This proposal seeks to reduce the size of the dYdX active validator set to 31.
Abstract
Following the successful reduction of the validator set to 42 in October 2025, the active set currently has only 38 filled slots, reflecting a natural contraction in validator participation. This proposal formalizes that reality by setting the `max_validators` parameter to 31, aligning the on-chain limit with the actual state of the ecosystem.
Motivation
The case for this reduction rests on a few key observations:
A smaller, more engaged validator group improves responsiveness during upgrades or incidents, reduces coordination overhead in critical situations, and keeps the governance process efficient.
dYdX operates with a designated proposer set of 8 validators responsible for block production. The active set size therefore primarily affects coordination overhead and governance participation rather than core chain performance, making periodic right-sizing a natural part of chain maintenance.
Specification
We propose to reduce the active set to 31 by changing the `max_validator` parameter to 31.
Acknowledgement
We recognize that validators currently ranked #32 and above have contributed meaningfully to the network. This reduction reflects the natural evolution of the validator ecosystem and is in line with ongoing efforts to maintain a well-coordinated and efficient validator set. We are grateful for their service, and validators exiting the active set remain welcome participants in the broader dYdX community.
NextSteps
We invite validators to provide feedback on this post, if there’s no significant objection, we will submit the on-chain proposal on Monday, March 9, 2026.
We are against this proposal by the MEV Committee. First of all, many validators are winding down voluntarily including top institutional validators like Blockdaemon. There is no need to reduce the active set like this because naturally many validators are leaving given the current 0 revenues for validators in dYdX. Nevertheless, some validators continue to support dYdX despite this and plan to stay for the long term. This proposal dangerously plans to remove some of these few validators left very negatively affecting dYdX for the long term. It seems the MEV Committee is unaware of the current situation in dYdX and irresponsibly suggests this proposal bringing nothing positive to dYdX but exacerbating the already difficult situation with a mass exodus of validators. This is clearly reflected by the fact that Jordi_r, the MEV Committee member who wrote this proposal in the forum, according to his forum summary the last time he visited the dYdX forum was over 3 months ago in early December last year clearly indicating both a total lack of interest in following dYdX developments while receiving important funding from dYdX and a total ignorance of the current situation in dYdX regarding revenues, validators winding down and performance of validators.
It is not a ‘natural’ correction. Validators are winding down because of the lack of revenue and high infrastructure and bandwidth costs in Japan. The top priority is to maintain and support the few remaining validators, not accelerate this process even further.
What reality? You are just making the situation worse and ‘31’ is a total arbitrary number, how did you choose 31? From 31 to 38 several validators have top performance, participate in governance and support dYdX for years, what is the benefit for dYdX to just eliminate such validators like that? 31 is totally not ‘aligned’ with the ‘current state of the ecosystem’.
Ironically some validators from 31-38 are some of the fastest to upgrade whether regular upgrades or critical situations so by reducing the set like that you are actually worsening the upgrade responsiveness. Even more ironic that you mention governance process without visiting the forum for months, and again some of the validators from 31-38 have some of the best governance participation in the forum and on-chain
This is absolutely incorrect. The primary reason for dYdX v3 to move to v4 was to fully decentralise also the orderbook and matching engine by having it decentralised across each validator. You are missing the whole point and your argument is flawed, because following your argument then dYdX should reduce the active set to 1 or 2 validators for maximum ‘coordination’ and ’ performance’. No, the active set affects DECENTRALISATION that was the reason and the whole point for dYdX to migrate from Ethereum to v4 and you completely ignore that fact. The biggest value of dYdX is its decentralisation, at the start there were around 60 validators so already the decentralisation has been reduced by a lot hence reducing the value of dYdX.
In this blogpost you can read (Introducing Designated Proposers: A Governance-Controlled Path to Reliability and Performance): ‘designated proposers — a governance-controlled subset of validators responsible for proposing blocks. This is a fully deterministic enhancement to CometBFT that brings increased resilience, network performance, and operational clarity — while preserving the full validator set, stake-based voting power, and decentralized governance of the network.’, read well while preserving the full validator set and decentralized governance, that’s the whole point to keep the orderbook and matching engine decentralisation, the core reason of the migration from v3 to v4.
What do you mean by ‘we’? This is just you, Jordi from MEV committee earning $5K monthly for checking the forum once in months to share proposals full of flawed arguments.
That is not the natural evolution, the natural evolution is to increase decentralisation and the number of validators like in other networks that from 200 they plan to raise to 300 by the end of 2026. dYdX is antinatural, because if the reason to move to v4 was to fully decentralise the dYdX orderbook and matching engine via 60 initial validators, the natural evalution should be to increase that decentralisation or at least maintain it! It has already decreased a lot and the centralisation increased a lot and you claim this is natural and should be further accelerated? I wonder if you collaborate with Hyperliquid to harm dYdX.
To make the set more efficient what you should do instead is update the designated set of 8! ‘designated proposer set P’:
-Governance can reconfigure P quickly via an expedited governance proposal if performance issues arise
-Governance can curate and update P as needed, based on performance or behavior
-These validators should be:
Highly reliable — strong uptime, minimal missed blocks.
Operationally mature — responsive to upgrades and infra coordination.
Network-optimized — capable of maintaining low-latency peering and fast propagation
According to the above, the MEV Committee (aka Jordi) failed in their job. Since the initial set P was proposed to governance, a lot has changed. So much that for example one of the validators in the set P has worse uptime and performance in the last 90 days than a validator suggested to be removed in this proposal this is outrageous and it is not making the validator set more efficient but less.
