Institutional crypto staking: the due diligence checklist

9 min read
Institutional crypto staking: the due diligence checklist
In this article
  1. Why “up to X% APY” is the wrong first question
  2. Where the yield actually comes from: protocol fees or token emissions
  3. Who holds the assets: the custody model
  4. What happens when you want out: redemption and lock-up terms
  5. What happens when something goes wrong: slashing and counterparty risk
  6. Is this legal where you operate: regulatory posture
  7. A short checklist for the next meeting

Institutional crypto staking due diligence means checking, before any capital is allocated, where the yield actually comes from, who holds the underlying assets, how redemption works, and who absorbs a penalty if a validator misbehaves. (A validator is a network participant that locks up tokens as collateral, helps propose or confirm blocks, and can be automatically penalized for breaking the network’s rules.) The headline APY (annual percentage yield) answers none of those questions. This checklist walks through the ones that do, using the current wave of custodian and ETF staking products as the working examples.

Crypto staking, in short, is the act of locking a token to help secure a proof-of-stake network (a type of blockchain that relies on locked tokens instead of energy-intensive mining), in exchange for a share of the rewards the network pays out. Institutional demand for it has grown fast enough in 2026 that the largest custodian bank in the world announced a product around it. What has not kept pace is a shared checklist for evaluating what is actually being offered.

Why “up to X% APY” is the wrong first question

A staking product’s advertised rate answers “how much,” not “why.” Two products can quote the same number and still carry entirely different risk: one funded by fees the network actually earns, one funded by creating more tokens, one that can be exited in days, one locked for a period whose length is set by network rules and may have no fixed end date. None of those differences appear in the headline percentage.

The first useful question is not the rate. It is what the rate is made of, who is holding the assets while it accrues, and what has to happen for the institution to get its capital back. Everything below answers those three questions before, not after, capital moves.

Where the yield actually comes from: protocol fees or token emissions

Every staking reward is paid from one of two places, and the difference changes what the yield is worth over time.

Real yield is paid from revenue a network genuinely earns, such as transaction fees, swap fees, or usage-based income. It exists whether or not the token’s price or supply changes, which makes it comparable to ordinary cash flow.

Inflationary yield is paid by creating new tokens. The reward looks like income, but it dilutes every holder of that token, including the staker. For example, a network that pays 8% while increasing the token supply by 7% is paying a real return close to zero, regardless of the headline number. We covered this distinction in more detail in why Ethereum staking rewards keep shrinking and in the difference between real yield and token inflation.

For an institutional allocator, the practical step is to ask the provider to name, in writing, the revenue source funding the reward, and to state what portion (if any) comes from newly created tokens. A provider that cannot separate the two has not done the work.

Who holds the assets: the custody model

There are, broadly, three structures on the market or announced, and they are not interchangeable:

ModelWho holds the assetsExample
Custodial staking, in placeThe existing custodian (assets do not leave)BNY adding staking to its digital asset custody platform with Galaxy as infrastructure partner (announced August 4, 2026)
Fund/ETP wrapperThe fund’s custodian, through the fund structureBitwise Hyperliquid Staking ETP (BHYP), listed on Deutsche Börse Xetra on April 9, 2026
Non-custodial delegationThe asset owner, who keeps the keys throughoutA RUNE holder staking with a THORChain node operator while keeping their own keys

On August 4, 2026, BNY, the custodian bank now holding approximately $63.4 trillion in assets under custody and/or administration as of September 30, 2026, according to its third-quarter financial supplement, announced it would add staking to its digital asset custody platform in partnership with Galaxy, a service it described as subject to regulatory review, according to BNY’s press release. As of November 2026, BNY had not announced a launch date. The point of that structure is that eligible clients’ assets do not have to leave BNY’s custody to earn staking rewards, as CoinDesk reported. This removes a real operational friction: the older pattern of moving assets to a separate staking provider to earn yield.

A wrapper product works differently. Bitwise’s Hyperliquid Staking ETP (BHYP), listed on Deutsche Börse Xetra on April 9, 2026 under ISIN DE000A4ARTJ5, tracks the Kaiko HYPE Reference Rate LDNLF index, carries a 0.85% annual total expense ratio, and targets a net staking reward of approximately 1%, according to Bitwise’s own product page for the ETP. Here the institution holds a listed security, not the underlying token, and the custody, staking and fee stack all sit behind the ETP issuer.

