U.S. spot Bitcoin funds have swung from $524.8 million in net inflows on October 21 to $635.6 million in outflows on October 22, followed by another $137.9 million in withdrawals on October 23.
Summary
- U.S. spot Bitcoin ETFs took in $524.8 million on October 21 and lost $635.6 million the next session, according to Farside.
- Another $137.9 million left the funds on October 23, based on Farside’s latest displayed fund entries.
- The October 19 to 23 total remained positive at $301.2 million despite two consecutive withdrawal days.
- BlackRock’s IBIT and Fidelity’s FBTC accounted for $471.6 million of the October 22 outflow.
- The SEC has permitted in-kind Bitcoin redemptions since July 2025, so an ETF withdrawal need not trigger an immediate fund sale.
Farside Investors’ fund-by-fund table shows the turn across BlackRock, Fidelity, Ark Invest and other issuers, but it cannot identify the buyer who took the other side of a Bitcoin sale.
An investor can sell an ETF share to another investor without a coin changing hands. A market maker can carry shares for a time, or an authorized participant can redeem a block of them for cash or Bitcoin. Only some of those routes require an immediate sale in the underlying market. The daily flow total combines the result at the fund level and leaves the trading path mostly out of view.
Bitcoin traded around $80,000 to $82,000 as the withdrawals arrived. The fund withdrawals coincided with a lower Bitcoin price. Neither figure shows whether $773.5 million worth of coins was sold on public exchanges over those two days. The issuers’ redemption procedures, the timing of hedges and the identities of eventual Bitcoin buyers determine how much pressure reached spot trading.
Two days erased most of a three-day inflow run
Farside recorded $526.7 million entering U.S. spot Bitcoin products on October 19, $23.2 million on October 20 and $524.8 million on October 21. Over three sessions, the funds added $1,074.7 million. Outflows of $635.6 million on October 22 and $137.9 million on October 23 removed $773.5 million, or about 72% of the previous three sessions’ gains. Subtracting the two amounts leaves a $301.2 million net inflow for the five sessions. The arithmetic uses Farside’s rounded daily dollar totals; the October 23 row still displays a dash for one issuer and can be revised.
The change between Wednesday’s $524.8 million inflow and Thursday’s $635.6 million withdrawal was a $1,160.4 million swing in the daily net flow. It was not a $1.16 billion redemption. SoSoValue put October 22 outflows at $635.49 million and October 21 inflows at about $528.4 million in crypto.news coverage of the price decline. Farside’s figures differ slightly, so the five-session calculation here uses Farside throughout.
BlackRock’s IBIT posted a $254.1 million outflow on October 22 and Fidelity’s FBTC lost $217.5 million, Farside’s table shows. Their combined $471.6 million made up about 74% of the day’s net withdrawal. Ark Invest and 21Shares’ ARKB lost $109.4 million, Bitwise’s BITB lost $27.4 million and Grayscale’s GBTC lost $29.5 million. Morgan Stanley’s MSBT had a $2.3 million inflow. A net figure can therefore conceal simultaneous inflows at one fund and redemptions elsewhere.
IBIT lost another $110.3 million on October 23, while FBTC lost $31.8 million. Morgan Stanley’s fund added $4.2 million. A negative daily total tells readers which fund shares were removed after creations and redemptions were netted; it says nothing on its own about the identity or motive of an investor selling shares in the open market. Institutional portfolios, retail accounts, arbitrage desks and market makers can all trade the same shares.
Bitcoin failed to hold a move toward $85,000 during the three-day run of ETF inflows. Lookonchain and Farside then published different daily tallies for the period. Lookonchain’s graphic labeled its figures October 22 without identifying a completed U.S. session, while Farside’s October 21 row listed $524.8 million. The two figures refer to differently labeled reporting snapshots.
