Bitcoin ETF Flows, Whale Wallets, and Futures: A Forensic Reconciliation Guide

Short answer: ETF inflow, fund creation, custodian transfer, exchange withdrawal, and futures positioning measure different parts of the Bitcoin market. None proves the others. Since US spot Bitcoin ETPs can use cash or in-kind creations, a positive flow estimate may involve open-market buying, delivery of Bitcoin already held by an authorized participant or its designee, or inventory and hedge adjustments. Reconcile shares, NAV, fund BTC, creations/redemptions, custody movements, and derivatives data before claiming a squeeze.

The original May 2026 article claimed $2 billion of April inflows, 26,000 BTC absorbed by two funds, 300,000 BTC withdrawn from exchanges, 20,000 dormant coins sold, 50,000 miner BTC distributed, and a 60/40 exchange-versus-OTC sourcing split. It provided no daily flow table, fund filings, transaction set, address labels, or methodology. Those figures have been removed.

This revision preserves the URL as a correction-led audit method. It explains what each dataset can establish, how to avoid double counting, and why “ETF demand neutralized derivatives volatility” is not a valid conclusion from flows alone.

Correction Ledger

Earlier claim	Evidence required	Audit result
$2 billion entered spot ETFs in April 2026	Complete daily fund table with source, revisions, and net total	Not supplied; removed
IBIT and FBTC absorbed 26,000 BTC	Shares and holdings change or issuer records	Not supplied; removed
ETF custodian balance rose 55,000 BTC	Verified fund-specific addresses and no double counting	Not supplied; removed
Exchange reserves fell by 300,000 BTC	Provider, covered venues, cluster method, snapshots	Not supplied; removed
Dormant holders sold 20,000 BTC	UTXO age analysis plus destination and ownership caveats	Not supplied; removed
Mining pools sold 50,000 BTC	Labeled pool/miner clusters and subsequent sale evidence	Not supplied; removed
60% of ETF BTC came from exchanges	Trade and settlement provenance	Not observable from public chain alone
40% came from OTC desks	Named desk inventory or transaction evidence	Unsupported; removed
ETF mechanism automatically buys against futures shorts	ETF premium/discount and creation orders, not futures direction alone	Mechanically false
Institutional demand created a one-way supply regime	Persistent net demand, limited supply, depth, and no reversal	Unsupported absolute claim

An audit should not replace false precision with vague optimism. It should show which claim cannot be tested and which evidence would change that conclusion.

Six Quantities Commonly Called “ETF Flow”

1. Secondary-Market Share Turnover

Investors trade ETF shares with one another on an exchange. High dollar volume can occur with no net creation or redemption and no change in trust Bitcoin. It measures liquidity and interest, not new capital into the fund.

2. Estimated Net Fund Flow

Data providers often estimate flow from the change in shares outstanding multiplied by NAV, adjusted under their methodology. This is useful but can be revised and should not be confused with executed spot volume.

3. Gross Creations

Authorized participants create baskets of new ETF shares. The trust receives cash or Bitcoin depending on the product's permitted process and the specific order.

4. Gross Redemptions

Authorized participants return baskets and receive cash or Bitcoin. Net flow can hide large gross activity if creations and redemptions offset.

5. Trust Bitcoin Holdings Change

The quantity of BTC held changes through creation/redemption activity and decreases through sponsor-fee payments or other permitted expenses. Price appreciation changes dollar AUM without changing BTC quantity.

6. Custodian Wallet Movement

Bitcoin may move among cold, hot, segregated, omnibus, trading, or sub-custody addresses without changing fund holdings or investor exposure. Wallet movement is operational evidence, not automatically a creation.

Observation	May change trust BTC?	Proves open-market BTC purchase?
ETF share trades between investors	No	No
Cash creation	Usually, after trust/agent acquires BTC	Purchase process exists, timing/source needs records
In-kind creation	Yes	No; BTC can be transferred from existing inventory
Cash redemption	Usually reduces BTC after sale	Shows sale process, timing/source needs records
In-kind redemption	Yes	No open-market sale required
Custodian internal transfer	No	No
BTC price rise	No	No

In-Kind Creations Changed the Plumbing

The SEC approved in-kind creations and redemptions for crypto ETPs in July 2025. Before that approval, recently launched US spot products were limited to cash processes under their original structure. In-kind capability permits authorized participants or designated parties to deliver Bitcoin for shares or receive Bitcoin when redeeming, subject to each fund's documents and operations.

This matters for market interpretation:

A cash creation can cause a fund or its trading agent to acquire Bitcoin.

An in-kind creation can move Bitcoin already owned or sourced by a market participant into the trust.

The participant may have bought that BTC earlier, hedged with futures, borrowed it, or sourced it OTC.

Public fund flow data rarely identifies the ultimate seller or venue.

A redemption can distribute BTC in kind without a contemporaneous market sale.