None of the three structures is automatically the safer choice. The question due diligence has to answer is which layers of custody, delegation and fee-taking sit between the institution’s capital and the validator doing the work, and what each of those layers is allowed to charge or decide.

What happens when you want out: redemption and lock-up terms

Staking is not typically instant to exit, and wrapping it in a fund or ETF does not change that. Even when the wrapper itself (an ETF share, for example) trades on an exchange, the underlying staked assets still have to be unstaked on the blockchain before the fund can raise enough cash to meet a large redemption. That on-chain step follows the network’s own timeline, not the fund’s.

The questions to put to any provider, in writing, before allocating:

  • Is there a fixed lock-up, a variable exit window, or no queue at all?
  • Who controls when that window opens: the network’s own rules, or the operator’s discretion?
  • Is there a cap on how much can exit at once, and could a large institutional redemption run into that cap?

An allocator that sizes a position without a written answer to all three is pricing a liquidity assumption, not a defined liquidity term.

What happens when something goes wrong: slashing and counterparty risk

Staking’s downside is not limited to the price movement of the underlying token. Most proof-of-stake networks can also penalize the staked position directly, through a mechanism called slashing. Slashing is an automatic, protocol-level fine for validator misbehavior, charged against the bond or stake the validator has locked. In plain terms, the network can automatically reduce the tokens a misbehaving validator has locked, as a penalty.

THORChain’s own documentation shows how specific these terms can be. A double-signing fault costs 5% of the minimum bond, and an unauthorized transaction costs 1.5 times the value moved. Both are charged against the bond posted by the node operator and the RUNE staked with it, according to THORChain’s documentation on risks, costs and rewards. We cover the full mechanism in slashing in THORChain: the fine that keeps the network honest.

The exact number matters less than the exercise. Every network sets its own slashing terms, and every delegation structure decides differently whose capital absorbs the hit: the operator’s, the delegator’s, or a shared pool. An institution should ask for that answer by name before it delegates, not infer it from a marketing page.

There is a second, structural risk worth naming. If several large custodians route their staking through the same small group of infrastructure providers, a technical failure or compliance problem at one provider could affect assets held across multiple institutions at once. That is a concentration question, separate from slashing. It is worth asking any custodian directly: how many other clients share this exact staking infrastructure?

In the US, an SEC interpretation issued together with CFTC guidance (Release 33-11412) addressed this directly. Self-custodial staking, custodial staking and liquid staking of “digital commodities” on public proof-of-stake networks can fall outside securities transaction rules, according to the release itself and Sidley’s analysis of it. The condition is specific: the custodian or staking provider must act purely as an agent. It cannot decide whether, when, or how much the client stakes, and it cannot fix or guarantee the reward amount. It can, however, charge a fee for the service.

That condition is the actual due diligence question, not a compliance detail to wave past. A provider whose product involves discretionary decisions about staking timing or amount, or that markets a guaranteed or fixed reward rate, may sit outside what that interpretation covers, regardless of how the product is marketed. Outside the US the picture is different again. Under the EU’s MiCA framework, staking has no license category of its own today. Staking offered as a service is captured indirectly, most commonly alongside custody, while staking that holders do on their own account falls outside MiCA, according to a legal analysis of the framework. The European Commission ran a consultation on reviewing MiCA from May to September 2026 that asked whether this treatment of staking is adequate. No legislative proposal has been published yet, so this is a framework under review, not a settled one. The regulatory posture of a staking product depends on the jurisdiction the institution operates under. It has to be confirmed for that jurisdiction specifically, not assumed from a US or EU precedent alone.

