Best Bitcoin ETFs: our picks by what you are actually doing

Almost every "best Bitcoin ETF" page you can find is a sorted table with the cheapest fund at the top. That is not a recommendation, it is a spreadsheet — and it quietly assumes every reader wants the same thing. They do not. A person buying once and forgetting about it for a decade, a person writing covered calls every month, and a person who lies awake worrying that four fifths of all ETF-held bitcoin sits with one custodian are three different problems with three different answers.

So this page is organised by situation. Each pick gets an argument and an honest statement of what you are accepting in exchange. If what you want is the underlying data rather than our reading of it, the full Bitcoin ETF list has every ticker, issuer, exchange and coin count, and the fee and spread page works the cost arithmetic properly. This page assumes you have those and want a view.

The picks at a glance

One table, then the argument. Read the table for orientation and the prose for the reasoning, because the reasoning is the part that decides anything.

How we chose, and what we refuse to do

Three inputs, weighted differently depending on the use case. Cost means the standing sponsor fee after waivers, not the promotional rate an issuer advertised at launch. Tradability means the published median bid-ask spread, average volume and, where relevant, the depth of the listed options market. Structure means who custodies the coins, whether in-kind creation is available, and whether the fund is large enough that closure is not a live risk — a question that stopped being theoretical when Hashdex's DEFI wound down in August 2026.

What we do not do: no issuer pays for placement, ranking or inclusion. There are no affiliate arrangements with any fund company on this site, and no fund is here because someone asked. Every figure is checked against issuer material or an SEC filing where one exists, and marked as aggregator-sourced where it does not. Our methodology page spells out the confidence level attached to each field, including the several places where we have simply written "not published" rather than borrow a number we cannot stand behind.

Best for long-term buy-and-hold — Grayscale Bitcoin Mini Trust (BTC)

If you intend to buy bitcoin exposure and then largely leave it alone, the expense ratio is the dominant term and almost nothing else matters. The Mini Trust charges 0.15% a year against IBIT's 0.25%. Ten basis points sounds like rounding. Over a decade on a position that compounds, it is not — it is a permanent, guaranteed drag applied daily to your entire balance, in a category where every fund holds the identical asset.

The reason we prefer BTC to the two funds that are nominally cheaper is scale you can verify. It held $4.78bn and 61,888 bitcoin on 1 September 2026, quotes a 0.03% median spread, and has in-kind creation and redemption available — which its own parent fund GBTC, oddly, still does not. That combination of a low fee with genuine size is rare here. Most of the cheap funds in this category are cheap and tiny.

The trade-off you are accepting. The Mini Trust uses exactly the same custody arrangement as GBTC — Coinbase Custody Trust Company, with Anchorage named as an additional custodian — so it does nothing to reduce the category's custody concentration. Its options market is far shallower than IBIT's, and it was excluded from Cboe's October 2024 options approval batch that covered its ten peers. There is also a structural wrinkle worth knowing: the fund was created by spinning roughly 10% of GBTC's bitcoin off pro rata on 31 July 2024, and it ran a 1-for-5 reverse share split on 19 November 2024, which breaks any per-share history you might try to chart.

The compounding term

Best for traders and options users — IBIT

Everything that makes IBIT expensive relative to the Mini Trust also makes it the only serious answer for anyone who trades. Its 30-day median bid-ask spread was 0.03% on 31 August 2026, it has quoted in penny increments under the Penny Interval Program since 2 January 2025, and its 30-day average volume has run above 54 million shares. When you are moving size, or moving often, that is the number that decides your cost.

The options market is where the gap becomes a chasm. IBIT is the deepest listed bitcoin options market anywhere, with open interest around 7.8 million contracts in late August 2026 and daily volume near 2 million. Position limits have been raised twice, from 25,000 contracts to 250,000 in July 2025 and then to 1,000,000 in April 2026. If you want to write covered calls against a bitcoin position, buy protective puts, or build a collar, the strikes and expiries simply exist here and are thin or absent everywhere else. It is no accident that the covered-call funds on our Bitcoin income ETF page build their positions around this one ticker.

