[ACP-247] Delegation Multiplier Increase & Maximum Validator Weight Reduction #248
Replies: 14 comments 30 replies
-
|
Great work, and I love to see people teying to tune these parameters. My reading of these measures won't increase the number of small validators, by increasing their capacity (4× to 24x), it would just mean large validators end up attracting much more delegations and start running extra validators to avoid the issue of max weight. It definitely makes running validators much more profitable, but at the cost of decentralization imo. It's an issue with user behavior that a lot of people just end up staking with larger validators/known brands. Think increasing the multiplier makes sense, but maybe by 3 x (to 12, just a guess) as opposed to 6x , while sticking with the 3x drop in max weight of a validator. Anyway, it's great to see these efforts |
Beta Was this translation helpful? Give feedback.
-
|
Thank you for your feedback, though:
|
Beta Was this translation helpful? Give feedback.
-
|
Nice incentive to get more validator on Avalanche, but maybe having ventilated reward to force them to stay active for many years |
Beta Was this translation helpful? Give feedback.
-
|
You mean vested? |
Beta Was this translation helpful? Give feedback.
-
|
Yess vested unlock 🔑 |
Beta Was this translation helpful? Give feedback.
-
|
That is tricky, though it is in my thoughts But it's hard to get consensus on two tweaks, let alone on 3 or 4. My goal is to get consensus on this so that others can be done swiftly after. |
Beta Was this translation helpful? Give feedback.
-
|
With ACP-247's 24x multiplier, it would be attractive for smaller validators to get into the space and in order to fill up the nodes, they would need to work for it. Thus, that would mean more of us will actually start promoting Avalanche. Also, i noticed that on other chains, validators are running their own loyalty or reward program (random airdrops or points) in other to win over delegations. Not saying that this will or should happen but it does show that with enough incentives, validators are motivated. |
Beta Was this translation helpful? Give feedback.
-
|
Yes, I just also read today about 'arcade tokens', and I believe they could be used by L1s to incentivise delegations - see my comment below. |
Beta Was this translation helpful? Give feedback.
-
|
ACP-247 also fundamentally changes validator economics for L1 builders. Here's why it matters: The Problem TodayWith the current 4x delegation multiplier, a validator with 10,000 AVAX self-stake can only accept 40,000 AVAX in delegations. This severely limits revenue potential from delegation fees—making validator operations barely viable as a business. What ACP-247 ChangesThe proposed progressive delegation multiplier (up to 24x for smaller validators) transforms this. That same 10,000 AVAX validator can now accept 240,000 AVAX in delegations. The impact depends on your delegation fee: At 2% Delegation Fee:
Income multiplier: 1.37x At 5% Delegation Fee:
Income multiplier: 1.83x At 10% Delegation Fee:
Income multiplier: 2.43x Why This Matters for L1 OperationsDirect operational funding: Here's the crucial part—L1 builders can now self-fund operations through Primary Network staking. Instead of burning through venture capital to pay for node infrastructure and development, the delegation fee becomes a revenue stream that sustains the business. Token incentive mechanism: L1 builders can set their delegation fee at 10% (or higher) and offset it by distributing their native L1 tokens as rewards to delegators. This creates a powerful tool: delegators get AVAX staking rewards plus L1 tokens, while the L1 builder captures enough AVAX fee revenue to fund operations. At 10% fees with ACP-247, that's 1,340 AVAX annually—real operational runway. Bootstrap alternative to VC: This fundamentally changes how L1s get funded. Instead of needing Series A/B capital to bootstrap validator infrastructure, they can start with a smaller self-stake (10,000 AVAX) and use the delegation mechanism to generate immediate revenue while distributing governance or utility tokens to their validator community. It's a form of self-sustaining bootstrap capital. Validator market competitiveness: Different L1s can now compete for delegations by offering different L1 token incentives. One L1 might offer 2% of their token's value annually; another might offer governance rights. This creates genuine differentiation and makes the validator market competitive rather than dominated by established players. LST integration: Liquid staking protocols like Benqi randomly delegate AVAX to validators. L1 validators become much more attractive targets when they can offer something extra (their token) in exchange for delegations. This creates a flywheel where LSTs direct capital toward L1 validators, and L1s use that capital to fund operations. Bootstrap funding alternative: With 240 million AVAX currently unstaked, there's massive capital waiting to be deployed. ACP-247 makes L1 validators the natural destination for that capital, turning what was a cost center into a sustainable revenue engine that funds the entire L1 operation. The bottom line: ACP-247 transforms L1 validator participation from a capital drain into a sustainable, self-funding business model where staking fees and L1 token incentives create a virtuous cycle of growth and ecosystem development. |
Beta Was this translation helpful? Give feedback.