An ETF share can change hands without a Bitcoin trade
An ordinary investor sells ETF shares through a broker. A buyer acquires those shares on the stock exchange, and the fund’s Bitcoin balance does not change simply because the trade occurred. The secondary market can process heavy selling if another investor, a dealer or a market maker is willing to hold the shares at an acceptable price.
The next question is what the buyer does with them. A market maker may sell its inventory to someone else, hedge Bitcoin price exposure with derivatives or assemble enough shares to redeem a basket. Authorized participants, the financial firms that deal directly with the trust, can submit those baskets and receive the specified redemption proceeds. Several participants may serve a fund, and a single participant can act for multiple clients. The route from an investor’s sell order to a redemption is therefore neither automatic nor visible in a daily ETF flow table.
BlackRock’s IBIT prospectus filed with the SEC says only authorized participants can redeem baskets. Its trust can make redemptions using partial cash, all cash or, at the sponsor’s discretion, Bitcoin in kind. The document says the choice can depend on participant capabilities and preferences as well as tax considerations. A shareholder who is not an authorized participant cannot demand Bitcoin from the trust directly.
A $100 million IBIT share sale to another investor would raise trading volume without changing the trust’s Bitcoin holdings. No basket needs to be redeemed in that exchange transaction. Shares outstanding can fall if an authorized participant later gathers a basket and returns it to the trust for redemption proceeds. Even then, the proceeds may be Bitcoin or cash, and the counterparty’s next trade is not disclosed by the daily flow number.
Existing ETF shares represent a claim on a trust that already holds Bitcoin. A trade in those shares does not require the trust to buy another coin. Creations and redemptions change the number of shares and the corresponding assets. Secondary trading transfers exposure between owners. Confusing these layers turns an observation about who held the security at the close into an unsupported claim about a simultaneous order in a crypto exchange’s spot book.
The redemption route determines where selling occurs
Cash redemptions and in-kind redemptions can produce different immediate trades. With an all-cash redemption, a trust needs cash to pay the authorized participant and may instruct an execution agent to sell Bitcoin. The execution, its venue and its timing are separate from the public tally of share redemptions. With an in-kind redemption, the trust can transfer Bitcoin to the authorized participant or its designated agent. The trust’s wallet balance falls, but the recipient decides whether to hold, sell, hedge or deliver the coins to another customer.
The Securities and Exchange Commission approved in-kind creations and redemptions for spot crypto exchange-traded products in July 2025. Its decision ended the cash-only requirement imposed when spot Bitcoin products first launched in January 2024. The commission described in-kind processing as a way to reduce costs and improve efficiency. Permission to use it does not establish that every fund, on every day, used that route.
Commissioner Mark Uyeda said the former cash-only structure could force issuers to sell Bitcoin to raise redemption cash. In-kind settlement removes that specific forced sale at the trust, though a recipient of the Bitcoin remains free to sell later. Daily flow dashboards do not report how many of the October 22 redemptions were paid in cash, partial cash or Bitcoin.
BlackRock’s prospectus sets an in-kind redemption order cutoff at 2 p.m. Eastern on the trade date. Its other cash and partial-cash orders follow their stated deadlines, and proceeds are generally expected on the next business day after an accepted redemption request. Market makers may begin managing exposure before a basket is submitted or settle positions later. Matching every October 22 withdrawal dollar to a Bitcoin sale during the same New York trading session would ignore those timing differences.
Some funds have revised their procedures since the SEC’s 2025 decision. An in-kind transfer can separate a fund outflow from a same-day Bitcoin sale by the trust. Each fund’s prospectus describes the options it offers, but public daily flow figures do not say which option it used for a particular October 22 basket.
Who could take the other side?
Every executed Bitcoin sale has a buyer, but a public flow number does not label that buyer. If a trust sells coins for a cash redemption, the direct counterparty may be a trading firm, an exchange customer or a dealer filling an institutional order. A dealer can subsequently sell the position to several customers. If the trust delivers Bitcoin in kind, the authorized participant receives the coins first and can keep them as inventory, use them to cover a hedge, pass them to another client or sell them. The eventual buyers may not transact at the same hour as the ETF redemption.
