Bitcoin ETF risks: what can actually go wrong

Price risk, above everything else

Everything below this section is a footnote to this section. A spot Bitcoin ETF holds one asset. That asset is among the most volatile things you can buy on a regulated exchange, and it has fallen more than 70% from a peak on multiple occasions in its history — not once, as a freak event, but repeatedly, as a feature of how it trades. The fund wrapper does absolutely nothing about this. There is no hedge, no floor, no diversifying holding, no cash buffer. The trust owns bitcoin, and you own the trust.

The category's own recent history makes the point better than any warning label. Total assets across the US spot funds peaked around $158bn in October 2025 and stood at roughly $99.6bn on 31 August 2026 — down about 37%. The instinctive reading is that investors fled. They mostly did not. Cumulative net flows since launch were still positive at about $54.8bn, and August 2026 was in fact the strongest inflow month of the year at roughly $3.03bn. The assets fell because the price fell. Which is exactly the point: in a fund holding one commodity, assets under management is a price chart wearing a different hat.

Nothing about the ETF structure moderates that. If anything it makes the volatility easier to act on badly, because the position now sits in the same account as your index funds and updates in the same red numbers. People who would have shrugged at a 40% drawdown in a wallet they rarely opened sell it out of a brokerage app at the low. That is a behavioural risk, not a structural one, and it is still the most common way people lose money here.

The risk that dwarfs the others

What the wrapper does not give you

Spot Bitcoin ETFs are grantor trusts registered under the Securities Act of 1933. They are not registered investment companies under the Investment Company Act of 1940, which is the statute behind essentially every other ETF a US investor has ever bought. That is not a defect and it is not hidden — the sponsors say so in their own annual reports — but it does subtract a layer of protection that most buyers assume is present because it always has been before.

Concretely: there is no independent board of directors owing fiduciary duties to shareholders, and no statutory action for an unreasonable advisory fee. If a sponsor charges 1.50% for holding a commodity, nobody inside the structure is obliged to argue with them on your behalf. Your only remedy is to sell. There are no statutory leverage limits of the kind Section 18 imposes on registered funds. There is no mandated diversification, no mandated liquidity risk management programme, and none of the affiliated-transaction restrictions that govern a 1940 Act fund's dealings with its own affiliates. In a single-asset trust some of those matter less than they would elsewhere. The fee governance gap is not one of them — it matters a great deal, and the ten-times spread in sponsor fees across funds holding identical bitcoin is what it looks like in practice. The full legal comparison sits on the regulation guide and the structure itself is explained on what is a Bitcoin ETF.

Custody concentration

Roughly 80.8% of all bitcoin held inside US ETFs sits with a single custodian, Coinbase Custody Trust Company. Nothing has gone wrong. Coinbase Custody is a New York-chartered limited-purpose trust company with years of institutional history, the coins are in cold storage, and at this scale the number of firms that could plausibly do the job is small. The concentration is a symptom of an industry that is still young, not of anybody cutting corners.

It is still a shared dependency, and it is worth naming as one. An operational failure, a legal seizure, a prolonged outage or an insolvency at that one firm would touch holders of most funds in the category simultaneously — which is precisely the kind of correlated exposure that diversifying across several ETFs does not fix. If you deliberately spread money across four different sponsors for safety, check the custodian column before you congratulate yourself, because you may well own the same operational risk four times over. Two funds break the pattern: Fidelity's FBTC self-custodies through Fidelity Digital Assets, and ARK 21Shares' ARKB splits across Anchorage Digital Bank, BitGo and Coinbase Custody. Whether that is worth paying for is a judgement, but it is at least a real choice. The issuers and custodians page has it fund by fund.

Closure risk, and it is no longer hypothetical

For two and a half years this category had a perfect survival record, and every risk page including ours had to describe fund closure in the abstract. That ended in August 2026. Hashdex's DEFI stopped trading on 17 August 2026 and completed liquidation later that month, holding about $14.7m and roughly 225 bitcoin at the end. It was the first US spot Bitcoin ETF ever to wind down. Shareholders were paid in cash, not coins.

