How to earn Bitcoin yield without giving up custody

Bitcoin Bonds are a new mechanism on Stacks, a network built on Bitcoin, that lets BTC holders lock their coins directly on Bitcoin’s own base layer, under a timelock they control, and earn native BTC yield without sending coins to an exchange or custodian. The mechanism went live with the PoX-5 hard fork on July 30, 2026.

Why “Bitcoin yield” has usually meant giving up custody

For most of Bitcoin’s history, earning a return on BTC has meant moving it somewhere else first: an exchange’s “earn” product, a lending desk, or a custodial platform that pools deposits and promises a rate back. That arrangement is a trade, whether it’s stated plainly or not: hand over the keys, receive a yield. When the entity on the other side of that trade mismanages risk or runs out of liquidity, depositors discover that the BTC they thought they owned was actually a claim on someone else’s balance sheet, and claims on a balance sheet are only as good as the balance sheet.

That history is a large part of why so much Bitcoin sits in cold storage earning nothing: for a long time, “yield” and “custody” were bundled together as a single, non-negotiable package. Bitcoin Bonds are notable specifically because they try to unbundle that pair.

What actually changed: the PoX-5 hard fork

Stacks activated a hard fork called PoX-5 at Bitcoin block 960,230 on July 30, 2026, according to an announcement on Chainwire. PoX-5 is what introduces Bitcoin Bonds, which pair BTC held on Bitcoin’s own layer 1 with STX, Stacks’ native token. The rollout is explicitly phased, starting with an institutional-only product rather than an immediate public launch.

Worth being precise about what a hard fork activation does and doesn’t mean here: it means the underlying protocol rules are live on-chain. It does not mean every product built on top of it is available to everyone yet, which matters for what comes later in this article.

How the self-custody claim actually works

The claim worth scrutinizing in Stacks’ product is what “self-custodial” means mechanically. Stacks’ own Bitcoin Staking page states: “BTC is locked on Bitcoin L1 using a standard timelock under your own keys. It is not bridged, wrapped, or held by a third party.”

Broken down, these are three separate claims:

• Not bridged. The BTC doesn’t cross to another chain to earn yield, the most common point of failure and hack surface in cross-chain crypto products.
• Not wrapped. There’s no synthetic BTC token standing in for the real coin somewhere else.
• Not held by a third party. The lock is a timelock enforced by Bitcoin’s own consensus rules, under keys the holder controls, not custody handed to Stacks, an exchange, or any intermediary.

Stacks’ own design claim about its protocol deserves the same scrutiny anyone should apply to any self-custody claim before committing significant capital: check it against the actual protocol mechanics rather than taking it at face value from marketing copy.

Where the yield actually comes from, and two different numbers

This is the part where it’s easy to accidentally merge two different products into one number, so it’s worth separating them cleanly.

According to Stacks’ Bitcoin Staking page, the underlying revenue source for both products is the same: Stacks miners commit real BTC roughly every 10 minutes to compete for STX block rewards, and that committed BTC is what gets distributed to stakers. It’s a real, ongoing flow of actual BTC rather than newly printed tokens. Stacks says more than 4,200 BTC has been distributed this way through the original PoX mechanism since January 2021.

But there are two distinct products drawing on that flow, with two distinct, non-interchangeable figures:

	STX-only stacking (pre-existing)	Bitcoin Bonds (new, PoX-5)
What you lock	STX only	BTC + a paired STX bond
Historical / target yield	~8-10% APY, per Stacks’ Bitcoin Yield Guide	~3% APY target, per Stacks’ Bitcoin Staking page
Paid in	BTC	BTC
Lockup	Stacking-cycle based (~14 days)	6-month lockup
Payout priority	Standard	Priority ahead of STX-only stakers in the payout waterfall

Both figures are Stacks’ own historical and target numbers, not guarantees, and both move with real network conditions rather than being fixed. The ~3% is also an annualized figure: over a 6-month bond it works out to roughly half that. And live data from Stacking Tracker shows the realized rate varying widely from one stacker to another within a single cycle, even before comparing across products. Treat every percentage in this article, on either product, as a historical range or a stated target, not a promise of what any specific holder will earn.

What it costs to participate

Bitcoin Bonds aren’t just a BTC deposit. Per Stacks’ Bitcoin Staking page, participating directly requires pairing the BTC with an STX bond worth approximately 5% of the BTC position’s value. Direct participation also carries a meaningfully higher STX minimum than pooling does; Stacks makes pooled participation available for holders who can’t meet the direct minimum on their own, similar in spirit to how liquid or pooled staking works on other proof-of-stake networks for people below a validator’s solo minimum. On the STX-only side of the older mechanism, Stacking Tracker’s live data shows a direct stacking threshold around 50,000 STX, again with pooling available below it.

