How to Choose a Solana Validator: A 10-Minute Checklist
Picking a validator is the one decision in Solana staking that you actually control. Everything else — the issuance rate, the epoch clock, the price of SOL — is handed to you by the network. Your validator determines what share of the rewards reaches your stake account, how reliably those rewards arrive, and whether your stake makes the network more decentralised or less. This guide is the checklist we would hand a friend: what to measure, what good looks like, what should make you walk away, and how to verify all of it yourself in about ten minutes. If you have not staked before, start with how to stake Solana and come back here for the selection step.
Key takeaways
- “0% commission” is only half a fee schedule. Inflation commission and MEV commission are set separately. A validator can charge 0% on staking rewards and still keep 100% of MEV tips.
- Performance beats size. Skip rate, vote credits, and delinquency history predict your yield. Total stake does not.
- Delegating to an already-huge validator costs you nothing and costs the network something. Stake concentrated in the superminority is the one systemic risk delegators can actually influence.
- Upgrade discipline is a real differentiator. Validators that lag on client releases miss rewards and occasionally miss consensus changes entirely.
- Every claim in this article is verifiable on Stakewiz, validators.app, Solana Beach, and SolanaFM. Including ours. Check, do not trust.
What does “commission” actually mean on Solana?
There are two commissions, and they are unrelated to each other.
Inflation commission is the validator’s cut of protocol staking rewards. The protocol mints new SOL each epoch, distributes it to stake accounts according to stake weight and validator performance, and the validator keeps the percentage it has set on its vote account. This is the number wallets display, and it is the number most validators advertise.
MEV commission applies only to validators running the Jito client. Traders pay tips for favourable transaction ordering, those tips are shared with delegators, and the validator takes a cut expressed in basis points: 10,000 bps is 100%, 1,000 bps is 10%, 0 bps is nothing. Because it lives in a separate configuration from the inflation commission, the two can diverge wildly. A validator running a “0% commission” banner while set to 10,000 bps of MEV commission is not lying, exactly. It is just showing you one of two prices.
So the first question is never “what is the commission?” It is “what are both commissions, and how have they moved over the past year?” Commission history matters as much as the current value. A validator that has sat at 0% for two years is making a different statement than one that dropped to 0% last month, harvested delegations, and has a track record of raising it back.
At the time of writing, Stakewiz showed the simple network average commission at roughly 15.4%, which sounds alarming until you remember that average includes private and closed validators that set 100% and never intended to take outside delegations. The distribution matters more than the mean. Most validators competing for public stake sit far below it.
Is a 0% commission validator too good to be true?
Not necessarily, but it deserves a direct question, and honest operators should give you a direct answer.
Running a validator costs money continuously. The largest recurring line item is not servers, it is vote transaction fees: validators submit vote transactions constantly and pay for them out of their own balance, every epoch, whether or not they earn anything. At 0% inflation commission and 0 bps MEV commission, a validator has no protocol revenue at all and is paying those fees from somewhere else.
That “somewhere else” is the thing to ask about. Legitimate answers exist: a foundation or grant delegation, revenue from other networks the team operates on, an explicit and time-boxed growth campaign, or a business that uses the validator as infrastructure rather than as a profit centre. An operator who cannot or will not explain the funding is telling you that the rate is a marketing number, and marketing numbers change.
Our own 0% is a campaign rate, and we say so. That is the honest framing: it is not a permanent promise, and any validator presenting a subsidised rate as forever is overselling.
How do you read validator performance?
Three public metrics do most of the work.
Skip rate is the share of a validator’s assigned leader slots where it failed to produce a block. Skipped slots are rewards that never existed, for the validator and for you. At the time of writing, Stakewiz showed the network average skip rate at roughly 1.7%. Well-run validators sit meaningfully below that, and many sit at or near zero over a recent window. Read it over multiple epochs, not one: a single bad epoch during a network incident tells you nothing.
Vote credits measure how consistently and how promptly a validator voted on blocks. Because staking rewards are proportional to vote credits earned, a validator in the top decile of credit performance is mechanically paying its delegators more than a mid-pack one at the same commission. Stakewiz and validators.app both rank this.
Uptime and delinquency history are the tail risk. A validator is marked delinquent when it stops voting for long enough that the cluster notices. Everyone has an incident eventually; hardware fails, data centres lose power, releases have regressions. What separates operators is frequency and duration. One delinquency event resolved in minutes is operations. A pattern of multi-hour delinquencies across months is a warning that nobody is on call.
