Castle Labs says Bitcoin has grown into a reserve-style asset without building a comparable native yield layer, leaving most BTC economically idle even as institutional adoption expands.
In the firm’s review of the on-chain Bitcoin finance market, BTC is described as the world’s 13th-largest asset with a market capitalization of about $1.3 trillion. Yet out of roughly 20.05 million BTC in active supply, only 311,000 BTC, or about 1.5%, is currently deployed in strategies that earn any yield at all.
That stands in sharp contrast with Ethereum. Castle Labs notes that 32.5% of circulating ETH is staked for an estimated 2% native yield, with a second layer of capital efficiency built on top through liquid staking. Lido alone accounts for 21% of all staked ETH and issues stETH, which can be reused across DeFi applications.
Why Bitcoin’s yield layer remains thin
The report ties the gap to both protocol design and holder behavior. Bitcoin is a proof-of-work network, so it does not have a native staking mechanism. At the same time, BTC holders are described as deliberately conservative, especially after the collapse of centralized lending platforms in 2022.
As a result, Bitcoin holders still mostly rely on price exposure rather than a built-in return mechanism comparable to lending income, dividends, or staking rewards in other markets.
The collapse of centralized BTC lending
Before on-chain crypto finance matured, firms such as Celsius, BlockFi, and Voyager offered BTC yield products in exchange for custody. Castle Labs says those platforms failed through nearly identical structures: they took legal ownership of customer deposits and re-lent them to opaque counterparties.
Celsius, according to the report, explicitly transferred “all rights and title” in depositor assets to the platform through its terms, allowing rehypothecation. Its failure was triggered by the liquidity squeeze that followed the Terra/Luna collapse. Castle Labs says 75% of Celsius withdrawals came after that event, forcing the company to halt redemptions.
Voyager collapsed after Three Arrows Capital defaulted on a loan of $350 million in USDC and 15,250 BTC. BlockFi was first hit by the same 3AC exposure and then by a second default tied to an approximately $680 million loan to Alameda Research.
The common pattern, in Castle Labs’ framing, was the use of retail deposits to fund undercollateralized institutional borrowing while stripping depositors of legal claims on the underlying assets.
DeFi opened new paths for BTC, but added new risk layers
Most active BTC yield now comes from mature DeFi ecosystems on EVM chains and Solana, where lending markets, decentralized exchanges, and vault structures already exist. But native BTC cannot interact with those environments directly. Holders first have to convert it into a representation that another chain can recognize.
Castle Labs groups those representations into custodial mints such as WBTC and Coinbase’s cbBTC, decentralized bridge assets such as tBTC, and liquid staking-style assets such as LBTC. Each route, it argues, adds another trust assumption.
Bridge and custody risk
To mint a represented version of BTC, users send native BTC to an address or custodian and receive a 1:1 receipt token on another chain. For WBTC and cbBTC, that process depends on centralized operators and carries key-management and social-engineering risk.
Threshold’s tBTC removes the single-custodian model through a signer network, but it still introduces signer-set and smart contract risk.
The report highlights a specific change at BitGo in August 2024. It says BitGo shifted from unilateral control over all three keys in WBTC’s 2-of-3 multisig to a split arrangement across BitGo Inc. in the U.S., BitGo Singapore Ltd., and BiT Global in Hong Kong. Even after the change, the third key remained with BiT Global, which Castle Labs identifies as a key risk factor because of its connection to Justin Sun’s Tron ecosystem and because BitGo is only a minority shareholder in that venture.
cbBTC, by contrast, is described as fully backed 1:1 but entirely controlled by Coinbase, making it a single-custodian structure that can freeze or seize funds under its user agreement.
Smart contract risk
Castle Labs argues that any DeFi deployment exposes users to smart contract risk, and that years of market use have not eliminated the problem. It cites a November 2025 Balancer exploit in which roughly $128 million was drained from Balancer V2 pools despite multiple audits, causing major losses for liquidity providers.
The report also points to a more recent vulnerability involving kelpDAO’s rsETH. In the first six hours after that issue surfaced, more than 86,000 BTC left the protocol as users reacted to possible contagion. Castle Labs says the attacker deposited rsETH as collateral and borrowed large amounts of wETH. Once wETH liquidity was drained, lenders began using their own deposits as collateral to borrow elsewhere in stablecoin markets. Users responded by pulling assets, and those withdrawals represented roughly 12% of the protocol’s BTC-based assets.