Bitwise's 2025 Form 10-K says its Bitcoin ETF may create and redeem shares in kind for Bitcoin or in cash. It describes baskets of 10,000 shares and ties the basket amount to Bitcoin attributable to each share. That is a specific legal mechanism, not evidence that every basket uses one method.

How ETF Arbitrage Actually Works

ETF arbitrage seeks to keep share price close to net asset value.

Shares Trade Above NAV

If ETF shares trade at a premium large enough to cover costs, an authorized participant can create shares and sell them. The creation requires delivering cash or Bitcoin under the fund's process. More share supply tends to narrow the premium.

Shares Trade Below NAV

An authorized participant can buy discounted shares, redeem a basket, and receive cash or Bitcoin. Removing shares tends to narrow the discount.

Futures Selling Is a Different Market

A trader selling CME futures can pressure futures prices and influence spot through arbitrage, hedging, and expectations. It does not directly create an ETF premium that forces a new ETF basket.

The chain may be:

1.Futures fall relative to spot.

2.Arbitrageurs trade futures and spot or ETFs.

3.ETF shares develop a premium or discount relative to NAV.

4.AP economics justify a creation or redemption.

5.Cash or Bitcoin moves under the fund process.

Every step is conditional. The old claim that a $100 million futures short automatically causes $100 million of ETF spot buying skipped the actual price relationships and decisions.

Worked Cash-Creation Example

Assume:

ETF NAV per share: $40.00

Market price: $40.20

Basket: 10,000 shares

Gross market premium: $2,000 per basket

Creation, financing, trading, and hedging costs: $1,200

Potential arbitrage margin before risk is $800. If an AP places a cash creation, the trust's agent may need approximately $400,000 of Bitcoin exposure for the basket, adjusted for fees and exact basket terms.

But the flow provider may report about $400,000, the exchange may report $402,000 of ETF share sales, and the spot venue may record a different BTC notional because of execution timing. Adding all three would triple count one economic process.

Worked In-Kind Creation Example

Assume the same basket requires 6 BTC.

AP or designee delivers 6 BTC to the trust.

Trust issues 10,000 shares.

AP sells or holds the shares.

Trust BTC rises by 6.

The 6 BTC may have been purchased that day, held as inventory, borrowed, received through an OTC trade, or offset by a short elsewhere. The trust's holding increase is known; the original source is not unless trade records or counterparties disclose it.

Calling all 6 BTC “new spot demand” is an inference.

The Fund-Holdings Reconciliation

For each fund and period:

Closing BTC = opening BTC + BTC received through creations - BTC delivered/sold for redemptions - BTC used for fees/expenses ± other disclosed adjustments

Dollar AUM is:

AUM = BTC quantity × valuation price + cash/receivables - liabilities

AUM can rise with zero net flow when Bitcoin rises. Flow can be positive while AUM falls if Bitcoin price drops enough.

Reconciliation Table

Field	Source priority
Shares outstanding	Issuer daily page, filing, transfer-agent data
BTC quantity	Issuer holdings disclosure or financial statement
NAV and valuation price	Issuer and benchmark documentation
Creations/redemptions	Filing, daily basket records where available
Sponsor fee paid in BTC	Filing and holdings reconciliation
Estimated flow	Consistent specialist dataset with methodology
Wallet transfers	Blockchain data with documented labels

If closing holdings do not reconcile, do not fill the gap with a whale story. Record it as an unresolved timing, data, fee, or scope difference.

AUM Is Not Flow

Suppose a trust holds 100,000 BTC at $60,000, worth $6 billion. Bitcoin rises to $66,000 with no share creation or redemption.

New AUM: $6.6 billion

AUM change: +$600 million

Net flow: approximately zero, before fees and other assets

Now suppose investors create shares corresponding to 2,000 BTC while Bitcoin falls from $60,000 to $55,000.

Opening BTC value: 100,000 × $60,000 = $6.0bn

Closing holdings: 102,000 BTC

Closing BTC value: 102,000 × $55,000 = $5.61bn

Holdings increased by 2,000 BTC

AUM fell by $390m

Fund demand can be positive while AUM declines. This is why issuer BTC quantity and share count matter.

ETF Filings Are Better Than Wallet Guessing

Fidelity's 2025 annual report states that its custodian holds the trust's Bitcoin and that quantity changes reflect purchases/acquisitions, fees, and redemptions. Its March 2026 Form 10-Q reports financial statement activity, including realized results on Bitcoin sold for redemptions.

Bitwise's annual report describes in-kind and cash processes and includes audited Bitcoin balances. KPMG's report says audit work included evaluating transaction authorization, trust control through the third-party custodian, and evidence from the public blockchain.

These disclosures establish legal trust holdings better than a social-media address label. They still do not provide a public tick-by-tick map of every sourcing trade.

Custodian Concentration and Address Attribution

Several spot Bitcoin funds may use the same custodian. One custodian can also serve exchanges, corporations, private funds, and individuals. A large Coinbase Custody transfer cannot be assigned to IBIT, BITB, or another trust merely because those funds use Coinbase.

