Ethereum staking hit a record 34%. Read why yields fell.

Ethereum just locked up more of its own supply than at any point in its history, and the pay for doing it fell to a three-year low. About 34% of all ETH, roughly 41.4 million coins, now sits in staking contracts as of early August 2026, up from around 29% at the start of the year. Most coverage reads that as conviction. The number under it reads as a squeeze.

Both are true at once, and the reason is mechanical. Ethereum pays stakers out of a fixed issuance schedule that scales with the square root of the total staked, so the yield per validator drops as more validators pile in. More people share the same slow-growing pool, and each slice gets thinner. The 7-day staking APR is now 2.66%, down from a 5.06% peak in June 2023, close to a halving in three years. Nothing broke. The design did what it was built to do. And a proposal filed on August 4 would push the yield down further on purpose.

The record is the reason the yield fell

Read the two headlines together and the story flips. "Staking hits an all-time high" and "validator rewards hit a three-year low" are not a contradiction. The first causes the second.

Every ETH that joins the validator set dilutes the reward for every ETH already there. The protocol mints a roughly fixed amount of new ETH per year for consensus, growing only with the square root of the staked total, and splits it across everyone staking. So the yield is not a rate the network promises. It is a dividend that shrinks each time someone new claims a share of it.

The staked total climbed from about 29% of supply in January to 34% by early August, and the APR fell in lockstep. Validators who entered in June 2023 earned 5.06%. Validators entering now earn 2.66% on the same work, a real decline of roughly 47%. Same duty, half the pay, because the crowd doubled.

This is the part that gets lost when staking is sold as a yield product. Traditional yield rises when demand for it rises. Staking yield does the opposite. The more capital chases it, the less it pays. It behaves less like a bond coupon and more like a room that gets more crowded every time someone hears there is money in it.

The square root nobody prices in

The exact shape matters, because it tells you where this goes. Consensus issuance tracks the square root of total staked ETH, which means the yield per staker tracks one over that square root.

Work the math forward. If the staked total doubles from here, aggregate issuance rises only by about 41%, so the yield per validator falls by roughly 30%. A 2.66% APR becomes something closer to 1.9% with no change in participation quality, no bug, no policy shift. Just more coins in the pool. The curve is convex against the staker: each new wave of stakers pushes the yield down by less in absolute terms, but it never turns around.

There is no built-in stopping point either. The current design has no staking ratio at which the reward for adding more ETH switches off. In theory the network could drift toward half its supply staked, or more, with the yield grinding lower the whole way. That open-ended drift is exactly what a group of Ethereum researchers decided to close.

EIP-8361 would turn the drift into policy

On August 4, six authors including Ethereum Foundation researcher Justin Drake published EIP-8361, a "tapered issuance burn." It burns a growing fraction of validator rewards as the staking ratio rises, and the burn reaches 100% at roughly 60.25 million ETH staked, about half of today's supply. Past that point, net new issuance for stakers goes to zero.

At today's participation, the proposal would cut consensus yield from about 2.6% to roughly 1.2%, phased in over 18 months to avoid a sudden rush for the exit. The point is not to punish stakers. It is to remove the one thing the current design lacks: a level at which the incentive to stake more finally turns off. Right now there is none.

Two caveats keep this honest. The proposal carries Draft status and has not been slated for any upgrade, so it may never ship in this form. And it has opposition. Liquid staking and DeFi operators are pushing back, since a lower base yield cuts straight into their product. But the direction of travel is set by the math whether or not the EIP passes. The square root is already doing quietly what EIP-8361 would do openly.

Who actually eats this

The people most exposed are the ones who sold the yield as the reason to hold. Two groups stand out.

First, the ETH treasury companies. Firms like Bitmine and Sharplink raised capital to buy ETH and stake it, pitching the staking return as the engine. Bitmine alone holds about 4.9 million ETH staked, near 12% of all staked ETH. Under EIP-8361, modeling shows the revenue on that position roughly cut in half, and worse as the ratio climbs toward 50%. When part of that stake is funded by preferred stock carrying a fixed dividend, a shrinking yield turns a spread into a gap.

Second, anyone holding a staking ETF or a liquid staking token for the payout. The staking ETFs that promise to stake on your behalf skim a management fee off a gross yield that is already at a three-year low and structurally headed lower. Net of that fee, and net of ETH's own issuance, the real return you keep is thinner than the 2.66% headline suggests, in the same way a bitcoin income ETF's advertised yield sits well above what lands in your account.

The concentration adds a second risk on top of the yield one. With one entity controlling near 12% of staked ETH, a forced unwind, from a margin call or a dividend it can no longer cover, would push a wall of exits into a queue that already takes weeks to clear. The yield squeeze and the concentration are the same story from two angles: a lot of ETH crowded into one trade that pays less every month.

What to watch

Three things tell you where this lands. Watch the staking ratio itself: every point higher is a point lower on the yield, EIP or not. Watch EIP-8361's status, from Draft to any inclusion in the Glamsterdam upgrade slated for late 2026, because that is when a modeling exercise becomes real money. And watch the treasury firms' filings for how they describe staking revenue, since they will feel the compression first and in public.

The record staking ratio is not a strength signal for the people staking. It is the mechanism draining their yield, and a live proposal wants to finish the job. Conviction and dilution look identical on a chart. They are opposites in your wallet.

Ethereum staking APR

Sources

• The Block: Ethereum staking climbs to 34% as proposal targets validator rewards (Aug 12, 2026)
• Coinpedia: Ethereum staking hits 34% of supply, validator rewards at a three-year low
• Staking Rewards: Ethereum staking analytics and live APY
• The Defiant: EIP-8361 tapered issuance burn would burn validator rewards
• Coingape: Ethereum developers submit EIP-8361 to introduce tapered issuance burn
• CryptoTimes: Why EIP-8361 has DeFi leaders fighting back (Aug 8, 2026)

This is not financial advice.