I kinda left you with a pretty disappointing ballpark number of 2% and I wanted to make sure I made my point clear. I refer to 2% commission. This is realistically what you can leave on the table for the stakers to get from the delegators. The reason is that if you make staking too lucrative there will be plenty of service providers convincing large stakeholders get into the game via staking while those service operator will take care of the technical part so that the staker doesn’t feel the difference between delegating and staking.
It’s clear that there is a business to be exploited. Markets are the mechanism to efficiently attribute resources. In the end there are not many different implementations out there so that it’s relatively easy for someone who already run e.g. any substrate chain to just spin up another one for someone else. In some jurisdictions staking costs income tax and registering a business while delegating is just tax free dividends which is why most large stakeholders prefer to delegate but if the incentives are high enough there are always ways around those obstacles.
When is staking to lucrative? That’s probably every time you try to fix the staking rewards on a network level with a fixed percentage. Most networks who employ this do so on a quest to limit the downward spiral of commissions that sets in if you make this a variable and thus choose a value that is relatively high like 20%.
The usual way is to make the commission configurable. Sometimes with a minimum. Now the small node operators have the choice like everyone else to delegate or to stake. Staking in the long run only makes sense if they make more money than from delegating after all costs. There is a cost to being on alert 24/7 but apart from that there is a fixed cost for the hardware that the node is running on. That also means the node operators try to rent the cheapest server that is still capable of running the node software to get a higher return, but since everyone is doing the same it actually becomes a requirement to be able to reduce commissions in lockstep with everyone else while not operating at a loss. Most of the time the commissions go down to the lowest that the network allows because there is a demand for delegations. Stakes make temporary marketing campaigns to attract delegations by reducing commission to 0%. In the long run this might not be sustainable but since in most networks you can only successfully run a node long term if you are one of the biggest nodes it makes sense if you are serious to invest into delegations, even by paying your delegators something on top of 0% commissions. In some networks it is possible to go negative. The bigger you are the more organic delegations you will receive down the line.
Obviously a race to the bottom is something for large players because small node operators can not sustain to accumulate losses for long but they get increasingly invisible to delegators at the tail of the active set. If you drop out of the top 100 what’s your realistic appeal to a delegator who tries to make an informed decision about where to put his delegation in the best interest of himself or the network? So many networks make a cut in the 60-100 validators range and the battle at the edge makes sure that a large entity can relatively easy establish new validators periodically and make one smaller guy drop out who suffers from revocations and never comes back.
There are mixed life forms thinkable. Let’s say you have a fixed commission and a whale comes to the rescue and helps by delegating to the small guy to stay active, but he finds the chat group of the small guy and tells him he wants a share in the staking rewards. After a few iterations the small guy might be still in the active set but he has now many whale friends who regularly threaten to revoke on him and slowly eat up all his extra profit.
If you don’t have a limit on the number of nodes in your network after posting some tweets and growing a community of 20 people in a telegram group the small operator has very few options to become visible to additional delegators apart from lowering the commission to post higher APY, but what if he doesn’t want to operate at a loss? The only reliable source of delegations are staking pools. They aggregate the customer for the operator comparable to amazon but they also enforce rules in order to build their own brand. Often the ask you for very low commissions. Then they airdrop governance token to delegators to raise their APY above what any node operator could provide organically. Then They go on an aggressive marketing spree, usually with the help of lending platforms that allow those liquid staking tokens as collateral for loans. Then they make the node operators buy the governance tokens. The deal is that with $x worth of governance token you direct more delegations to your node than by buying the $x worth of the staked token directly. Leverage. You kinda confiscate your future staking rewards by buying those governance token from the delegators and give them higher APY, but at least you are cash flow positive and not loose money on operation every single day. This is a much better proposition that allows to dream of future success.
Back to the reason why someone would stake instead of delegating. 1) he doesn’t pay commission. Those 2% that he would need to pay someone else ideally this should cover the operating expenses. 2) he has the option to make more money from commission he get from others.
Avalanche has made some pretty good design decisions in that regard. But also bas ones. First of all the 2% minimum commission seems to be the standard now. But it’s not 0% and no one will run a node just to save those 2%. Also a node can only carry 4 times the own stake in delegations. This means you can invest into governance token to fill up your delegation pool, but you need to also increase your own stake in native coin in lockstep. Another clever thing that Avalanche did is to allow downtime of 25% without (much) penalty. This is because the cheapest hosting solution is to run the nodes at home. And the biggest downside of running at home is decreased reliability and bandwidth.
Let’s talk about bandwidth. The avalanche node is not light on bandwidth. You frequently fill 25 MBit/s up and down, which is already a problem. In Georgia this is the limit from where the providers will start canceling your subscription because you use the product too much. In Belgium that’s the fastest business line you can buy. The problem is that most of this bandwidth is international bandwidth and not local bandwidth. That means it’s cost to the provider and the resources are limited so that he can not sell them to higher paying business customers when residential lines use it up.
The main reason seems to be gossip traffic. The more nodes you have in the network the more bandwidth you need.
If you want to find the sweet spot for the number of nodes and profitability there is a simple logic you can follow:
You have 5% inflation per year. Let’s say that is $100 million. Now you have 2% commission so lets just assume that $2 million is the sum that has to pay for the operation of all the validators. You like to target a number of 2000 validators somehow which means $1000 per year per validator in commission income if they are all equally sized. That’s $80 each per month that has to cover everything. If you host at home you can say that $40 go to pay for the internet and $10 for electricity. Let’s assume you need to buy a few parts for the computer that you run your node on and you want to earn that back in 2 years then the machine can cost maximum $720 before you loose money as a operator compared to simply delegating.
And that assumes you have a network with $2 billion FDV. While you are bootstrapping those numbers can look a bit different. E.g. only 50% of the total supply participates in staking so the 5% inflation are airdropped to less stake and the APY is therefore higher. That’s the baseline. Starting from there people will try to increase their rewards by modifying the client e.g. to participate in pools that pay MEV income or to get a little bit more performance and get higher rewards if the network values that. Staking pools might force people to use those modified clients either directly as in MEV pools come with a staking pool as bundle, or indirectly by establishing performance criteria that can realistically only be met by patching the client software.
Another road to revenue is merge mining. Validators have an incentive to run on every subnet that whitelists them basically at no additional cost. But those subnets so far have their own ideas about who should be able to validate because they want to sell their token and create entry barriers. It would be perfect if those could be permission less and only require to have a validator on main net to generate some extra income, no matter how small. It’s extra dividends.
Additionally into that category I’d drop plugins that entitle you to collect extra rewards, e.g. by offering storage to apps like eigenlayer, sell resources to networks like graph protocol that indexes on chain data etc. I think running a validator on the base layer increasingly doesn’t need to generate a direct profit but may entitle you to participate in other opportunities down the line that validators make money from. If you run block explorers, also think about other services developers need to deploy apps to their new chains and how you can deliver such services decentralized. For instance it should be possible to make a block explorer based on a subgraph that is hosted on the validator nodes themselves without any centralized infrastructure but some gateways and then pay the validators something extra for answering queries to users of the block explorer.
Staking pools are a good way to enforce behavior. While it’s difficult to e.g. have a geo multiplier in the protocol level it’s pretty doable to fund a staking pool with foundation money and let a bot delegate according to criteria that you want to incentivize. Fill the small nodes in unique locations with good enough performance up to their maximum in delegation capacity and you get people to run the nodes from home. Don’t focus too much on network parameters but make sure that the inflation rewards support the infrastructure cost of your target number of nodes. Whales games every network so far. The outcome seems to be similar in every case. That’s why 2% is my magic number.