The paper says that episode damaged confidence around Aave, described here as the largest lending protocol. Castle Labs ties the fallout to a mix of external asset risk, Aave’s pool-based design, and governance oversight gaps during inclusion in E-mode, which pushed LTV to 93%. It says Aave’s market size later fell by more than 50%.
Bitcoin L2s and staking protocols form the second major route
The report places Babylon, Lombard, Stacks, Rootstock, BOB, and related protocols into a second branch of the BTC yield market, with combined total value locked of about $4 billion.
Babylon accounts for the largest share of that locked value, according to Castle Labs. Its core model is native BTC staking on Bitcoin L1 to help secure proof-of-stake networks using Bitcoin’s base-layer security guarantees. In return, users earn yield denominated in the native tokens of the secured chains.
Users seeking liquidity can deposit native BTC or WBTC into Lombard and receive LBTC, a liquid and yield-bearing representation. Castle Labs therefore classifies Lombard as a liquid staking protocol.
That said, these structures still expose BTC holders to slashing risk. The report adds that the rest of the locked value in this segment comes from BTC L2s trying to launch and sustain their own chains, something it describes as historically difficult because it requires on-chain activity, builders, stablecoin rails, strong user experience, and incentive programs that keep capital in place.
Trust assumptions differ from one BTC yield model to another
Castle Labs argues that market demand and technical infrastructure both exist, but not in one place and not under one unified trust model. EVM ecosystems offer deep liquidity and rich applications, but they force users to accept custody or bridge risk. Bitcoin L2s align more closely with BTC-holder preferences, but still lack broad yield diversity and polished retail UX. Centralized lending, meanwhile, was structurally broken by the 2022 blowups.
The report lists the trust assumptions behind each route:
- Wrapped BTC requires trust in custodians and/or bridge signer sets, plus the risks of the wrapper itself and downstream smart contracts.
- Babylon native staking relies on a covenant committee and an M-of-N multisig, currently 6-of-9, including three signers from Babylon Labs and one each from CoinSummer Labs, RockX, AltLayer, Zellic, Informal Systems, and Cubist, to co-sign unstaking and slashing transactions.
- Stacks’ sBTC bridge relies on a 15-institution signer set that includes Figment, Chorus One, and Stacking DAO. Bridge operation requires 70% honest participation, or 11 of 15 signers, while system safety can still hold and permit non-withdrawal if 33% of signing power remains honest, or 5 of 15.
For Castle Labs, the unmet need is a way to preserve as much of Bitcoin’s trust model as possible while still giving BTC capital productive uses and enough liquidity to move.
What Castle Labs means by “Bitcoin-native finance”
The report defines Bitcoin-native finance as financial products that use BTC as the primary asset while keeping settlement, security, and asset movement as close to Bitcoin L1 as possible and minimizing added trust assumptions.
It also says full Bitcoin-native finance cannot be built directly on Bitcoin L1 because of the chain’s design limits:
- Bitcoin Script can enforce simple spending conditions but is not Turing complete, so it cannot express the state machines required for DeFi protocols such as AMM curves, health-factor checks, or liquidation logic.
- The UTXO model does not provide shared persistent state, while lending markets and AMMs require shared mutable state.
- Bitcoin block space is scarce and blocks arrive roughly every 10 minutes, making high-frequency DeFi impractical on L1.
- Bitcoin has no native yield mechanism because it uses proof of work.
- Bitcoin cannot verify external chain state, so moving BTC into a programmable environment requires custodians or signer sets.
Within those limits, Castle Labs says every product on the market today depends on some trust assumption outside Bitcoin consensus. Some protocols still try to stay as close to Bitcoin as possible.
Babylon avoids a traditional bridge by keeping staked BTC in timelocked Bitcoin UTXOs, though users take covenant-committee risk and slashing risk tied to delegated finality providers securing external proof-of-stake chains. BOB pairs Bitcoin-finality ambitions with a full EVM environment through a BitVM-style bridge, but for now settles on Ethereum and treats full Bitcoin security as a staged roadmap item. Stacks anchors execution finality directly to Bitcoin consensus while accepting a signer set for its BTC-backed bridge asset sBTC, again requiring 11 of 15 signers to move funds.