The Genesis Bond and who can join today

The first bonding period built on Bitcoin Bonds, the Genesis Bond, was announced August 11, 2026, according to CryptoBriefing, and opened on September 10, 2026, at Bitcoin block 966,350. About 250 BTC was bonded at launch by four institutions, above the 100-200 BTC originally targeted, according to Stacks’ launch announcement. Direct participation is still whitelisted: Stacks describes it as open to institutions and individual holders committing durable BTC capital, with new bonding periods roughly monthly. CryptoBriefing’s reporting reiterates the same self-custody framing covered above: BTC never leaves layer 1.

That’s worth stating plainly because it’s easy to skim past: if you’re a retail Bitcoin holder reading this, direct self-custodial bonding is still gated by a whitelist. The same announcement says individual holders can participate through supported staking pools, but a pool is a different arrangement from locking BTC under your own keys, so check how it holds your coins before assuming the self-custody claim above still applies. Stacks says PoX-6 will remove the whitelist entirely, expected in 6 to 12 months. Until then this is a controlled rollout, the kind of cautious sequencing that’s reasonable for a mechanism this new and this consequential to get wrong.

What’s still unproven

Two things worth sitting with before treating any of this as settled.

First, the backstop behind the ~3% target rate is governance-ratified, but it’s new. Bitcoin Bonds use what’s described as a “waterfall” payout: bond holders get paid their target rate first, and any surplus miner revenue beyond that splits 85% to STX-only stakers and 15% into a protocol reserve fund, designed to cover roughly 1.2 years of yield commitments, according to CryptoBriefing’s coverage of the mechanism. That 15% allocation was validated through simulations covering 210 two-week cycles, roughly eight years of modeled data. The reserve-fund whitepaper itself was published May 13, 2026, and the waterfall mechanism it describes was formally codified in SIP-045, “PoX-5: Bitcoin Staking and Emission Schedule Alignment,” which went to a community vote on July 6, 2026, and passed with over 99.99% of votes in favor, according to Stacks, well ahead of PoX-5’s own activation on July 30, 2026. SIP-045’s status is recorded as Ratified in the Stacks Improvement Proposals repository. That settles the governance question: the mechanics backing the target rate are approved and already live on-chain, not pending review. What isn’t settled is operating history. The reserve fund has only been live since July 30, 2026, and the first bonds only started in September, so how it performs under sustained real-world stress, as opposed to the 210-cycle simulation, hasn’t been observed yet.

Second, “no slashing” is not the same as “no risk.” SIP-045 states that Bitcoin Staking “does not introduce slashing or protocol-level principal loss,” meaning stakers can’t lose principal purely because of a network penalty. That’s a real, meaningful design difference from proof-of-stake networks that do slash. But it’s Stacks’ own claim about its own protocol, and it says nothing about smart-contract risk, execution risk, or the simple fact that this is a genuinely new mechanism that hasn’t operated at scale yet. New, largely untested, and “no slashing” are not the same statement as “risk-free,” and nothing about this product should be read that way.

A related idea, a different mechanism: RUNEBond

The shift Bitcoin Bonds represent, yield sourced from real protocol activity instead of custody handed to a third party, is similar to the idea behind RUNE bonding on THORChain. THORChain is a separate blockchain whose main job is letting people swap native coins across other blockchains, such as Bitcoin for Ethereum, without a centralized exchange or a wrapped stand-in token in the middle. The network is secured by independent operators who run nodes and lock up RUNE, THORChain’s native token, as a security deposit. RUNEBond is an interface built around that system: it connects RUNE holders, called bond providers, to those node operators so that the holders can earn a share of the node’s rewards without running a node themselves. RUNEBond does not take custody of the RUNE being bonded.

The mechanism itself is genuinely different, though, and it’s worth being precise rather than implying the two are the same thing dressed up differently. Bitcoin Bonds lock BTC directly at the protocol level via a timelock on Bitcoin’s own layer 1. RUNE bonding works by attaching RUNE to a specific operator’s node, called a THORNode, after that operator has approved the holder’s address. The lock-up is tied to THORChain’s regular node rotation, called churn, rather than a fixed calendar timelock. And unlike Bitcoin Bonds, a THORChain bond can be slashed if the node it backs misbehaves. If the underlying question that brought you here is “can I earn real yield without handing custody of my crypto to someone else,” it’s worth reading how that works for RUNE specifically in how THORChain bond providers work, and how that compares to inflation-funded staking on other networks in real yield versus token inflation.

The bottom line

Bitcoin Bonds are a real, verifiable mechanism, not a marketing phrase: PoX-5 activated the underlying protocol rules on July 30, 2026, and the first bonding period built on it, the Genesis Bond, opened on September 10, 2026, with about 250 BTC bonded at launch and direct participation still whitelisted. The self-custody claim is specific and checkable against Stacks’ own documentation. The yield figures come in two different sizes for two different products, and both are historical or targeted rather than promised. The backstop behind the newer product is governance-ratified but untested at scale. None of that makes it risk-free, and none of it makes it hype.

If what’s pulling you toward this is the broader question of putting idle crypto to work without handing it to a custodian, you can compare current THORChain node operators and see live RUNE bonding terms on RUNEBond before committing anything.