Bitcoin yield and what is left of BTCfi

There are two ways to try to make bitcoin pay. The first is to lend it to a company and take their credit risk, which is what every CeFi Bitcoin interest account does and where the honest rates sit between 0.02% and 0.25%. The second is the subject of this page: a protocol layer that takes bitcoin, locks it or wraps it, and uses it as security or collateral somewhere else. That layer has a name — BTCfi — and in 2024 and 2025 it was the most confidently marketed idea in the market.

It has since contracted sharply. Bitcoin DeFi peaked at around $9.1bn in October 2025; Bitcoin layer-2 and sidechain total value locked was down 74% from that peak by the first quarter of 2026, and the ecosystem shrank from 101,721 BTC to 91,332 BTC. Only about 0.46% of circulating bitcoin is in DeFi at all. That contraction is the most useful fact about the sector, and it is the one its own marketing never mentions.

What "Bitcoin staking" actually means

Bitcoin has no staking. It uses proof of work, miners earn the block subsidy by spending electricity, and nothing in the protocol pays a holder for holding. Any product using the word is describing something else, and it is worth being precise about what.

The Babylon design, which is the one most of the sector is built on, works like this. You lock bitcoin on the Bitcoin chain itself, under a timelock, using scripts that make the coins spendable only by you after the lock expires — or by a slashing path if you misbehave. That locked position is then registered as economic security for a separate proof-of-stake network, which pays you in its own token for providing it. Your bitcoin never leaves Bitcoin and never validates anything on Bitcoin. It is collateral posted to a third party's consensus.

That is a genuinely clever construction, and it is meaningfully different from wrapping BTC and bridging it to another chain, because the coins stay under Bitcoin script control. It is also the source of a risk that does not exist when you simply hold bitcoin: you have accepted slashing conditions written by somebody else, for a chain whose governance you do not control. The same objection applies, with more force, to Ethereum restaking — covered onliquid staking.

Most of the products a retail reader will actually encounter sit one or two layers above that. A liquid staking token representing a Babylon position. A wrapped bitcoin on a sidechain. A yield-bearing token representing a basket of the above. Each layer is another contract that has to work.

The sector by the numbers

Put the last figure in perspective. More than 99.5% of all bitcoin is not in DeFi. After several years of building, an enormous amount of venture funding and a period of intense promotion, the entire category has attracted less than half a per cent of the asset it is built on. That is not a rounding error in adoption; it is the answer to the question.

The protocols, and what each does with your bitcoin

The Solv row deserves a second look, because it is the most informative number on the page. Roughly $41 a day of revenue against about $2.15bn of assets is not a business generating yield; it is a business distributing something else. When a protocol's fee income is that small relative to the capital it holds, the returns being paid to depositors are coming from token emissions, from a treasury, or from a partner chain's incentive budget. That is the organic-versus-subsidy test we apply throughoutDeFi yield farming, and BTCfi fails it clearly.

The concentration is also worth noting. Babylon at about $3.12bn and EigenCloud at around $6.3bn together make up roughly 97% of a $9.68bn restaking category. Two protocols are effectively the whole market for pledging an asset as security somewhere else, and one of them is the Ethereum one.

The Kraken reality check

The clearest test of whether BTCfi yield is real is what happens when a large, regulated venue packages it for retail. Kraken did exactly that on 19 June 2025, launching Bitcoin staking built on Babylon.

The resulting published rates are 0.02% flexible and 0.04% bonded, and Kraken states these are pre-commission estimates. Kraken takes 30% of rewards on flexible staking and 25% at the first bonded tier. It also discloses that flexible staking stakes only up to 50% of the balance you enrol. On a 0.02% headline, those deductions are almost too small to express, which is itself the point: at that level nothing about the fee structure matters, because there is nothing to divide.

Set that against the same exchange's other rows — SCRT at 9.64% flexible and 20.19% bonded, Cosmos at 9.24% and 19.32% — and the picture is unambiguous. When a venue with genuine validator infrastructure and an incentive to show attractive numbers puts Bitcoin staking on the same page as proof-of-stake staking, bitcoin comes last by three orders of magnitude. Details of the product sit in our Kraken review.

What the restaking cycle already demonstrated

BTCfi's core proposition — pledge an asset you already hold as security for something else, and collect an extra return — is not new. Ethereum tried it first, at far greater scale, and the results are now in.

EigenLayer went from $22.06bn at its August 2025 peak to between $5.10bn and $6.30bn. Symbiotic fell from $2.70bn to $342.8m. The whole liquid restaking category, worth more than $15bn in 2024, stood at about $1.32bn across 28 protocols — a contraction of more than 90% — with Kelp's rsETH at $1.029bn making up 78% of what remains and Renzo's ezETH down to $110m. On 6 August 2026 ether.fi removed restaking from weETH entirely, moving it to a separate opt-in token; under 1% of ether.fi's assets were still restaked with EigenLayer, down from about 50% earlier in the year.

There was also a demonstration of what the tail risk looks like. On 24 April 2024, Renzo's ezETH briefly printed around $700 on Uniswap after a disappointing token disclosure, causing more than $65m of liquidations across Morpho and Gearbox. The underlying ETH was fine. The receipt token was not.

The lesson transfers directly. In both cases the yield was driven by points and token incentives rather than by fee revenue from the services being secured; in both cases the extra return was small relative to the layers of contract, operator and peg risk stacked underneath it; and in both cases the capital left once the incentives thinned. BTCfi is running the same experiment on a less liquid asset with a thinner set of venues.

Five questions to ask before depositing

For readers who intend to look anyway, the useful diligence here is different from the diligence you would run on a savings account. The counterparty is code and an operator set rather than a balance sheet, so the questions change.

1. Does your bitcoin leave the Bitcoin chain? A Babylon-style timelock keeps it under Bitcoin script control. A wrapped token on another chain does not. These are different risks wearing the same marketing language.
2. What pays the yield? If the answer is a partner chain's incentive budget or a token distribution rather than fees paid by users of the secured service, the rate has an expiry date attached to it.
3. How do you get back to native BTC, and who has to cooperate? A redemption path that runs through a bridge, a minting contract or a single operator is the point at which the sector's losses have historically occurred.
4. What are the slashing conditions, and who can change them? They are written by the secured chain, not by Bitcoin, and they can be amended by governance you do not participate in.
5. How many layers are between you and the position? A liquid staking token representing a restaked position on a wrapped asset is three contracts deep, and each one can fail on its own.

An honest verdict

We do not tell readers what to do with their money, but we will state a reading of the evidence. On the numbers available as of 16 September 2026, the incremental yield available from BTCfi does not compensate a retail holder for the risk added. The verified retail rate is between 0.02% and 0.04%. The sector has lost roughly three quarters of its value from peak. Its largest protocols show fee revenue that is negligible against the capital they hold. And the specific failure mode — an unbacked mint on a bitcoin representation — produced the largest single loss of 2026 so far.

That judgement could change, and the conditions under which it would are worth naming. Genuine fee revenue from the chains being secured, rather than token emissions, would change it. A redemption path that does not depend on a bridge would change it. A sustained rise in the share of bitcoin held on-chain, rather than the current 0.46%, would suggest the market disagrees with us.

Until then, readers who want yield on bitcoin have the options set out onearn Bitcoin, and readers who want yield more than they want it specifically on bitcoin will find better-documented sources instablecoin yield andtokenised treasuries. Readers who hold bitcoin because of what it is may reasonably conclude that four basis points is not a reason to convert it into something else. The framework for that decision is oncrypto earn risks andcrypto staking.