Castle Labs provides scale data for some of these protocols: Babylon at $2.6 billion in TVL, Lombard at $800 million, and Solv at $480 million. It says users are drawn to them because they can combine with other DeFi products and stack extra yield on top of a base return.
LBTC earns through staking, with returns sourced from the Babylon network as it helps secure proof-of-stake blockchains. Solv’s solvBTC generates yield through delta-neutral strategies, lending, liquidity provision, staking, and real-world assets, or RWA.
The report says there is still no complete Bitcoin-native solution, but a closer approximation today would mean earning yield without giving up self-custody or introducing avoidable risk, obtaining a liquid representation of committed BTC when needed, and keeping all risk tradeoffs fully transparent.
It suggests users ask four questions when evaluating any offer: Does the structure preserve self-custody? How decentralized is the bridge signer set? Where does the yield actually come from? Can the holder exit or redeem freely, or is there a lockup?
Stacks as a case study
Castle Labs uses Stacks as its main case study for a closer-to-Bitcoin finance stack. It describes Stacks as a BTC L2 whose execution layer is anchored to Bitcoin. Current on-chain TVL is listed at about $90 million, with another $190 million in sBTC.
The report says Stacks is one of the most complete attempts so far to gather the main parts of a Bitcoin-native stack in one place: a Bitcoin-anchored execution layer, a trust-minimized BTC bridge, BTC-denominated native staking, a liquid staking representation, and live DeFi applications such as Zest, Bitflow, StackingDAO, and Hermetica.
Bitcoin-anchored finality after the Nakamoto upgrade
According to Castle Labs, the broader Stacks chain and sBTC are supported by a finality model anchored to Bitcoin. Since the Nakamoto upgrade in October 2024, each Stacks miner’s block-commit transaction anchors the full chain state of the previous tenure into a Bitcoin transaction.
That means reversing a finalized Stacks tenure would require reversing the corresponding Bitcoin block. The paper says a Stacks block reaches full Bitcoin finality roughly one Bitcoin block after its tenure ends.
sBTC and the bridge model
sBTC is the mechanism that lets BTC move between Bitcoin L1 and Stacks without a traditional custodian. It serves as the bridged representation of Bitcoin on Stacks and currently holds about $190 million in TVL.
All bridged BTC sits in a single peg wallet, a Bitcoin Taproot address. Movement of those funds is controlled by a 15-member signer set elected by the community, with a 70% threshold. In practical terms, 11 signers are needed to move funds.
PoX-5 and self-custodied BTC staking
The report gives special attention to Stacks’ proposed PoX-5 upgrade, which it says would let BTC holders earn BTC-denominated yield while their BTC remains locked on Bitcoin L1. Launch is expected in late August.
The mechanism uses standard OP_CHECKLOCKTIMEVERIFY timelocks. In Castle Labs’ description, BTC stays committed and yield-bearing while remaining under the holder’s own custody, with lockup enforced by Bitcoin consensus itself rather than by custodians, bridges, or signer committees.
Yield in PoX-5 comes from the Stacks consensus mechanism, where miners bid BTC for the right to mine Stacks blocks. Castle Labs says that mechanism has distributed more than 4,200 BTC to STX stakers since January 2021. Under PoX-5, BTC stakers would be able to earn through that same process.
Participation requires a BTC commitment locked on L1 and an STX position worth at least 5% of the BTC value for a six-month term. The paired BTC-STX commitment forms a senior tranche with first claim on each cycle’s miner BTC pool and a target APY of 3%. Remaining yield goes to STX-only stakers, which form a junior tranche, and to a reserve fund that accumulates excess miner revenue for periods when miner bids may fall short of payout obligations.
Castle Labs says there is no forced exit in this model. Holders can leave with their BTC within a few Bitcoin blocks, but they forfeit the remaining term’s yield, while the paired STX stays locked for the full period. The structure does not include slashing risk, meaning stakers are not exposed to principal loss through that mechanism.
From native yield to lending and liquid staking
Making BTC productive in a native way is only the first step in the Stacks roadmap, the report says. The team is also studying self-custodied Bitcoin lending. In that model, users would be able to post BTC as collateral in lending protocols on Stacks such as Zest while the BTC itself never leaves Bitcoin L1.
