There’s been a lively discussion on the forums regarding Arbitrum’s first governance proposal, AIP-1. The level of engagement has been remarkable and a testament to the strength and dedication of the DAO community. There have been some requests for comments from the Foundation about AIP-1 and requests for further clarification, and that’s the intention of this post – to engage with the questions being raised and also to provide some additional context that may be helpful in understanding and taking a stance on AIP-1.
One acknowledgement before diving in is that in hindsight it’s clear that the Foundation definitely could have done a better job on communication – both by providing more clarity directly in AIP-1 and also by being quicker to respond. We are in the process of staffing up the Foundation and will strive to improve our communication going forward, and we hope that this post provides the necessary clarifications for AIP-1.
When it comes to setting up a DAO, there’s a chicken and an egg problem. Let’s put aside the question of funding the Foundation for just a second (we’ll get there shortly), and focus on the handoff of the core technology and upgradability of the Arbitrum chains. In order to do that, certain parameters need to be decided including transferring the code as-is to the DAO, creating the security council, setting the time delays for code upgrades, establishing the initial textdraft of the Constitution, establishing the AIP proposal mechanism, the initial set of validators and DAC on Nova. Some have suggested that even the initial values of these parameters should have been chosen by the DAO, but that’s not technically possible. When the system was given over to the community, these numbers had to be set with some value – i.e. the DAO governance smart contract had to specify the thresholds that allowed for a proposal to be passed and had to specify a security council that would be live during the initial handoff. There simply was no community that could have voted on these numbers, and the very act of creating the community required these parameters to be specified.
The Foundation was created in order to kick off and facilitate the DAO decision making process in the first instance – if the Foundation didn’t exist, simple operational questions such as “who hosts the various governance forums where these discussions occur” as well as broader questions such as “who kicks off governance” would go unanswered and revert back to the chicken and the egg issue.
And with regards to setting up the Foundation, there were similarly lots of decisions that needed to be made ahead of time in order to get it off the ground. These include basic questions like where the Foundation is domiciled, who its initial directors and employees are, as well as questions about its mandate, mission, bylaws, and funding. Now of course, there’s room for discretion about what these values should have been, particularly with how much initial funding the foundation should receive, but two things that we’d like to make abundantly clear in this post are (1) these values were set with the sole goal of making sure that Arbitrum is positioned to succeed and compete with its peers and (2) the decisions made are well within market range when compared to Arbitrum’s peer projects and foundations.
Nothing here is intended to minimize that there was very clearly a shortfalling in communication. One of the mistakes in the drafting of AIP-1 was a failure to note at the outset that this proposal was intended to act as a ratification of the initial setup of both the Arbitrum DAO and the Foundation that has been created to serve the DAO; this was spelled out at the end of AIP-1 in the section “Steps to Implement”, but it should have been made clear earlier on. Moreover, the post was written in the future tense (a reflection of the fact that it was in the future when we were drafting the post and did not catch this) and understandably this likely led to some of the confusion on the point that this was a ratification. As discussed above, there is a chicken and the egg that needs to be solved when decentralizing a network, and the point of AIP-1 was to inform the community of all of the decisions that were made in advance. The communication of those decisions was clearly not articulated correctly. We hope that this post will properly delve into the initial thought processes that went into making these decisions.
The ratification of AIP-1 was also intended to allow for the Foundation to assume some obligations relating to infrastructure and service providers that have been supporting the network in the past, establish its relationship with service providers, contractors and hire personnel on a moving forward basis and ensure that the ecosystem has what it needs in order to continue to thrive in an extremely competitive environment.
The establishment of an initial distribution of tokens to the Foundation is a normal course of practice in setting up ecosystems like Arbitrum, and this was always the intent when AIP-1 was submitted for ratification concurrently with the announcement of the Arbitrum DAO. While our focus is not competing ecosystems, it is useful to compare to understand how this figure was originally chosen and whether allocating 7.5% of the token supply to the Arbitrum Foundation is reasonable. The three closest peers with live or announced tokens are Optimism, Starkware, and Polygon, and indeed 7.5% is lower than all of these:
- The Polygon Foundation and Ecosystem Growth contracts currently hold more than 10% of the token supply. This is after years of significant grantmaking and other distributions which has led to the successful growth of the Polygon ecosystem. In other words, there remains more funds available despite years of historical distributions.
- The Starknet Foundation has earmarked all 50.1% of the community token supply into different buckets, with 10% of the token supply earmarked for its foundation to maintain a strategic reserve to fund ecosystem activities.
- When the Optimism Foundation was established, the token allocations were also much more clearly defined, similarly to the Starknet Foundation. The Optimism Foundation established that 20% of the tokens would be distributed for retroactive public goods and that 25% of the supply would be established for an ecosystem fund. 5.4% of the token supply were allocated to the Partner Fund which explicitly is to “be distributed strategically by the Optimism Foundation to grow the Optimism ecosystem for the benefit of the entire Collective.” Additionally, an additional 8.8% of the token supply remains unallocated for future growth programs.
While not as directly comparable because they are Layer-1 blockchains, the Avalanche Foundation originally received 9.26% of the token supply and the Solana Foundation received 12.5% of its token supply, respectively.
Looking at the above data points, Arbitrum is unique in that it is the only one among its peers in which the vast majority of the community tokens are currently controlled by the DAO on-chain. And also uniquely, The Arbitrum Foundation is the only one that even asked for the community’s ratification of these decisions.