The trade-off you are accepting. 0.25% — eleven basis points above MSBT and ten above the Mini Trust, forever, on the whole position. If you buy quarterly and hold for a decade you will cross the spread perhaps forty times and pay the fee roughly 3,650 times. In that scenario you are buying a liquidity premium you will never use. We go through the same argument fund-by-fund on the IBIT profile.

Best on headline fee — MSBT, with a caveat we will not bury

Morgan Stanley's Bitcoin Trust charges 0.14%, the lowest standing sponsor fee of any US spot Bitcoin ETF. It is the first fund in the category to carry a major US commercial bank's own name on the door, it uses Coinbase Custody with BNY as cash custodian, administrator and transfer agent, and it tracks the CoinDesk Bitcoin Benchmark 4pm New York settlement rate. On a pure fee comparison it wins, and if the fee were the only variable this page would be one paragraph long.

It is not, and here is the honest position. MSBT launched on 8 April 2026, which makes it young by any measure. No reliable assets figure is published for it — the last thing we can verify is $183.6m of cumulative flows at thirteen trading days, on 29 April 2026, which tells you about the launch and nothing about today. We have no published median spread, no verified coin count, and it has no listed options. Its debut ranked in the top 1% of all ETF launches by early volume, which is encouraging, but early volume is not the same as a durable quote.

Say it plainly: the cheapest fee on a fund you cannot size or price properly is not automatically the best deal. If you buy MSBT, you are making a reasonable bet that a distribution machine the size of Morgan Stanley's will fill the gaps in time. That is a real argument. It is just a different argument from the one the fee table appears to be making.

Best if you want custody away from Coinbase — FBTC

Roughly 80.8% of all bitcoin held inside US ETFs sat with Coinbase Custody Trust Company as of April 2026. That is not a scandal, and Coinbase Custody is a New York-chartered limited-purpose trust company with a serious institutional record. But it is a concentration, and you own it whether or not you thought about it when you placed the order.

Fidelity's FBTC is the only fund in the group that removes the dependency entirely. Fidelity Digital Assets holds the bitcoin in-house, making FBTC the sole fund whose custodian is an affiliate of its own sponsor rather than a third party. It is also a serious fund on its own terms — roughly $13.6bn and 175,483 bitcoin as of the end of August 2026, second in the category, with a 0.03% median spread and a live options chain.

The counter-argument, which is real. The whole point of an independent custodian is that somebody outside the sponsor's own corporate tree is holding the assets and reconciling them. FBTC gives that up in exchange for removing the Coinbase dependency. Which risk you prefer is a genuine judgement rather than an obvious call, and anyone telling you self-custody by the sponsor is unambiguously safer is skipping the interesting half of the question. The issuers and custodians page lays out each arrangement in full.

Best for custody diversification — ARKB

If concentration is the thing that bothers you, ARKB is the most direct structural answer available. It is the only US spot Bitcoin ETF that spreads its holdings across three custodians: Anchorage Digital Bank, BitGo and Coinbase Custody. Nothing else in the category does this. It held roughly 33,957 bitcoin on 1 September 2026, worth about $2.6bn at the reference price, so it has the scale to be taken seriously.

Two details most directories get wrong and you should know. The sponsor is 21Shares US LLC, with ARK Investment Management acting as sub-adviser — the billing order in the fund's name misleads a lot of comparison tables. And the fund's benchmark switched to the FTSE Bitcoin Index effective 27 August 2026, so any tracking analysis that straddles that date is comparing two different reference rates.

The trade-off you are accepting. You pay 0.21%, six basis points above the Mini Trust, and you cross a 0.05% median spread — the widest of the big five, and roughly double IBIT's. Splitting custody across three providers has an operational cost and you are paying part of it. Whether three custodians is genuinely safer than one very good one is also arguable: more relationships means more surface area as well as less single-point risk.