Stacks would continuously verify the L1 collateral by reading Bitcoin state. That would let borrowers access Circle’s USDCx and Hermetica’s Bitcoin-backed stablecoin USDh for additional yield strategies while still earning PoX-5 staking rewards.
There is a tradeoff, though. BTC locked on L1 and paired with STX is productive because it earns a base yield and remains redeemable, but it cannot be used elsewhere at the same time. For users who want DeFi composability, Castle Labs points to a liquid staking token structure.
StackingDAO is preparing to launch stBTC, described in the report as a Bitcoin analogue to stETH. It is a liquid staked BTC token whose main advantage is that BTC holders do not need to source their own 5% STX pairing. StackingDAO provides that from STX already backing its stSTX and stSTXbtc products, allowing holders to earn BTC yield on BTC without taking direct STX risk while retaining the highest-priority claim on rewards.
In the model outlined by Castle Labs, users deposit native L1 BTC or sBTC and receive stBTC. As sBTC rewards generated by the PoX-5 mechanism accumulate and are paid out at the end of each cycle, stBTC appreciates through automatic compounding. Part of those rewards goes to supporters supplying the STX pairing through stSTX and stSTXbtc, while the rest of the sBTC is reinvested into the pool backing stBTC, increasing the stBTC-to-sBTC ratio.
For withdrawals, stBTC exits can be handled through a reserve buffer for a small fee. If that buffer is exhausted, withdrawals move to a one-cycle cooldown process using a withdrawal NFT.
The application layer: Zest, Granite, Hermetica, and Bitflow
Castle Labs also details the DeFi applications built around that stack.
Zest is described as the largest DeFi application and lending market on the chain. It allows users to deposit BTC-denominated collateral, including sBTC, STX, and stSTX, and borrow stablecoins in isolated-risk markets. Zest has also announced “Stacks Vaults,” which are set to launch alongside Stacks’ BTC staking. The first vault will run a fully automated stBTC looping strategy with a target APY of 6% to 8%, which the report calls the first example of putting liquid staked BTC to work through the Stacks financial stack.
Granite is another lending protocol that lets users borrow stablecoins against BTC or sBTC collateral with isolated risk.
Hermetica issues USDh, which the report defines as a synthetic dollar backed by Bitcoin and built around basis-trade mechanics.
Bitflow is the DEX on Stacks. It uses a Curve-style Stableswap AMM designed for BTC-anchored assets such as sBTC, xBTC, aBTC, and WBTC, as well as Stacks-native stablecoins.
A funnel model for Bitcoin capital
Castle Labs frames Bitcoin staking as the top of the funnel for Stacks. It is meant to attract BTC capital looking for yield without giving up custody. Once that capital is committed, the surrounding application layer can route it toward borrowing on Zest, swapping on Bitflow, structured basis-trade yield on Hermetica, and added liquidity through StackingDAO once stBTC goes live.
Usage of those applications, the report says, drives on-chain activity and fee generation. Over time, those fees could become a meaningful and self-sustaining source of consensus yield alongside miner bids into the PoX pool.
Castle Labs’ closing view
The report concludes that Bitcoin’s idle-capital problem is both real and structural. Every route built to compensate for the lack of native BTC yield — wrapped BTC, BTC L2s, or CeFi lending — comes with its own trust assumptions.
BTC currently represents about 7% of DeFi TVL, or roughly $5.2 billion, according to the paper, but Castle Labs says that remains low relative to Bitcoin’s overall size. Most BTC is still held primarily for price appreciation.
Stacks, in Castle Labs’ view, has spent years trying to reduce those extra assumptions. With the PoX-5 upgrade approaching, BTC holders may be able to earn yield natively from miners bidding BTC to produce Stacks blocks. Whether that model becomes widely adopted, the report says, will depend on how tightly the application ecosystem grows alongside native BTC staking, because maintaining the targeted 3% APY requires miners to keep bidding.
Castle Labs adds that the mechanism already has more than five years of mainnet activity behind it. Bitcoin remains a reserve asset and is widely adopted by institutions, but greater productivity, the report argues, would only expand its ability to become part of more balance sheets and make the asset more useful for new holders.