What Staking Is, What It Pays and What Can Go Wrong

Almost every explanation of staking opens the same way: "it's like a savings account that pays interest." The comparison is comfortable and it breaks on all three points that matter. There is no debtor. The percentage is neither guaranteed nor contractual. And you are paid in the same asset whose price can drop 44% in a year, which is exactly what ETH did in the twelve months to August 2026.

What does exist is a specific, measurable mechanism with public rules. A validator performs work the network needs, the network pays for that work with new issuance and with tips from people transacting, and if the validator cheats it loses part of the capital it posted as a bond. Understanding that sentence in full — especially the word "bond" — is understanding staking.

What a validator does, and why the network pays for it

Since The Merge on 15 September 2022, Ethereum has run on proof of stake. There are no longer miners burning electricity to propose blocks; there are validators who have deposited ETH and who, in exchange, get to vote on which chain is the real one. If you want the comparison with the older model, it is in proof of work versus proof of stake.

The actual job is dull and relentless. Every 12 seconds (a slot) the network picks one validator to propose a block and assigns a committee the task of attesting: signing a message that says "I have seen this block and it looks like the valid head of the chain." A validator spends nearly all its life attesting. Proposing comes around a handful of times a year.

The network pays for this because it needs thousands of independent machines switched on and in agreement, and it needs lying to be expensive. That is where the deposit comes in: locked ETH guarantees that whoever signs something contradictory has something to lose. Under proof of work the cost of attacking is electricity. Under proof of stake it is capital the network itself can destroy.

Where the 32 ETH figure comes from

The number is an engineering compromise. Every validator emits signatures that all nodes must receive, verify and aggregate, so consensus cost scales with the number of validators, not with the ETH deposited. A 1 ETH minimum would have multiplied that traffic by 32. Raising the threshold, on the other hand, centralises. Even 32 ETH is a serious barrier: at the roughly $2,505 price of August 2026, around $80,000 immobilised.

Pectra (7 May 2025) introduced consolidation: an operator can merge validators up to 2,048 ETH, which cuts signatures without cutting the ETH at stake. The activation minimum is still 32.

The numbers, as of 28 August 2026

Source: validatorqueue.com, 28 August 2026.

Two readings are compulsory. The first: the APR falls as more ETH is deposited, because issuance is shared among more validators. A 2.67% rate with 34.79% of supply locked is not a rate anybody decided. It is the arithmetic result of a public formula applied to how many people have joined.

The queue asymmetry, and why it inverts

Getting in currently takes over a month. Getting out takes barely an hour plus the sweep that actually returns your balance, roughly 7.8 days. That describes a market where a lot of people want to stake and almost nobody wants to unwind.

The mechanism itself is symmetric: the churn rate caps activations and exits at the same pace, 256 validators per epoch. What is not symmetric is demand. In a stress episode — a sharp price drop, cascading liquidations across DeFi protocols, a regulatory scare — the queue that fills is the exit queue, and the wait to get your ETH back goes from an hour to weeks. Precisely when you would want it available.

Slashing: what really triggers it, and what does not

Slashing is the severe penalty: the network confiscates part of the deposit and ejects the validator. Three behaviours trigger it, and all three are contradictory signatures: proposing two different blocks for the same slot, issuing two attestations that conflict within the same epoch (a double vote), or issuing one that "surrounds" an earlier one (a surround vote).

All three only happen if the software is misconfigured or if someone is deliberately attacking the chain. The most common practical cause is not malice. It is starting the same validator key on two machines at once, typically during a migration performed "to be safe."

What does not cause slashing is being switched off. If your validator goes down you stop earning and accrue inactivity penalties, small and proportional to the time offline. An offline validator bleeds slowly; one that signs twice loses a lump.

• Do not copy validator keys onto a second server "just in case": that is the exact recipe for slashing.
• Do not confuse inactivity penalties with slashing: the first stops the moment you are back online.
• Do not assume a provider with "no slashing history" will never have one; ask whether it covers that loss contractually and up to what limit.

The four ways to stake

With solo staking you run everything: nobody charges a fee and nobody covers your mistakes. You need an execution client and a consensus client, plus the discipline to update ahead of every hard fork (Glamsterdam is targeted for 4 November 2026). In a pool, keys are split among several operators and none can sign alone, at the cost of an extra smart contract layer. On an exchange you tick a box: the most convenient option and the one that most resembles a deposit, with no control over the keys. Coinbase, for instance, charges nothing to stake or unstake but withholds a commission on rewards that varies by asset. If that is your route, read how to choose a safe exchange first.

Liquid staking: the fix that creates another problem

You deposit ETH and receive a token representing your position, which you can move, sell or post as collateral while the underlying ETH stays locked. It solves something real — the capital stops being dead — and in exchange it relocates the risk to three new places.

The first is the smart contract. Your ETH no longer sits only in Ethereum's deposit contract; it sits behind a protocol's logic. Audited, heavily used logic, but still code.

The second is concentration. As of 28 August 2026 Lido is the largest DeFi protocol in the world by TVL at $24.20B, ahead of Aave V3 ($17.46B) and SSV Network ($12.57B). When that share of validating ETH runs through a single set of governance rules, the uncomfortable question is not whether the protocol is honest. It is what happens if one bug or one vote simultaneously affects a meaningful percentage of the network's validators.