What does good look like, and what is a red flag?
| Metric | What good looks like | Red flag |
|---|---|---|
| Inflation commission | Stable value, disclosed history | Recent drop to 0% with no explanation, or frequent changes |
| MEV commission | Disclosed in bps next to the headline rate | Not shown anywhere, or 10,000 bps behind a “0% commission” banner |
| Skip rate | At or below the network average over many epochs | Persistently above average, or wild epoch-to-epoch swings |
| Vote credits | Consistently in the upper ranks | Mid-pack or falling ranking over time |
| Delinquency history | Rare, short, explained | Repeated multi-hour outages, no public comms |
| Client version | Current stable release, upgraded promptly | Several releases behind, or upgrading only after incidents |
| Stake position | Outside the superminority | Already among the largest validators by stake |
| Operator identity | Named team, public tooling, reachable | Anonymous, no history, no channel to ask questions |
| Jito | Enabled, MEV commission published | Not enabled, or enabled with undisclosed terms |
Why does stake concentration matter to me?
Because the network you are being paid to secure has a failure mode, and delegators collectively decide how close it gets to it.
The superminority is the smallest group of validators that together control more than one third of active stake. That threshold matters because a third of stake is enough to prevent the network from finalising blocks. The smaller that group, the fewer independent parties it takes to halt Solana, whether through coordination, coercion, or a shared dependency failing at the same time in the same cloud region.
Adding your stake to a validator that is already among the largest moves the network in the wrong direction. Adding it to a competent validator outside that group moves it back. The concrete cost to you of choosing an independent validator is zero: rewards are proportional to your own stake and the validator’s performance, never to how much total stake it holds. A 500 SOL delegation earns the same rate at a small validator with a 0% skip rate as at a giant one with the same commission and performance.
Stakewiz and Solana Beach both flag superminority membership directly on the validator page. It is a one-glance check.
Does the Jito client matter?
For most delegators, yes, in a mildly positive direction. Validators running Jito capture MEV tips and pass a share to delegators, which shows up as an additional component of total APY on top of the inflation-based staking APY. The size of that component varies with market activity and is usually small relative to base staking rewards, so treat it as a bonus rather than the deciding factor.
What matters more than “Jito or not” is disclosure. A validator running Jito and publishing its MEV commission in bps is giving you a complete fee schedule. A validator running Jito and staying quiet about the bps figure is not.
How much does client version and upgrade discipline matter?
More than most delegators realise. Solana ships frequently, and releases carry performance improvements, consensus-relevant changes, and occasionally urgent fixes. A validator several releases behind is usually leaving rewards on the table through worse block production, and in the worst case risks falling out of consensus after a network-wide change.
You cannot audit an operator’s runbook, but the version string is public on every explorer, and it is a decent proxy. Check it, then check it again in a few weeks. An operator running the current stable release both times is doing the boring work. Related signals: does the team publish tooling, monitoring, or failover kits publicly? Do they post about incidents? Public engineering artifacts are weak evidence individually and strong evidence in aggregate.
Own hardware or rented?
This one is genuinely nuanced, and anyone who tells you there is a single right answer is selling something.
Bare metal you own gives full control of the hardware lifecycle, no noisy-neighbour effects, and no dependence on a provider’s policy decisions. It also means you are responsible for physical redundancy, and a single-site operator with no failover is fragile no matter who owns the box.
Rented bare metal or cloud buys you rapid replacement, geographic flexibility, and someone else’s on-site staff. The cost is correlation risk: if a large share of Solana’s stake sits with the same few providers in the same few regions, a single provider outage becomes a network event.
What you actually want to know is not the ownership model but whether the operator has thought about failure domains: is there a tested failover path, is the setup in a region that is not already saturated with stake, and can they swap identity to a hot spare without a long outage? Ask. The answer, or the absence of one, is the signal.
What should I check about the people, not the machine?
- Public identity. A named team with a website, a GitHub organisation, and working contact channels can be held accountable. Anonymous operators are not automatically bad, but you are trading away recourse.
- Track record and length of operation. Multi-year operating history across market cycles and network incidents is the hardest thing to fake.
- Governance participation. Validators are not just block producers; they vote and they participate in network discussion. An operator engaged in governance is one with a stake in the network’s direction beyond fee income.
- Communication during incidents. The best predictor of how an operator will behave during your next bad epoch is how they behaved during the last one.