A liquidation is not a disaster and it is important not to dress it up as one. The process is orderly and supervised: the trust sells its bitcoin, the proceeds are distributed, and nobody loses their investment because the fund closed. The real cost is the timing. You receive cash on a date the sponsor picked, which in a taxable account means a disposal you did not plan, at a price you did not choose, possibly crystallising a gain in a year when you did not want one — or crystallising a loss you would rather have carried. If you then want the exposure back you buy a different fund, pay a spread and reset your holding period. None of that is catastrophic. All of it is avoidable by not owning the smallest fund in a category for no particular reason.

It was not an isolated event either. Six Bitwise option-income crypto funds were wound up on 31 July 2026. And several spot funds in the tail of the category sit in a range where a sponsor has to think seriously about whether the fee revenue covers the audit, the custody agreement, the listing and the legal work. The smallest survivor has been running in the low hundreds of millions. At a 0.25% fee, that is a business generating a few hundred thousand dollars a year of gross revenue against real fixed costs. Sponsors are patient, but they are not infinitely patient.

Fees, and waivers that expire while you are not looking

Sponsor fees in this category run from 0.14% to 1.50% a year for exposure to the same asset, frequently held at the same custodian. On a $25,000 position that is the difference between $35 and $375 a year, every year, in rising markets and falling ones. Because the fee is paid in bitcoin, the coin entitlement behind each share erodes continuously — which means the expensive fund is not merely costing you more, it is holding less bitcoin for you with every passing day.

The more insidious version is the waiver. Several funds launched with introductory fee waivers and most have quietly lapsed. The instructive case is VanEck's HODL, which waived its fee entirely on the first $2.5bn of assets — a genuinely aggressive offer — but wrote the waiver to expire on 31 July 2026 whether or not the fund reached that threshold. It did not: HODL held roughly $1.1bn as the waiver lapsed, about 43% of the way to the target. Everyone holding it started paying the full 0.20% on 1 August without any notice landing in front of them and without doing anything. If you selected a fund partly on the strength of a waiver, go and read the current prospectus today rather than the comparison table that persuaded you. We keep the standing rates current on the fees page.

Spreads, tracking, and the hours you cannot trade

Two costs sit outside the expense ratio and both are real. The first is the bid-ask spread you cross on every trade, which in this category runs from about 0.02% at the most liquid funds to roughly 0.11% at the thinnest we track. If you buy once and hold for a decade that barely registers. If you contribute monthly you pay it twenty-four times over a dozen years, and at the wide end it quietly overtakes the difference in sponsor fees you were optimising. The second is tracking difference — the gap between the fund's return and bitcoin's — which is driven by the fee but also by the trust's own trading costs, cash drag and the reference rate it prices against. Some smaller funds have historically trailed by a good deal more than their headline fee would suggest. The mechanism is on how Bitcoin ETFs work.

Then there is the risk nobody mentions until it bites: bitcoin trades continuously and the ETF does not. The coin market runs every hour of every day, weekends and holidays included. The fund trades during US exchange hours. When something happens at two in the morning on a Sunday — a policy announcement, a large liquidation cascade, an exchange failure somewhere — the price moves without you, and the ETF simply gaps at Monday's open. You cannot sell into it, you cannot buy the dip, you can only watch. Investors arriving from equities consistently underrate this until the first weekend it costs them something. It is one of the clearest arguments for holding at least some exposure as the coin itself, which we set out on ETF versus owning bitcoin.

Leveraged, inverse and income funds are a different category of danger

Everything above concerns plain spot funds. The structured products around them carry risks of a different kind, and they are routinely mis-sold to people who read the headline and not the prospectus.

Leveraged funds target a multiple of bitcoin's daily return and reset every session. Over any period longer than a day, the compounding of daily resets means your return is not two times the period return — in a choppy, directionless market it can be meaningfully worse than that, and in a sustained drawdown the decay is brutal. These are trading instruments with a holding period measured in days. The arithmetic is on leveraged Bitcoin ETFs, and the mirror image, with the added problem that a short position's losses are theoretically unbounded, is on inverse and short funds.