Best for transparency-minded investors — BITB

Bitwise does something none of its competitors does: it publishes the wallet addresses holding the fund's bitcoin, so anyone can check the chain rather than take the audit on trust. It also directs a share of the fund's profits to Bitcoin open-source development, which is either a meaningful contribution to the network you are buying exposure to or a marketing gesture, depending on your temperament. We think it is the former, and either way it costs you nothing.

On the numbers, BITB charges 0.20%, held roughly $3.0bn and about 41,500 bitcoin at the end of August 2026, and quotes around a 0.03% spread. Note the listing venue carefully: BITB trades on NYSE Arca, not Cboe BZX. A surprising number of comparison tables have this backwards. The authoritative source is the SEC's approval order of 10 January 2024, Release No. 34-99306, which places Bitwise under NYSE Arca Rule 8.201-E.

The trade-off. Third-party data puts BITB's tracking difference around 0.15%, wider than IBIT's or FBTC's roughly 0.03%. If that holds, the disclosure premium is costing you more in practice than the five basis points of fee difference suggests. Verify it against current data before you decide it does not matter.

The one to think hard about — GBTC at 1.50%

Grayscale Bitcoin Trust charges 1.50% a year: six to ten times its peers, for an identical asset in comparable custody. On a $25,000 position that is $375 a year against $35 for MSBT. The market has responded exactly as you would expect. Roughly $17.5bn has left the fund since it converted in January 2024, and it has shed something like 447,000 bitcoin — it now holds 130,481. Grayscale's own answer to the problem was to launch the Mini Trust at a tenth of the price, using the same sponsor and the same custodian.

And yet. GBTC prints the tightest bid-ask spread in the entire category, 0.02% on 1 September 2026, tighter even than IBIT. It still holds $10.1bn. Whatever you think of the fee, this is not a neglected husk. It is also the only fund in the group still restricted to cash-only creation and redemption — its FY2025 10-K and Q2 2026 10-Q both state that authorised participants may submit cash orders only, while the Mini Trust does offer in-kind. That is an underreported quirk with real implications for how tightly the fund can be arbitraged.

The decision splits cleanly on account type. In an IRA, a Roth IRA or a 401(k), there is no argument for staying. Selling GBTC and buying the Mini Trust inside a tax-sheltered account triggers nothing, costs you one spread crossing, and saves 135 basis points a year on the same bitcoin at the same custodian. Do it.

In a taxable account with a large embedded gain, it is genuine arithmetic. Switching means realising the gain now and paying tax on it, which permanently shrinks the capital that compounds. Against that you save 1.35% a year for as long as you hold. Roughly: the larger your unrealised gain, the higher your marginal rate, and the shorter your remaining horizon, the more staying makes sense. The longer you plan to hold, the faster the fee saving overtakes the tax bill. Work it out with your own numbers rather than accepting either the "always switch" or the "never realise a gain" reflex — and read our guide to Bitcoin ETF taxes before you act, because the wash sale rule does apply to these shares and that changes the sequencing.

Where we see people go wrong

The most common error is treating the fee table as the whole decision when you contribute monthly. Twelve purchases a year means twelve spread crossings, and on the wider funds that adds up faster than a ten-basis-point fee gap. The mirror-image error is chasing liquidity you will never touch: if you buy once and never look again, IBIT's options depth is worth exactly nothing to you.

The second error is buying the smallest funds without a reason. WisdomTree's BTCW held somewhere around $122–142m in August 2026 and quotes roughly a 0.11% spread — nearly four times IBIT's. DEFI's liquidation showed what the tail looks like when a sponsor decides the revenue no longer covers the audit, the custody and the listing, and shareholders received cash on a date they did not pick, which in a taxable account is a tax bill you did not plan.

The third is confusing the fund with the asset. Everything on this page is a grantor trust registered under the Securities Act of 1933, not a 1940 Act fund, and none of it gives you bitcoin you can withdraw, send or hold yourself. If that distinction matters to you, it is a different decision entirely and we set it out in full at ETF versus owning bitcoin. If you are still deciding whether the structure suits you at all, start with what a Bitcoin ETF actually is and the honest list of risks worth understanding.