-
|
That's terrible with half of the validators not having any/ having tiny amount of delegations. Obviously one needs to focus on balancing the delegation fees, but at a modest 5% fee i think they should receive some. This ACP however will incentivize more validators to lower their fees and seek delegation, effectively balancing incentive to talk about node space and for delegators to delegate (as they earn more fee). Overall very supportive of this and would like to see this passed. |
Beta Was this translation helpful? Give feedback.
-
|
OK. I finally get it. The use case of L1's rewarding small validators with their token makes sense. One potential is that many 2000 AVAX validators with large delegations could outweigh larger validators. Perhaps that would be good though because it's democratizing effect, people flock to validators who share their ACP viewpoints. Or it could be bad because a Ponzi could gain a lot of ACP votes and maliciously try to change the protocol. Is this a concern? Maybe it's just a problem we wish we had ;) I'd be curious to know how many L1's are running Avalanche validators? I wonder what percentage of them would find it sufficiently useful lock up that much capital to use it to attract investors via rewards. As a delegator, I'd be tempted by getting L1 rewards. (This echo's the original envisioned reward mechanism of subnets being validated by Avalanche validators. I'd hoped subnets would incentivize avalanche validators to validate subnets for their token's rewards.) Separately, I'm not sure how you chose the 1 million hard cap? Nice work on this ACP! |
Beta Was this translation helpful? Give feedback.
-
This number wouldn't make sense, big or small. The scenario is completely disincentivizing L1s to run Primary Network validators, so it's misleading to think it would help or not help the argument imho. |
Beta Was this translation helpful? Give feedback.
-
|
How is it disincentivizing L1s to run Primary Network validators? |
Beta Was this translation helpful? Give feedback.
-
|
Because they won't make much money from it right now.. they'd need to have hundreds of thousands of AVAX to make a few tens of thousands of dollars per year from self-stake and delegation rewards combined. But with this ACP, it starts to make much more sense! |
Beta Was this translation helpful? Give feedback.
-
Strong Opposition: Proposal degrades security and incentivizes centralization theaterI strongly oppose ACP-247. While the stated goal is to help small validators, the specific parameters chosen (24x multiplier + 1M cap) create perverse incentives that harm network security and encourage artificial node splitting. Below are the critical risks this proposal introduces, followed by verifiable inaccuracies found in the specification. 1. Security Risk: Drastic Reduction in "Skin in the Game"
This proposal claims to reduce concentration but effectively allows entities to control the network with significantly less collateral.
By increasing leverage to 24x, you are making it 80% cheaper for an attacker to acquire consensus weight. This degrades the economic security of the network by limiting the self-stake liability required to wield influence. 2. Economic Risk: Incentivizing "Sock-Puppet" Node Splitting"The 24x multiplier change has minimal impact on P-Chain resources" The proposed parameters create a distinct economic incentive for large holders to split their stake into chunks of exactly 40,000 AVAX.
This encourages Decentralization Theater—bloating the validator set and P-Chain load with multiple nodes controlled by the same entity, purely to bypass the weight cap and maximize leverage. 3. Market Risk: The "Field of Dreams" Fallacy (Oversupply)"Unlocks Dormant Validator Capacity... converts uneconomical validators into active, capital-efficient operations" This claim relies on a fundamental economic error: confusing Validator Capacity (Supply of slots) with Delegator Capital (Demand for staking). The supply of AVAX is finite (720M Hard Cap). Increasing the delegation multiplier from 4x to 24x does not create new AVAX to be delegated; the pool of delegatable funds is inelastic and finite.
In a market with massive oversupply of capacity, liquidity does not "trickle down" to unknown, dormant validators. Instead, it consolidates around trusted market leaders and incumbents who can now leverage their existing reputation to capture 500% more of the fixed supply. This creates a "Winner Take All" dynamic that will starve smaller validators rather than "unlocking" them. Additional Technical InaccuraciesFurthermore, the proposal relies on mathematical errors and potentially flawed assumptions in its justification:
|
Beta Was this translation helpful? Give feedback.
