Running a node
Two different things share the name. One is infrastructure you run for yourself; the other is a business with staff costs and no employees.
"Running a node" describes two activities that have almost nothing in common, and the confusion between them is used to sell things.
A full node: infrastructure, not income
A full node downloads the blockchain, verifies every rule for itself and relays transactions. It earns nothing. Nobody pays you.
What it gives you instead is not needing to trust anyone about the state of the chain. Your wallet asks your own machine rather than somebody else's server, which means better privacy and no dependence on a company's uptime or honesty.
It costs a machine that stays on, a terabyte or so of fast storage, and an internet connection with no strict data cap. Enthusiasts run them on hardware costing less than a phone.
People run full nodes for the same reason people host their own email: independence, and a working understanding of the thing they rely on. Framed as an investment it makes no sense, because it is not one.
A validator: a job with an uptime requirement
A validator participates in producing blocks on a proof-of-stake network and is paid for it. This is a real business, and the parts that get advertised are the pleasant ones.
- CapitalMany networks require a substantial minimum stake before you can validate at all.
- UptimeDowntime costs money directly. Serious operators run redundant infrastructure and monitoring, which is a cost and an ongoing obligation.
- CorrectnessSigning contradictory blocks is punished severely, and a misconfiguration can do it without any bad intent. A common cause is running two copies of the same validator "for redundancy".
- AttentionClient upgrades, hard forks and security patches on somebody else's schedule, including at inconvenient hours.
The return, net of hardware, hosting and your own time, is often not far from what delegating to somebody else's validator would have paid — with substantially more that can go wrong. People do it for control and for the technical interest, and both are decent reasons. "Passive income" is not one, because none of the above is passive.
Delegating gives most of the return with none of the operations
If the goal is the yield rather than the involvement, native delegation pays close to the same and cannot be broken by a failed upgrade at two in the morning. The page on staking covers it.
Where the word gets abused
A whole category of schemes sells "nodes" as products. You buy a licence, an NFT or a subscription, and are promised daily rewards for a node the seller operates on your behalf.
Ask what the node actually does. On a real network, a validator is paid from issuance and fees, and those are public numbers — you can check what the network pays in total and how it is divided. If the promised return has no relationship to that, or if the "network" has no independent existence outside the product being sold, then the rewards are funded by later buyers.
The tell is usually the sales structure rather than the technology: a fixed daily figure, a referral tree, and a price for the licence that only makes sense if more licences are sold. The page on mini-app schemes describes the same arithmetic under a different name.
Real node rewards are public and checkable
Every proof-of-stake network publishes what it pays validators. If a seller's promised return cannot be traced back to those numbers, it is not coming from the network.
If you want to run one
Start with a full node on a network you care about, on hardware you already own. It costs almost nothing, teaches you more about how this works than any article, and removes your dependence on somebody else's server.
Validating is the step after that, and it is a decision about how you want to spend your evenings as much as about money.