Price is part of how those counterparties meet. At a quoted price, bids may be insufficient for a seller’s desired size; the next available bids sit lower. A trade clears when a buyer accepts the offered amount and the seller accepts the price. A $635.6 million fund outflow does not tell us how many coins, if any, were offered through public order books, or how deep those books were at each moment. It cannot by itself assign the whole Bitcoin price move to one class of seller.
Glassnode’s October 21 market study described buying around $80,000 to $81,500 after an October 15 dip. Its analysts placed the estimated cost basis of U.S. spot ETF holdings and corporate treasury holdings in that region, with potential selling and liquidation clusters between about $85,000 and $87,500. A cost basis is an average calculated for a cohort, not a standing bid from every member of the cohort. The research showed where buyers had previously appeared; it did not identify who bought the October 22 or 23 redemptions.
The same study cited an October 19 ETF inflow of $653 million on thin volume, a different number from Farside’s $526.7 million for that day’s fund table. The discrepancy appears in published datasets and should not be smoothed away. Provider coverage, preliminary timestamps or methodology may differ; neither source offers transaction-level identities for the buyers on the other side. The five-session calculation in this article consistently uses Farside.
Direct spot demand can come from private clients, exchange users, corporate treasuries or trading firms arbitraging prices across venues. Their motives cannot be reconstructed from the aggregate ETF totals. The ledger shows transfers between addresses, yet exchange and custodian wallets often pool many customers’ coins. An address receiving Bitcoin is not necessarily a newly bullish investor, and an exchange inflow is not automatically a market sell order.
The case for treating outflows as real selling pressure
Redemptions do remove shares and assets from a fund, and the process can still create selling pressure. Someone who takes Bitcoin in kind might sell it to raise dollars, while a fund paying cash might sell the coins itself. Market makers buying ETF shares from departing investors may hedge by selling Bitcoin futures or spot. Selling and hedging can press on price even when the trades do not line up neatly with a single daily fund-flow print.
Bitcoin’s decline and ETF withdrawals occurred together during October 22 and 23. Bitfinex analysts had warned of liquidation risk as futures open interest grew despite falling prices. The October 23 price report put BTC near $81,450 after a retreat from the $84,000 to $85,000 area and recorded more than $400 million in broader crypto liquidations from CoinGlass. Liquidations can add forced sales or buybacks independently of fund redemptions. Weak technology stocks and other risk assets were moving at the same time, according to contemporary market coverage. The available data do not show how much of Bitcoin’s drop came from ETF redemptions.
Farside recorded $773.5 million in net withdrawals over two sessions. Dealers or other investors had to take on the shares or the resulting Bitcoin inventory as the funds shrank. Demand for fund shares weakened, although the public figures do not show how much Bitcoin funds or market makers sold in response.
An authorized participant can sometimes warehouse coins while seeking another counterparty, but inventory has a cost and risk. A rapidly moving Bitcoin price makes that risk larger. If spreads widen or dealers step back, a sale that might have been absorbed quietly in a calmer market can move through available bids more visibly. The size and capacity of those intermediaries on October 22 are not reported in the fund flow table. Price movement is consistent with pressure, though it does not reveal the identity of its source.
Glassnode recorded buyers near $80,000 after the October 15 dip. Their orders on that day say little about how much they would buy during another selloff. Macro news, liquidations and further redemptions can change the price buyers are willing to pay. The estimated ETF cost basis around that level offers no guarantee that it will hold.
The fund totals hide a mixed week
Thursday’s $635.6 million withdrawal was bigger than Wednesday’s $524.8 million gain, but the preceding Monday added $526.7 million. Friday’s $137.9 million loss then reduced, without erasing, the weekly net inflow. The last two sessions therefore show net withdrawals, while the five-session total shows a net inflow. The choice of window changes the description of ETF demand without telling us how much Bitcoin traded outside the funds.