The 10-minute checklist
Open the validator’s vote account on Stakewiz and work down this list.
- Inflation commission and its history. Stable? Explained?
- MEV commission in bps. Published? What is the actual number?
- Skip rate over the last several epochs against the network average.
- Vote credit ranking. Upper ranks or mid-pack?
- Uptime and delinquency history. Frequency and duration of past incidents.
- Superminority flag. Inside or outside?
- Client version. Current stable, or lagging?
- Jito enabled, with terms disclosed.
- Operator identity. Named team, public tooling, reachable contact.
- Cross-check on a second explorer. validators.app for performance rankings and commission history, Solana Beach for stake distribution and superminority context, SolanaFM for raw account activity.
If a validator passes all ten, the difference between it and the next candidate that also passes all ten is noise. Pick the one that is smaller, or the one whose team you can actually reach.
How we score on the same checklist
We run this validator, so treat what follows as a claim to verify rather than a conclusion. Every number below is a snapshot taken on 2026-07-02; live values are on our Stakewiz profile and will have moved since.
At that snapshot: inflation commission 0%, MEV commission 0 bps, total APY 5.62% (5.51% staking, 0.11% Jito) against a network average of 4.24%, uptime 99.76%, skip rate 0%, Jito enabled, client version 0.1004.0-rc.40101, around 103,044 SOL of activated stake. We are outside the superminority by a wide margin, we have operated proof-of-stake infrastructure since 2018, we publish our Solana tooling on GitHub, and the 0% rate is a campaign rate funded by our other operations, not a permanent guarantee.
The honest caveat: that stake figure is a fraction of what it was thirty epochs earlier, because delegation flows move constantly and foundation programmes rotate. Stake trend is worth watching on any validator, ours included, but read it as a signal about delegation programmes rather than about performance.
Do not take our word for any of it. Paste the vote account ENVaKoD7ytn58xJ8s5htFfQ8hqQt1G9dcPUDqbSwVcgB into Stakewiz, validators.app, Solana Beach, or SolanaFM and read the same data from a source with no incentive to flatter us. Our homepage dashboard pulls from the Stakewiz API, which makes it convenient, not independent.
FAQ
What is a good commission for a Solana validator?
Anything from 0% to about 8% is defensible if performance is good. Stability matters more than the number: a validator that has held a rate for a long time is more predictable than one that just cut to 0%. And always read the MEV commission alongside it, because a 0% inflation commission tells you nothing about what happens to MEV tips.
Is a 0% commission Solana validator actually worth choosing?
It can be, but ask how it is funded. Vote transaction fees are paid out of the validator’s own pocket every epoch, so a 0% operator is running at a loss on protocol revenue and covering it from somewhere else. A grant, a foundation delegation, another business line, or an explicit time-boxed campaign are all reasonable answers. No answer at all is the problem.
What is MEV commission and why does it matter?
It is the validator’s cut of Jito MEV tips, set separately from the inflation commission and expressed in basis points, where 10,000 bps is 100%. Because the two fees are configured independently, checking only the headline commission shows you half the fee schedule.
Does delegating to a small validator lower my rewards?
No. Your rewards scale with your own stake and with the validator’s vote credit performance, not with the validator’s total stake. A small validator with a near-zero skip rate and the same commission pays the same rate per SOL as a large one, while doing less damage to stake concentration.
How do I check whether a Solana validator is reliable?
Look up the vote account on Stakewiz, validators.app, Solana Beach, and SolanaFM, and compare skip rate, vote credits, uptime, delinquency history, commission history, client version, and superminority status. It is all public on-chain data. A validator’s own dashboard is a starting point, never the verification.
Before you delegate
Two structural reminders worth repeating. Delegation is non-custodial: your SOL sits in a stake account only your wallet controls, the validator receives voting power and never touches your funds, and you can redelegate or deactivate at any time without asking permission. And as of mid-2026 Solana does not automatically slash delegated stake for validator misbehaviour, which means the realistic cost of a bad choice is missed rewards rather than lost principal. That rule could change, so re-check it rather than treating it as permanent.
Everything else settles at epoch boundaries, roughly every two days: stake activates, rewards are paid by the protocol and compound into your stake account, and deactivation cools down the same way. Run the ten-minute checklist once, and the next few years take care of themselves. More questions are answered in the FAQ on our homepage, and if you want to see whether we hold up under the same scrutiny, the Stakewiz profile is the place to start, not this page.