Covered-call and income funds sell options against bitcoin exposure to generate a distribution. The yield headline is seductive and the structure is honest about what it does, but the trade is real: you cap your upside in exchange for premium, which in an asset whose entire historical return comes from a handful of violent rallies is a genuinely expensive thing to give away. And a large part of what these funds pay out is frequently return of capital — your own money handed back, reducing your cost basis rather than representing income earned. Six of these funds from one issuer were wound up in July 2026. Read the distribution breakdown before the yield number on Bitcoin income ETFs.

Regulatory risk has not gone away

The approvals are done and the listing framework has settled — the SEC's generic listing standards of September 2025 made new commodity-trust listings routine rather than case-by-case. But the wider legal foundation for crypto markets in the US is still being built. The CLARITY Act, which would set out market-structure rules and divide jurisdiction between the SEC and the CFTC, passed the House in July 2025 and remains pending in the Senate as of September 2026. It was reported out of Senate Banking in June 2026 and a cloture motion was presented in August, but it is not law.

That unsettledness feeds through in ordinary ways rather than dramatic ones: tax treatment without ETP-specific guidance, custody rules that shifted once already when SAB 121 was rescinded, and broker reporting requirements that are still phasing in. None of it threatens the existence of the funds. All of it means the rulebook you buy under may not be the rulebook you sell under. Our regulation guide tracks the current position and the tax guide covers what is settled and what is not.

What SIPC does and does not cover

This is the most misunderstood line on any brokerage statement, so it is worth being blunt. SIPC protects the account, not the investment. If your broker fails and customer assets are missing, SIPC coverage steps in up to $500,000 per customer, of which no more than $250,000 may be cash. That is the entire scope. SIPC states in its own words that it does not protect against a decline in the value of your securities, and that sentence is not boilerplate — it is the whole distinction. If bitcoin halves and your fund halves with it, the protection is untouched and irrelevant, because nothing went missing. You simply own something worth less.

It is worth adding the comparison, since people reach for it. Bitcoin held on a crypto exchange has neither SIPC nor FDIC protection, and customers in a platform insolvency have historically been treated as general unsecured creditors. Bitcoin held in your own wallet has no protection of any kind, and no recourse whatsoever if you lose the keys. Every option here has a failure mode. The ETF's is the one most likely to be mistaken for having none.

Managing all of this without pretending it away

None of the above is an argument against owning a Bitcoin ETF. It is an argument for owning one deliberately. A handful of practical decisions handle most of the risk that is actually manageable, and the ones that are not manageable — the price, mostly — are handled by owning an amount you can watch fall by 70% without changing your life or your mind.

Size the position first, before you choose the fund. If a 70% drawdown on the position would force you to sell, or would stop you sleeping, the position is too big, and no amount of selecting the cheapest ticker fixes that. Where the option exists, hold it in a tax-sheltered account — an IRA, a Roth or a 401(k) that permits it — because that removes the taxable consequence of rebalancing, of a fund liquidating on you, and of the small annual disposals that arise from the sponsor fee being paid in bitcoin. Prefer the liquid funds unless you have a specific reason not to: liquidity buys you a tighter spread, a more reliable arbitrage loop, a deeper options market and a far lower chance of waking up to a wind-down notice. Treat leveraged and inverse products as short-horizon trades and nothing else. And read the current prospectus rather than a comparison table, including ours — waivers lapse, fees change, benchmarks get switched, and the document filed this quarter is the only one that binds anyone.

If you conclude that the fund's constraints — no coins, no weekends, an annual charge — are the problem rather than the solution, that is a legitimate conclusion and the alternative is straightforward. Buying bitcoin directly removes the sponsor fee and the market-hours limit, and replaces them with the responsibility for keeping it safe. Plenty of people sensibly do both.