Bitcoin ETF issuers, custodians and authorised participants

Who actually sponsors each fund

A spot Bitcoin ETF is a grantor trust, not a company with employees. There is no board of directors in the ordinary sense and no investment manager picking holdings. What there is, is a sponsor — a legal entity that establishes the trust, appoints the custodian and the administrator, files the paperwork and takes the sponsor fee. Everything else is outsourced.

That distinction between sponsor and brand catches people out. IBIT's sponsor is iShares Delaware Trust Sponsor LLC. FBTC's is FD Funds Management LLC. Neither name appears in the marketing. If you are reading a 10-K or trying to work out who is legally on the hook for a disclosure, the sponsor entity is the name you need.

A few other pairings are worth knowing. BTCO is a joint effort between Invesco and Galaxy Digital, pairing a traditional fund house with a crypto-native trading firm. BRRR is run by CoinShares, Europe's largest listed digital asset manager, following its acquisition of Valkyrie. MSBT is the first US spot Bitcoin ETF to carry a major American bank's own name, launched on 8 April 2026. And Grayscale runs two funds side by side — GBTC and the Mini Trust — which turns out to matter more than it sounds, as the creation mechanics section below explains.

Who holds the bitcoin

The custodian is the part of this structure that has no equivalent in an ordinary stock ETF. Nobody worries about who is holding the share certificates for an S&P 500 fund. With bitcoin, the custodian holds private keys in cold storage, and the quality of that arrangement is the difference between a fund and a very expensive story.

Three arrangements break the pattern and each answers a different question.

FBTC self-custodies. Fidelity Digital Assets, an affiliate of the fund's own sponsor, holds the coins. That removes any dependency on Coinbase entirely and is the only such arrangement in the category. It also means custody is not independent of the sponsor, which is precisely the separation that third-party custody is meant to provide. Whether that is a feature or a trade-off depends on how much weight you put on Fidelity's own institutional controls versus structural independence. It is a real choice, not a marketing point.

ARKB splits across three. Anchorage Digital Bank, BitGo and Coinbase Custody all hold portions. This is the clearest structural answer to concentration risk available in the category, and it is available today at a 0.21% fee. We have not been able to verify the precise legal entity names or the split between them from filings, so treat the three-custodian arrangement as confirmed and the allocation as unknown.

HODL is reported as Gemini, and the sources conflict. Third-party data lists Gemini Trust Company as VanEck's custodian; other sources describe a multi-custodian arrangement. We have not resolved it against the prospectus and we are not going to publish a figure we cannot stand behind. If HODL's custody matters to your decision, read the current prospectus directly rather than trusting any directory, this one included.

The Coinbase concentration question, taken seriously

Roughly 80.8% of all bitcoin held inside US ETFs sat with Coinbase Custody Trust Company as of April 2026. Against a category holding something like 1.23 million bitcoin, that is a very large number of coins behind one set of operational procedures.

The alarmist reading of that statistic is wrong, and so is the dismissive one. Take the case against alarm first. Coinbase Custody Trust Company is a New York-chartered limited-purpose trust company, supervised by the New York Department of Financial Services, holding client assets in segregated cold storage that are not on its balance sheet and are not available to its creditors. The funds using it are subject to Exchange Act reporting, audited financial statements and independent verification of holdings. This is not a bitcoin left on an exchange; it is a regulated custody relationship with disclosure obligations attached.

Now the case against complacency. Concentration is not primarily a solvency risk — it is a correlation risk. If one custodian suffers an operational failure, a legal freeze, a cyber incident or a supervisory intervention, it does not affect one fund. It affects most of the category at once, and it does so on the same day. Every diversification benefit you thought you had from spreading across four different sponsors evaporates the moment those four sponsors share a single set of keys. That is worth knowing before you decide it does not bother you.

The practical response, if it does bother you, is not to avoid the category. It is to notice that FBTC and ARKB exist, that they solve the problem in opposite ways, and that neither charges a meaningful premium for doing so. That is an unusually cheap way to buy a diversification you actually want. We cover the broader failure modes — custody, tracking, liquidation, and the protections a 1933 Act trust does not carry — in what can go wrong with a Bitcoin ETF.

The arbitrage layer

Authorised participants: the firms nobody markets to you

An authorised participant is a broker-dealer with a signed agreement letting it create and redeem shares directly with the trust, in large blocks, at net asset value. Ordinary investors cannot do this. APs are the only parties who interact with the fund itself; everyone else trades shares with each other on an exchange.

The names below come from the funds' own SEC filings, not from a data vendor. They change over time, and each list carries the date of the filing it came from.

Read down the columns and a pattern falls out. Jane Street, Virtu Americas and Macquarie appear in every list. ABN AMRO Clearing USA appears in most of them. These are the firms that actually do the work of keeping Bitcoin ETF share prices tied to bitcoin, and the size of a fund's AP roster is a reasonable proxy for how competitive its market making is. IBIT has thirteen; the smaller funds have four or five. That difference shows up directly in the quoted spread, which is why we treat spread as a cost in its own right on the fee analysis.

ARKB is the one fund that does not publish a clean count. Its filings name Jane Street, Macquarie, Virtu Americas, Marex and ABN AMRO Clearing USA "among others", and its FY2025 10-K names only Macquarie in connection with an in-kind AP agreement dated 16 December 2025. We have left it out of the table rather than guess at a number.

How creation and redemption keeps the price honest

This mechanism is the reason an ETF share tracks its underlying asset at all, and it is worth understanding in plain language because everything else on this page depends on it.

