepped up to run relayers, hoping to achieve the user oriented goals of SynapseRFQ and STIP.
Overall, Synapse accomplished the goals it set out in the STIP application, and consistently posted about our progress on the Arbitrum Forum. Some major goals we are particularly proud of seeing the protocol accomplish:
- Bridging speeds under 15 seconds on Arbitrum – 100x faster than earlier this year.
- The cheapest bridging option on 75% of routes between ETH <> L2s.
- $3.05bn in bridge volume throughout STIP campaign
Given the short timeline, the intent based market dynamics, and engineering issues, Synapse DAO, throughout STIP, did not achieve its goal of a wide number of relayer participants. At the end of the program, the Synapse DAO executed on its SIP-32 passed proposal, and the 750k allocated ARB was transferred to the relayers who participated at the time of STIP, run by Synapse Labs.
The first relayer started running on January 12, funded in part by an external LP. Synapse Labs reached out to a number of participants to onboard them to run relayers using the reference implementation built. Throughout the course of 1.5 months, progress was being made to onboard relayers, but timelines were not fast enough in conjunction to achieve the STIP goals. Given the STIP mandate of “faster and cheaper” bridging, Labs stood up a second relayer to scale up RFQ liquidity, while additional relayers were onboarding. This second relayer went live March 13. ARB was distributed to Synapse Labs for its two relayers, and to the external LP that funded the first relayer.
Neither Synapse DAO nor Synapse Labs intended such relayer concentration. A combination of the shortened STIP timeline and unexpected engineering challenges led to the concentration in relayers, and thus RFQ grantees. In hindsight, we at Synapse Labs wish we had recognised the timelines of potential relayers’ and engineering implementation sooner, and went back to the Synapse DAO to propose a new vote to re-allocate the ARB sooner. Additionally, there were oversights in both the original STIP application as well as the Synapse DAO proposals to specify not only goals of the grant, but also more specific parameters around the grant process and mechanics, especially when unexpected issues arise.
Ultimately, Synapse Labs is aligned with the Arbitrum ecosystem and with the original intention of the Synapse DAO STIP grant. To that end, both Synapse Labs and the external LP are happy to return the ARB to the Arbitrum DAO, if the DAO/Foundation believes this wasn’t in the spirit of the STIP program.
Please reach out with next steps or any additional questions.