Farside lists a cumulative $57.251 billion in net inflows to the covered funds in its October 23 table. This historical total is not their current Bitcoin holdings or the amount investors would receive if every share were sold. Asset values move with BTC’s price, and Grayscale’s GBTC alone carries roughly $28 billion in cumulative net outflows in the provider’s series. Cumulative dollars flowing into and out of shares answer a different question from coins held at the custodian or current market value of the shares.
The issuer breakdown complicates any claim that all institutions sold in unison. On Thursday, IBIT and FBTC dominated withdrawals while MSBT gained $2.3 million. On Friday, MSBT gained $4.2 million as the larger funds remained negative. Simultaneous directions can reflect different clients, mandates, fees or execution choices; the table cannot say which explanation applies. It shows enough to rule out a single, uniform trade across all funds.
An ETF share’s discount or premium to its net asset value can signal whether the exchange price and underlying holdings have moved apart. Authorized participants may be motivated to create or redeem baskets when the gap covers their costs, helping align the two markets. A small discount alone does not prove a completed redemption, and a large one may persist if trading or settlement becomes difficult. Issuer price and NAV disclosures are useful alongside fund flows, but they are still not a tape of the authorized participant’s Bitcoin trades.
The 2025 SEC approval gave market participants another way to close those gaps. BlackRock’s current prospectus nonetheless leaves the redemption form to the sponsor’s discretion, and authorized participants have no obligation to submit orders. The daily fund table records the outcome but does not disclose each basket’s settlement method or the authorized participant’s trades.
What price and on-chain data can establish
Market data offer fragments of the answer. Exchange trade records identify the time, price and size of transactions, and some venues report whether aggressive buyers or sellers crossed the spread. Order books show posted bids and offers that can disappear or change before they execute. Neither dataset labels a Bitcoin trade as an IBIT redemption or tells readers whether a market maker is closing an earlier hedge.
On-chain transfers can show coins leaving a known custodian address, subject to the quality of address identification and the custodial settlement timetable. A wallet transfer between a custodian and an authorized participant may represent an in-kind redemption, an internal reorganization or another operational movement. Transfers through omnibus addresses require particular care. Without transaction-level confirmation from the fund or a participant, a matching dollar amount and timestamp are suggestive but not proof.
The ETF’s own holdings, shares outstanding and NAV can provide a more reliable fund-level cross-check. A decline in shares outstanding accompanied by lower Bitcoin holdings is consistent with a net redemption. Comparing consecutive reports needs adjustment for the fund’s fee accrual, pricing timestamp and any other assets or liabilities. The count of coins moved is more informative about trust inventory than a chart of share trading volume, but it still stops short of identifying whoever ultimately bought the coins.
Glassnode’s October study adds the holders’ estimated cost basis, the distribution of on-chain coin ages and areas where leveraged positions may be forced out. Such models describe investor positioning, not individual identities. Its observation that bids refilled near the October 15 low is evidence of buying at that point. It does not prove that those same buyers remained there when October 22’s ETF outflow was announced.
At a hypothetical fixed Bitcoin price of $81,000, Farside’s $773.5 million two-day outflow is equivalent to about 9,549 BTC. The calculation divides $773.5 million by $81,000 and measures scale only. Fund share valuations, fees, cash balances and redemption methods vary, so 9,549 BTC is not an observed count of coins sold.
A buyer’s identity remains private even when the flow is public
Public disclosures can show that Bitcoin left a trust; they generally cannot reveal the names behind every share order and subsequent coin trade. Broker records identify customers to their intermediaries, but those records do not appear in Farside’s aggregate table. Market makers can offset positions across several venues or derivatives contracts. A dealer who receives coins in a redemption may already have a customer order for them, avoiding any visible exchange sale at the point of delivery.
The publicly available October 22 and 23 fund data do not name the buyers. Another investor or a dealer may have bought the ETF shares. If a basket was redeemed, an authorized participant could have received Bitcoin and either held it or sold it to a client or on an exchange. Naming a final buyer would require transaction records or a disclosed trade, neither of which is available in the fund table.