Suppose demand pushes IBIT shares to trade slightly above the value of the bitcoin behind them. An authorised participant can now buy bitcoin in the open market, hand it to the trust, receive newly created shares worth marginally more than what it paid, and sell those shares into the demand. It pockets the difference. In doing so it has increased the supply of shares and bought bitcoin — both of which push the two prices back together.

The reverse works identically. If shares trade below the value of the bitcoin, an AP buys the cheap shares, redeems them with the trust for bitcoin, and sells the bitcoin. Share supply falls, the discount closes.

Nobody is doing this out of public spirit. The arbitrage is profitable, which is exactly why it is reliable — and it is why the number of firms competing to do it matters. With thirteen APs racing each other, the mispricing that survives is tiny. With four, it is bigger. This is also the structural difference between a spot ETF and the old closed-end trust model: GBTC traded at a persistent discount of as much as 40% before it converted, precisely because no redemption mechanism existed to close it. The wider mechanics, including how net asset value is struck each day, are set out in how Bitcoin ETFs work, and the basic structure in what a Bitcoin ETF is.

The move from cash to in-kind

When the SEC approved these funds on 10 January 2024, it did so on cash-only terms. Footnote 77 of Release 34-99306 is explicit that the proposals "only contemplate cash creation and redemption by authorized participants" and that in-kind arrangements were outside the scope of the order. In practice that meant an AP wanting to create shares handed over dollars, and the fund itself went and bought the bitcoin — adding a conversion leg, a spread and a slice of slippage that ultimately came out of shareholders' returns.

That changed on 29 July 2025. The SEC approved in-kind creation and redemption in Release 34-103571, published at 90 FR 36248 on 1 August 2025. It was the first major crypto action under SEC Chair Paul Atkins, and it arrived through accelerated approval of exchange rule changes rather than through exemptive relief. The operational groundwork had been laid by a joint statement from the SEC's Division of Trading and Markets and FINRA on 15 May 2025, which addressed how broker-dealers could handle crypto assets in this context.

Eight spot Bitcoin ETFs were covered: IBIT, ARKB, FBTC, HODL, BTCW, BTCO, EZBC, BITB, alongside the equivalent ether products.

Those first two numbers matter because they answer the obvious sceptical question: is in-kind actually being used, or was it a rule change that nobody acts on? For IBIT, at least, it is live at genuine scale. BlackRock's FY2025 Form 10-K reports $5.76bn of bitcoin purchased in kind against shares issued and $845m paid in kind on redemptions, and the H1 2026 10-Q shows multi-billion-dollar activity on both sides. Operational status for FBTC and ARKB we have not been able to verify from filings, and we are not going to assume it.

The third number is the constraint. Only four of IBIT's thirteen authorised participants — Jane Street, Virtu Americas, JP Morgan Securities and Marex — have executed in-kind agreements. An AP cannot deliver bitcoin unless it or an affiliate can actually hold bitcoin, which is a regulatory and operational hurdle rather than a commercial preference. Nine of the thirteen still work in cash. The rule changed faster than the plumbing.

What in-kind actually buys you, and what it does not

The genuine benefits are cost and spread. Removing the cash leg removes the fund's own execution risk: the trust no longer has to buy or sell bitcoin at whatever price it can get on the day, so there is less fund-level slippage feeding into tracking difference. It also lowers the risk an AP takes on between pricing a basket and settling it, and lower AP risk shows up as a tighter quote for everyone else. These are real, and they compound quietly in the tracking difference numbers.

The Grayscale split nobody reports

Here is the genuine oddity in all of this. GBTC was not included in the in-kind order. No separate order has followed through September 2026. Grayscale's FY2025 Form 10-K and its Q2 2026 Form 10-Q both still state that authorised participants may only submit cash orders. The largest fund in the category by age, and one of the largest by assets, is still running on the mechanism everybody else has left behind.

And yet Grayscale's own Bitcoin Mini Trust does offer in-kind. Same sponsor. Same custodian. Same listing exchange. Two funds sitting next to each other in the same product line, one of which can accept bitcoin from an authorised participant and one of which cannot.

We have not found a public explanation, and we are not going to invent one. What we can say is what it implies: a GBTC holder is paying 1.50% for a fund running on older plumbing, while the same sponsor's Mini Trust charges 0.15% and has the newer mechanism. If you hold GBTC, the switching arithmetic — which turns entirely on your embedded capital gain and on whether the account is taxable — is worked through in detail on the fees page.

What any of this means for you

Most of this machinery is invisible if you simply buy a fund and hold it, and that is by design. But three practical conclusions fall out of it.

• A fund with more authorised participants competing to arbitrage it will generally quote a tighter spread — which is a real cost you pay on every purchase, not an abstraction.
• Custody concentration means your diversification across sponsors is thinner than it looks. If that matters to you, FBTC and ARKB are the two funds that address it, at no meaningful cost premium.
• Creation mechanics are not marketing. Cash-only creation adds fund-level slippage that ends up in tracking difference, which ends up in your return.

None of it changes the fact that these funds all hold the same asset. If you want the roster with assets and coin counts, that is the Bitcoin ETF list. If you want our view on which one suits which kind of investor, that is best Bitcoin ETFs. If you are ready to place the order, how to buy a Bitcoin ETF covers the steps and the regulation guide covers how the SEC got here, including the generic listing standards adopted in September 2025 that opened the door to everything that followed.