-
|
OK. These papers change things for me. It puts the cost of a liveness attack close to $190M at today's price. I think the chain has to small of a market cap to withstand an attack under this ACP. Perhaps we could revisit it once PMF is achieved and some real price floor is there. |
Beta Was this translation helpful? Give feedback.
-
|
Can we consider a significant reduction to the delegation weight in consensus in tandem with this ACP if it is to proceed? |
Beta Was this translation helpful? Give feedback.
-
|
That is not possible @MoonBoi9001 , as the network will be paying for more security by paying less, so the only way to do this wold be to also reduce staking rewards overall, but I believe that would be too much of a change for validators, and it won't be seen well. |
Beta Was this translation helpful? Give feedback.
-
|
If I'm reading this correctly, you're saying that since delegation staking rewards are proportional to delegation weight, then a reduction in delegation weight would mean a reduction in staking rewards for delegators (and subsequently validators due to less fee collection) If I've understood that concern correctly, then this could be solved by increasing the rewards payout for delegators by the same magnitude as the reduction in the delegation staking weight. So for example 1/x delegation weight would receive a reward of x fold the current reward. Then it nets out to no difference. |
Beta Was this translation helpful? Give feedback.
-
|
I've completed a detailed security analysis of Sybil attack vectors against Avalanche under both the current system and proposed ACP-247. Bottom line: ACP-247 does not materially increase Sybil attack risk. While the upgrade enables a more operationally feasible attack path, multiple practical barriers make successful attacks highly improbable. Key Findings1. Capital cost unchanged (~$1.4B for BFT control)
2. The Authentication Paradox destroys the 40K approach
Both paths fail. 3. Pattern recognition detects all strategies
4. Delegation market provides built-in defense
5. Operational burden remains extreme
Success Probability by Scenario
No scenario exceeds 10% success probability when practical constraints are included. RecommendationsBefore deployment:
6 weeks post-deployment (if no threats):
Ongoing:
Full AnalysisFor a comprehensive deep dive covering all five attack scenarios, the authentication paradox, delegation market dynamics, timeline analysis, and probability reasoning: Verdict: ACP-247 is safe for deployment with standard community monitoring. The upgrade improves staking economics without introducing vulnerabilities more severe than the current system. Proactive entity-level constraints within 6-12 months post-deployment are prudent risk management. |
Beta Was this translation helpful? Give feedback.
-
|
Very good idea, thanks |
Beta Was this translation helpful? Give feedback.
-
|
I strongly support ACP-247. Avalanche Foundation Delegation Strategy Instead of the Foundation running its own nodes at scale, it would be far more decentralizing and economically supportive to delegate to community-run validators, especially those who: have a delegation fee ≤10%, have a proven history of reliable validation, have maintained high uptime and performance across previous subnets or P-Chain validation, and are geographically diverse Redirecting Foundation stake towards community validators would amplify the positive effects of this proposal. It also creates a strong incentive mechanism: validators who perform well get increased Foundation delegation, which further encourages best practices and long-term participation. ACP-247 is a simple, low-risk economic improvement that unlocks idle validator capacity and strengthens decentralization. Combined with a community-focused delegation strategy, it would create the healthiest validator ecosystem on Avalanche Strongly agree with moving this forward. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you @mahofmahof ! Yes, this is something that I've been discussing and pushing at the Avalanche Foundation. Hopefully this ACP will make it more doable! |
Beta Was this translation helpful? Give feedback.
-
|
@iJaack, as u know validator can stake upto 1 year, is it make sense to encourage them to validator for longer period like 2, 3, or 5 years and get more rewards like 1.2x, 1.4x, or 1.7x respectively? Very high level analysis done on avascan validator list, there are roughly 200 validators currently staked for 1 year. In nutshell i am thinking to encourage deep pocket validators to lock their stake longer and create some kind of investment vehicle hoping big institutions might build something on this. |
Beta Was this translation helpful? Give feedback.
-
|
@MoonBoi9001 your Sybil attack hypothesis is more than valid, in theory for sure. But on deeper thought, I want to run some modeling to understand if these Sybil are actually possible given the current conditions and potential changes in AVAX distribution post-ACP-247. What I want to validate is that even though your preoccupation is correct, it may not apply to this scenario based on the current dynamics and stake distribution. I'll run some models in the next few days, I'll write here my findings. |
Beta Was this translation helpful? Give feedback.
-
|
ICYMI: The developer community call has been announced. Sign up to join the discussion! #252 |
Beta Was this translation helpful? Give feedback.