Holding an ETF share and buying a coin are different forms of exposure. One person buying shares from a seller can keep the fund’s Bitcoin where it is. Another can receive Bitcoin in kind from the trust and transfer it without touching a public exchange. Direct buyers at a lower price can absorb a later spot sale. Each route can contribute to the same observed combination of ETF outflows and a falling Bitcoin price, which is why a price chart cannot reconstruct the transaction chain by itself.
Farside’s table will show whether net redemptions continue after the next trading close, subject to corrections. Fund holdings and share counts can show which trusts shed assets. Exchange volume, spreads and the direction of active trades can show how Bitcoin markets responded. Keeping those readings separate preserves the distinction between money leaving an investment product and coins changing owners in the market underneath it.
What to watch
- The next completed fund rows: Check Farside’s issuer entries and any revisions to October 23 before treating the $301.2 million five-session net inflow as final.
- Shares and holdings: Compare each issuer’s shares outstanding, Bitcoin holdings and NAV at consistent timestamps to see whether redemptions changed trust inventory.
- Redemption disclosures: Check current prospectuses and any issuer updates for the cash, partial cash or in-kind methods used; permission to redeem in kind does not establish daily usage.
- Spot liquidity: Watch executed volume and bid depth near $80,000 to $81,500, the area where Glassnode recorded buying on October 15, without treating it as a guaranteed floor.
- Price versus fund demand: Compare completed ETF flows with Bitcoin’s trading range and liquidation data, while keeping other market drivers in view.
FAQ
Did Bitcoin ETFs sell $635.6 million of Bitcoin on October 22?
Farside reported $635.6 million in net fund outflows, not an audited total of Bitcoin sold on exchanges. Cash and in-kind redemptions can send the resulting inventory through different hands and on different schedules.
How much left Bitcoin ETFs over October 22 and 23?
Farside’s displayed totals were $635.6 million and $137.9 million, respectively, or $773.5 million combined. One October 23 issuer entry was displayed as a dash when the figures were reviewed, so the later day’s total could change.
Were Bitcoin ETF flows negative for the whole week?
Using Farside’s October 19 to 23 entries, the three inflow days totaled $1,074.7 million and the two outflow days $773.5 million. The difference was a $301.2 million net inflow across five sessions, subject to any revised fund entries.
Who buys an ETF share when an investor sells it?
Another investor or a market intermediary takes the other side of the stock exchange trade. Only authorized participants can later create or redeem shares directly with a trust in baskets.
What is an in-kind Bitcoin redemption?
An authorized participant returns a basket of ETF shares and receives Bitcoin, subject to a fund’s procedures. The SEC permitted that structure for spot crypto products in July 2025; the participant’s later decision to hold or sell the coins is separate.
Can ETF outflows push Bitcoin’s price down?
They can contribute to selling or hedging pressure when funds or intermediaries dispose of Bitcoin exposure. A daily outflow total alone cannot measure that pressure or separate it from other selling, including derivatives liquidations.
Does the $80,000 cost basis guarantee support?
No. Glassnode estimated that ETF and treasury holders’ average costs sat around the $80,000 to $81,500 area and observed bids after an October 15 dip. An average historical cost and a past episode of buying do not guarantee future demand.
Can public data identify who bought the redeemed Bitcoin?
Public ETF tables, issuer disclosures and on-chain records can narrow the route and measure changes in fund holdings, but they do not generally name the final buyer. A direct answer requires transaction-level information from the relevant fund or intermediary. This is educational analysis, not investment advice.
Disclaimer: This article is for information and educational purposes only and does not constitute financial or investment advice. Figures reflect regulatory filings and reporting available at the time of writing and change with each disclosure. Nothing here is a recommendation to buy, sell, or hold any security or asset. Always do your own research. Information is accurate as of October 25, 2026.