The third is subtler: the liquidity of a liquid staking token is secondary-market liquidity, not protocol liquidity. When you want out fast you are not asking the network for your ETH, you are selling to another person. If everyone wants the same thing at once, the price detaches from parity. The exit queue puts a theoretical floor under that divergence, but it takes days or weeks to materialise.

What the "yield" actually is

The APR has three components, and only the first is stable.

1. Issuance. New ETH the network creates to pay for consensus. This is the predictable part, and it declines as staked ETH rises.
2. Priority fees (tips). Since EIP-1559 (5 August 2021), every transaction pays an algorithmic base fee that is burned plus a tip that goes to the block proposer. With gas at 0.151 gwei in August 2026 these contribute very little; in a congested market, a great deal. We unpack this in what gas fees are.
3. MEV. The value extractable by reordering, including or excluding transactions within a block. Validators typically auction the right to build the block and keep a share. This is the most volatile component.

None of that is interest. There is no counterparty obliged to pay you, no schedule and no guaranteed principal. It is variable compensation for a service, paid in the same asset you contributed.

Worked example: $5,000 in liquid staking for a year

Suppose you convert $5,000 into ETH and, so the arithmetic stays clean, that you end up with 2 ETH (the exact figure depends on the price on the day; ETH was trading around $2,505 on 28 August 2026). You deposit them into a liquid staking protocol charging 10% of rewards, like Lido.

• Gross reward: 2 ETH × 2.67% = 0.0534 ETH per year.
• Protocol fee: 10% of 0.0534 = 0.00534 ETH.
• Net reward: 0.04806 ETH, that is 2.40% net in ETH terms.
• Final balance: 2.04806 ETH.

You have 2.40% more ETH. Now the part that does not make the marketing page: what is that worth in your own currency?

• If the price does not move: $5,000 × 1.0240 = $5,120. Gain: $120.
• If ETH repeats the year it just had (−44.2% year on year to August 2026): $5,000 × 1.0240 × 0.558 = $2,857. Loss: $2,143.

The staking reward is the thin line in the calculation; the price is the whole chart. A 2.40% annual yield offsets a 2.4% price fall and not one cent more. Anyone selling this as "passive income" is describing the numerator and hiding the denominator.

Tax: the concept, not the form

Tax treatment of staking rewards is set nationally, and the differences are not cosmetic. Broadly, tax authorities have taken one of two approaches, and you need to know which one applies to you before you start rather than after.

The first approach taxes rewards as income when you gain control of them, valued at market price on the day of receipt. Under that model, each reward creates a taxable event even though you never sold anything, and it also sets the cost basis for the tokens: when you eventually sell, the difference between that recorded value and the sale price is a separate capital gain or loss. The second approach defers everything to disposal, treating rewards as having a zero or acquisition-linked basis and taxing only when you sell.

Two practical points hold in either case. Rebasing liquid staking tokens — which grow your balance without any identifiable payment on a specific date — are an unresolved edge case in several jurisdictions, and some tax authorities have published nothing on them at all. And most regimes that tax receipt do not tell you which price source to use for the conversion, which means your own consistent, documented method is the best defence you have.

Checklist before locking a single unit

• Check today's entry and exit queues, not the ones you read about months ago.
• Work out the APR net of fees, not the headline gross: at a 10% cut, 2.67% becomes 2.40%.
• Decide how long you can have that capital locked without needing it, then double the number.
• With liquid staking, assume your fast exit is a secondary market, not a redemption right.
• With an exchange, check it holds the licence required in your jurisdiction before depositing.
• Record the local-currency value of every reward on the day you receive it; you will need it at tax time.
• If you run your own validator, have a key migration procedure that makes two simultaneous instances impossible.

Where things stand in 2026

In the European Union, MiCA now governs who may offer these services. Spain's transitional period ended on 1 July 2026, after which only firms authorised by the CNMV or passported from another member state may operate there; equivalent deadlines applied elsewhere in the bloc. Staking offered by a platform is a third-party service, and the first sensible question is no longer how much it pays but who provides it and under what licence.

Across the Atlantic, the joint SEC and CFTC interpretation of 17 March 2026 published a taxonomy of crypto assets and expressly addressed protocol staking, mining and wrapping, recognising that most crypto assets are not themselves securities.

Regulation changes the wrapper, not the mechanism. The validator does the same work, the network pays the same, and the price of ETH remains the variable that decides whether your year was good. If you plan to trade that volatility actively, it is worth understanding the basics of trading first.

Frequently asked questions

Sources and references

1. Validator Queue — Ethereum entry and exit queues
2. Ethereum.org — Staking
3. Lido — Protocol Fee
4. DefiLlama — Lido TVL
5. Coinbase Help — Earn rewards with staking
6. CNMV — Crypto-asset regulation (MiCA)
7. SEC — Joint SEC/CFTC clarification on the application of federal securities laws to crypto assets (17 Mar 2026)
8. CFTC — Joint press release with the SEC (March 2026)

Related guides