-
|
With the introduction of ACP-236 (continuous staking), stake and delegation will be much more sticky. This will have a multiplicative effect on the importance of the delegation-to-self-stake cap. I do think 4x is quite low but I also agree that 24x seems somewhat excessive. A compromise here would be a satisfactory solution IMHO. In addition it is quite clear to me that delegators have little to no governance interest (even among validators I would claim this is true unfortunately). So, another set of changes that could help mitigate the "too much (voting) power for too little skin in the game", which does exist and happen in other chains, is to simply not count delegation weight for consensus. As far as I know this was never done before but I would like to hear arguments for/against it as I don't see an inherent risk. Together with the fact that running Avalanche validators are among the easiest, lightest and most stable validator clients around (talking from experience here), I would argue that delegators that want to participate in consensus/governance should run their own validators. I'm also in favor of raising the minimum delegation fee from 2% to something more like 10%. This is a rush to the bottom and with the current Avalanche prices this is one of the main tools to make it economically feasible. As a bonus this is another incentive for delegators to become validators. I know this might sound controversial (taking away "vote") but in reality is that with high probability 99% of delegators have no idea and do not care about vote-power. |
Beta Was this translation helpful? Give feedback.
-
|
Following up from the ACP-call, it would also be possible to "discount" the delegation weight. Instead of full-or-zero but a percentage. |
Beta Was this translation helpful? Give feedback.
-
|
A delegation weight reduction proportional to the increase to the 6x increase in the delegation multiplier (6x increase as we're going from from 4x to 24x) would be a great place to start to essentially make no net difference for the potential to game consensus weight and perform a 5.2% liveness attack. Delegation weight would therefore need to be reduced to 1/6 of the current weight. So a 6k delegation stake would have as much consensus weight as 1k AVAX validator self stake. This would also be a first step towards a future ACP where delegation consensus weight is reduced to 0 and the delegation multiplier completely unbounded. Then a validator could attract as much delegation as they like, at no risk to the network. |
Beta Was this translation helpful? Give feedback.
-
|
Following-up on the ACP Community Call, the first thing that I would do is decoupling the max stake limit from the delegation multiplier, by removing the former from the ACP and proposing it in another ACP to submit at a later stage. As we discussed, increasing the multiplier alone should not create any issue, and it can flip the narrative for a lot of validators who are either leaving the network or are not properly incentivized to run a business out of their operation. |
Beta Was this translation helpful? Give feedback.
-
|
As a complementary idea to ACP-247, I believe it would be beneficial to explore an optional, cycle-based lock-up reward boost for validators and delegators. Allowing participants who voluntarily commit to not unstake for a defined number of cycles to earn a higher APR would naturally extend average staking durations and make the validator set more predictable and stable. This approach is fully incentive-based, non-coercive, and non-punitive, and it encourages long-term participation while strengthening the network’s operational sustainability. |
Beta Was this translation helpful? Give feedback.
-
|
Yes, this is a good idea of course, but it needs to be paired with and included in a broader restructuring of how staking APR changes based on a number of factors. I’m currently analyzing various proposals and trying to understand how to break down the roadmap to have a step-by-step changes timeline that’s highly actionable. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
This is a discussion for the proposed ACP-247, I'll rewrite the summary below.
ACP PR Link: #247
Proposed Changes
Why This Matters
Real validator distribution data (Oct 28, 2025) reveals a critical inefficiency:
Root Cause: Current 4x multiplier creates marginal profitability (~$180/month net) that makes validator operations uneconomical, especially for smaller validators.
Solution Impact
At current conditions ($20 AVAX, 8.25% APY, 5% delegation fees):
Network Benefits
Activates Dormant Validators: 451 validators with zero delegations become economically viable (+$3,300 annual profit each = ~$1.5M network-wide wealth creation)
Enables Single-Node Operations: Validators no longer need multiple nodes to achieve profitability
Improves Capital Efficiency: Better utilization of validator infrastructure capacity
Prevents Centralization: 1M AVAX cap (0.14% per validator) is more conservative than current 3M cap (0.42%), with real data showing no validators exceed 2.5% currently
Maintains Backwards Compatibility: Existing validators, delegations, and reward calculations unaffected
Considerations
Validation
Discussion Points
This proposal prioritizes validator capital efficiency and network decentralization over maximum accessibility. It's complementary to (but separate from) proposals to lower validator entry barriers or improve geographic diversity.
Beta Was this translation helpful? Give feedback.
All reactions