A short checklist for the next meeting

  • Name the revenue. Ask the provider to state, in writing, what portion of the yield comes from protocol fees versus new token issuance.
  • Map the custody chain. List every entity that holds, delegates or has discretion over the asset between the institution and the validator.
  • Get the redemption terms in writing. Fixed lock-up, variable window, or none, and who controls when it opens.
  • Ask who absorbs a slashing penalty. By name: operator, delegator, or a shared pool, and at what rate.
  • Ask about infrastructure concentration. How many other clients share the same staking provider or validator set.
  • Confirm the regulatory fit for your jurisdiction. Does the provider’s discretion, or lack of it, match what a regulator in your jurisdiction requires for the structure you are considering?
  • Reject a fixed or guaranteed rate. A real, revenue-funded reward is variable by design; a promised fixed rate is a red flag, not a feature.

None of the checklist items above are specific to any one network, and that is the point: they apply whether the product under review is a custodian’s new staking service, an ETP, or a non-custodial delegation model like staking RUNE on THORChain (THORChain calls it bonding), where a RUNE holder (a bond provider) stakes with a node operator and keeps their own keys throughout, rather than moving the asset to a custodian at all.

For institutions evaluating that non-custodial route specifically, including the practical mechanics of bond provider delegation, node operator selection and slashing exposure, our institutional page explains the mechanism in more detail and puts you in touch with our team.

Frequently asked questions

What is institutional crypto staking due diligence?

It is the process of checking a staking product before allocating capital to it: where the yield actually comes from, who holds the underlying assets, how redemption works, what happens if a validator is penalized, and whether the structure fits the regulatory framework the institution operates under. It is the same discipline applied to any yield-bearing product, adapted to what makes staking different: the assets are on-chain, the penalties are automatic, and the yield is variable by design.

Is crypto staking legal for institutional investors in the US?

An SEC interpretation issued together with CFTC guidance (Release 33-11412) states that self-custodial, custodial and liquid staking of digital commodities on public proof-of-stake networks can fall outside securities transaction rules, provided the custodian or staking provider acts only as an agent: it does not decide if, when or how much the client stakes, and it does not fix or guarantee the reward amount. If a provider exercises that kind of discretion, or guarantees a return, the interpretation no longer covers it. This is not a blanket approval of every staking product; it is a test each structure has to pass.

What's the difference between real yield and inflationary staking rewards?

Real yield is paid from revenue a network actually earns, such as transaction or swap fees. Inflationary yield is paid by creating new tokens, which dilutes every holder, including the staker, so a high headline rate can still be a negative real return once dilution is priced in. Due diligence has to separate the two, because the source determines whether the yield is sustainable if new token issuance slows or stops.

Can institutions stake crypto without giving up custody?

Yes, in more than one way. A custodian can add staking to assets it already holds, so nothing moves out of the existing custody account, which is the model BNY announced with Galaxy in August 2026. Separately, a non-custodial model lets the asset owner keep their own keys throughout and delegate only the staking or bonding role to an operator. Both remove the older pattern of sending assets to a third party to earn yield, but they are not the same thing: one keeps a custodian in the loop, the other removes one entirely.

What happens to staked assets if a validator gets slashed?

It depends on the network and the delegation structure, and this is exactly what due diligence has to confirm before allocating, not after. On THORChain, for example, the network's own documentation sets a penalty of 5% of the minimum bond for a double-signing fault, and 1.5 times the value of any unauthorized transaction, both charged against the bond the operator and their delegators have posted. The question to ask any provider is direct: whose capital absorbs a penalty, the operator's, the delegator's, or both?

How liquid is an institutional crypto staking position?

Rarely as liquid as the marketing implies. A staking ETF may keep part of its assets unstaked to meet ordinary redemptions, but a redemption larger than that depends on assets unstaking on-chain before the fund can raise cash, and that can take longer than same-day settlement. Direct staking and bonding positions typically have their own lock-up or exit-window mechanics tied to the network's own rules. The redemption terms, in writing, are a required piece of due diligence, not an assumption to make from the product name.

What should disqualify a crypto staking product during due diligence?

A provider that cannot name where the yield comes from, that guarantees a fixed reward rate, that cannot explain in writing who absorbs a slashing penalty, that offers no verifiable redemption timeline, or whose custody and staking discretion make the SEC/CFTC agent test a stretch, not a fit. Any one of those is a reason to keep looking, before capital moves.

Put your RUNE to work

Compare THORChain nodes by fee, minimum stake and track record, and request a spot from your own wallet.

Find a Node