Why are you talking as if you were the ‘ruler’ and decision maker of dYdX? It is trully remarkable to not even visit the forum for months or follow any updates while pocketing huge monthly revenue from dYdX grants, and then appear months later in the forum as the ruler of dYdX dictating what has to be done based on flawed arguments and as if dYdX governance was whatever you had to say rather than decentralized governance
Despite repeated attempts to contact the MEV Committee who selected the set of 8 many months ago no reply was received. The MEV Committee failed to monitor the performance of the set of 8 over time and update this set. Even worse, some validators below the 31 position that the MEV Committee is suggesting to remove have better performance than some in this set of 8. This is again very irresponsible, not only is dYdX grants paying a large monthly amount to some validators selected by the MEV Committee in the set of 8 with worse performance than other dYdX validators, but now the MEV Committee instead of updating the set of 8, suggests to eliminate from the active set some validators with better performance than some within this set of 8.
With clear data, for the last 90 days performance (https://analytics.smartstake.io/dydx/validators), validators in current positions 33 and 34 had better performance (99.6% and 99.4% uptime) than a validator in the designated set of 8 with only 98.7% uptime. The validators in the set of 8 receive large monthly funding to finance better infrastructure to maintain a top uptime, yet some validators without any large funding have better performance than some validator in the set of 8. And the MEV Committee is suggesting to eliminate some of the best dYdX validators by performance that have no funding to sustain such top performance. Meanwhile, the MEV Committee continues to burn dYdX funds in the current difficult situation while not fullfilling its responsibility of monitoring and updating the set of 8 according to performance.
According to the above, we are totally against this proposal and it represents an irresponsible threat to dYdX when already many validators are leaving due to the current economics. Instead, the funding of the MEV Committee should be reviewed according to their current tasks. The community demands clarity about the current monthly tasks of the MEV Committee and their total funding from dYdX. In addition, we demand that the MEV Committee updates the set of 8 according to the current and recent performance rather than a selection done and not updated since almost a year.
Once again, this shows that the MEV Committee is totally disconnected with the current reality of dYdX. Since the set was reduced to 42, many validators left voluntarily giving the lack of revenue and high costs of the infrastructure. More validators are leaving and the set is being reduced gradually without the need of any proposal. The main risk for dYdX is that if this trend continues the chain will continuously centralize and may even need to find validators. Some validators remain committed for the long term, upgrading fast, participating in governance and with top performance and the MEV Committee blindly suggests this proposal effectively eliminating some of the few top validators left that plan to be here for the long term supporting dYdX. This is highly irrresponsible and against the best interest of dYdX. Moreover, some of the validators suggested for elimation received delegation from @kpk based on performance so this is also negatively affecting the work of Kpk to maintain a strong and committed validator set.
Regarding the funding of the MEV committee (dYdX Grant: dYdX MEV Committee (2025 Extension)), according to this for example $30k was received for 6 months meaning $5k monthly revenue, this is huge considering validators are getting just $1k monthly since due to the fee holidays there was no revenue for months. But even more, it says that ‘Jordi will be the only contributor compensated for the committee’s work’, so Jordi is receiving $5k funding monthly, 5x more than what dYdX validators are receiving but was the value brought to dYdX? As mentioned, according to Jordi’s forum profile he hadn’t even visited the dYdX forum for over 3 months since early December. According to this blogpost (dYdX Grant: MEV Committee Renewal) ‘Jordi, a research contributor at Imperator’, it is important to mention that Imperator was managing before the OEGS deployment for a long time (dYdX Grant: Order Entry Gateway Service (OEGS) Operation), but last December dYdX Grants mentioned 'The previous OEGS deployment was hosted on Amazon Web Services (AWS), costing approximately $35,000 per month. Following an evaluation by the Grants Program, OEGS was migrated to bare metal servers, reducing operating costs to $6,000 per month and generating over $360,000 in annual savings (https://dydx.forum/t/dydx-grants-program-community-update-december-2025) indicating a lack cost optimisation by Imperator and a large unnecessary cost for dYdX for a long time since the new OEGS deployment at much lower cost ‘has consistently demonstrated best-in-class uptime and low latency, strengthening the performance and reliability for traders.’
In the previous proposal to reduce the set to 42 the main arguments were:
-Those ranked #43 and below in the active set now carry very minimal stakeweight, collectively representing only around 0.01% of the total stakeweight. This raises similar security concerns from the inactivity and apparent disengagement of these validators from the dYdX protocol (https://dydx.forum/t/drc-reduce-active-validator-set-to-42)
The above is very different to the current proposal. In the current proposal validators proposed for elimination are very active and engaged, in fact some of the most engaged and active which goes directly against the arguments in the previous proposal to reduce the set from 50 to 42. Also, some validators proposed for elimination have significant stake even delegation from both Kpk and the dYdX Foundation which also goes directly against the arguments in the previous proposal to reduce the active set. Currently it seems there are 38 validators (soon 37 after blockdaemon winds down) active, engaged and with significant stake, therefore following the previous active set reduction arguments it would make sense to reduce the set currently to 38. Above 38 the validators have significant stake and engagement in the network and hence they should not be eliminated
In summary:
-We are totally against this proposal and it should be seen as highly irresponsible and a threat to dYdX. Following arguments previously used to reduce the active set, currently it could be reduced to 38 but not more
-The MEV Committee/Jordi should stop ignoring community requests about the set of 8 update and based on recent performance update this set of 8 which is their responsibility and present the new set of 8 for governance approval
-The MEV Committee/Jordi should update the community transparently about their current tasks and total funding from dYdX