Now, even though it’s an established practice, one may fairly wonder why The Arbitrum Foundation (or any of its peer foundations) needs a significant token budget and why this cannot all happen via on-chain voting. And this is again a place where we should have provided more clarity. It’s not only because of “voter fatigue”, which was intended to describe a more effective way to process smaller grant requests, but often it’s more fundamental and necessary for Arbitrum’s ability to efficiently and effectively fulfill its mission.
At times, it may be in the best interest of the DAO to partner with a company or organization that either won’t or doesn’t have the ability to publicly negotiate deals on-chain due to confidentiality and other operational requirements. Without a strong foundation that has the ability, mandate, and resources to engage with these entities in real life, the Arbitrum ecosystem would be ineffective and at a serious disadvantage as compared with its peers. Please take note of Polygon’s partnership with Nike and Starbucks and Optimism’s partnership with Coinbase’s Base chain. In all of these cases, there was no public process at all and the first time that the public became aware of these relationships was after they were agreed upon.
While it would be incredible if all traditional companies agreed to do everything on-chain, this is not realistically going to happen. Not having a well-funded Arbitrum Foundation that has limited discretionary power to engage and enter deals with large institutions on behalf of the DAO could be severely damaging to Arbitrum as Arbitrum would be the only ecosystem that could not compete for these deals.
Indeed we are aware of similar conversations going on with fintech, social media, and gaming companies right now, and we at the Foundation, are ramping up to engage with and educate these players on Arbitrum, and the initial distribution at inception was intended to provide the Foundation with the resources it needed. It’s clear that it’s within the DAO’s best interest to have a voice in these arenas, and the DAOs ability to choose the Foundation’s directors give it direct control to replace, remove, or augment its representatives. The current (and any future) directors of the Foundation are (and must be) signed fiduciaries of the Foundation and committed to acting in the DAO’s best interest in accordance with the Foundation’s mission.
The Foundation’s purpose is to fulfill its mission to foster the development and growth of the Arbitrum ecosystem – this is its mission as stated in its governing documents and the Foundation cannot act in ways that are counter to these documents. As a Foundation Company, it does not have shareholders or partners that it is accountable to, and instead, significant governance authority over the Foundation has been given to the holders of the $ARB token. In other words, the Foundation exists to effectively facilitate and serve as a steward for the DAO and the Arbitrum ecosystem.
The specific governance authority that the Foundation’s governing documents provide to the DAO include:
- Appointment and removal of directors (who are fiduciaries to the Foundation and are tasked with carrying out its mission), including the ability to expand or reduce the number of directors (see Bylaws § 6 and Articles of Association § 13); and
- Amending the Foundation’s governing documents (subject to certain limitations like compliance with Cayman Islands law, see Bylaws, § 3(a)) (in fact, the Foundation is prohibited from making any amendments to its governing documents that would adversely affect the rights of the DAO without the DAO’s explicit approval, see Bylaws, § 3(b)).
The DAO can also cause the Foundation to take any other action subject to compliance with Cayman Islands law (and in fact, according to the Foundation’s governing documents, in the event that the decisions of the Foundation conflict with those of the DAO, the DAO will prevail unless prohibited under Cayman Islands law, see Bylaws § 7(c)).
Simply put, the Foundation works for the DAO.
We believe that a lot of the negative sentiment around AIP-1 was driven by confusion around the notion of AIP-1 being a ratification and not a request. For those that didn’t realize that this was a ratification, they may have been surprised to see that the Foundation’s tokens have already been separated and begun to be utilized.
In hindsight, we should have made it clearer that this was indeed a ratification of a decision that had been made, which explains why there has been movement in the treasury that had been set aside for the Foundation. The Foundation treated this as a ratification of its initial setup, not an initial grant request from the DAO Treasury, and indeed has begun to use these tokens in the interest of the DAO, including conversion of some funds into stablecoins for operational purposes.
The most significant miscommunication is likely the unclear classification of the initial distribution to the Foundation in the original pie chart in the governance documentation. The documentation described the allotment to the DAO treasury address and to the Foundation in one bucket as “DAO treasury.” It would have been more clear had that been bifurcated into two, distributing the bulk of the funds to the on-chain DAO-controlled treasury and a portion to the Foundation, in service of the DAO.
In hindsight, there was clearly a lot that could have been better communicated. We do want to stress, however, that AIP-1 was posted on the forums simultaneously with the governance announcement on March 16 (one week before the airdrop went live), and in the “Steps to Implement” section it did make clear that this was indeed a ratification of funding that had already occurred. Again, we could have done a lot better with regards to making sure that this was clear to all, but lest anyone think there was any change of intent, we do want to point out that this was announced at the outset. And we will strive to communicate better going forward.
P.S. There have been some questions for clarity as to the role of Lemma LTD, the poster of AIP-1, and their relationship with the Foundation. Setting up a foundation, managing the governance process and setting up and managing the DAO tooling takes a lot of work, and indeed this is one of the reasons that it was so critical to get the Foundation up, running, and funded prior to launch. To assist in these roles as it staffs up, the Foundation is working with a few different service providers in order to bootstrap itself and be in a position to properly serve the DAO. One among those is Lemma Ltd which provides several services to the Foundation including monitoring and participating in governance forums.