Address attribution needs:

first-party publication or cryptographic proof;

fund-specific segregated account evidence;

a transaction matching disclosed creation/redemption amount and timing;

subsequent behavior consistent with the fund;

no conflicting service-client explanation.

Even then, beneficial ownership and key control differ. A custodian may control keys for assets legally owned by a trust.

Use the <a href="/insights/whale-wash-2026-institutional-custody-rotation">Bitcoin whale transfer evidence audit</a> for the full label-confidence method.

Exchange Reserves Are Model Outputs

Exchange reserve providers cluster addresses believed to belong to trading venues. Their totals can differ because they cover different exchanges, discover addresses at different times, and classify institutional custody or internal wallets differently.

An “exchange outflow” can be:

customer self-custody withdrawal;

transfer to an ETF or institutional custodian;

exchange cold-wallet reshuffle;

movement to an affiliate;

OTC settlement;

change-output misclassification;

address-label revision.

An outflow does not prove the BTC is unavailable for sale. The holder can trade through OTC desks, derivatives, lending, or return the coins to an exchange.

When citing an exchange-balance number, publish provider, venues, snapshot time, entity-adjustment method, and revision policy. Never merge absolute levels from one provider with changes from another.

Dormant Coins: Movement Is Not Sale

UTXO age measures time since an output last moved, not the owner's holding period or cost basis. A five-year-old output can move because of:

key rotation;

estate planning;

custody migration;

UTXO consolidation;

collateral transfer;

OTC delivery;

exchange deposit;

direct sale.

To support a sale interpretation, follow the output into a labeled exchange or counterpart and examine subsequent transactions. Even an exchange deposit indicates ability or intent to use the venue, not confirmed execution.

The original “20,000 dormant BTC sold to ETFs” combined two unproven attributions. No public blockchain field links an old UTXO to an ETF basket and beneficial seller without corroboration.

Miner Flows Need Entity and Time Controls

Mining pool payout addresses are not identical to mining-company treasury wallets. A pool can distribute block rewards to participants without selling. Miners can transfer BTC to custodians, lenders, OTC desks, exchanges, or internal addresses.

For miner selling analysis:

1.Separate pool coinbase and payout flows from miner-company holdings.

2.Identify public-company disclosures where available.

3.Track exchange deposits, not gross wallet movement.

4.Compare BTC sold with BTC produced and opening treasury.

5.Account for debt, power costs, hedges, and hardware purchases.

6.Avoid summing the same BTC through pool, miner, custodian, and exchange hops.

Our <a href="/insights/bitcoin-mining-economics-2026">Bitcoin mining economics model</a> explains how cash costs and balance sheets influence actual treasury sales.

Futures Positioning: What CFTC Data Can and Cannot Show

The CFTC Commitments of Traders reports publish positions for CME Bitcoin and Micro Bitcoin futures by trader categories under defined reporting methods. They show long, short, and spreading positions at a weekly snapshot.

They do not reveal:

the motivation for a position;

linked spot, ETF, option, or offshore hedges;

intraday changes after the Tuesday snapshot;

whether one short is directional or basis arbitrage;

all crypto derivatives outside regulated US reporting.

A large futures short can belong to a market maker hedging ETF or spot inventory. Calling it an attack on Bitcoin confuses gross position with net economic exposure.

Basis Trade Example

A fund buys $10 million of spot Bitcoin or ETF exposure and shorts $10.3 million notional of futures trading at a premium, seeking convergence.

The CFTC data may show a large short. Economically, the strategy is approximately price-neutral before basis, financing, and execution risk. If analysts count the spot leg as bullish demand and the futures leg as bearish pressure independently, they misstate the combined position.

ETF Options Add Another Hedge Layer

Options on Bitcoin ETFs can create dealer hedging flows. A dealer short calls may buy ETF shares as price rises and sell as price falls; a dealer's gamma profile depends on the full option book. Put buying can be protection for a long ETF holder rather than outright bearish exposure.

To analyze options:

use open interest by strike and expiry;

distinguish volume from outstanding positions;

estimate dealer positioning cautiously;

include implied volatility and skew;

avoid assuming every purchased put is speculative shorting;

track whether hedges use shares, futures, or spot.

ETF flows can coexist with short futures and put demand because investors and intermediaries hedge.

Why ETF Inflows Cannot “Neutralize Volatility” Permanently

Volatility is a function of new information, leverage, liquidity depth, market concentration, options, collateral, forced flows, and investor behavior. Persistent spot demand can absorb sellers and reduce some downside pressure. It can also increase upside gaps if available liquidity is thin.

ETF flows can reverse. Redemptions can distribute or sell BTC. A concentrated custodian or AP network can create operational dependencies. The same mechanism that transmits inflows can transmit outflows.

Absorption Ratio

A useful descriptive calculation is:

Demand absorption ratio = estimated net BTC acquired by funds ÷ estimated net BTC offered by measured seller cohorts

The ratio is only as reliable as the seller denominator. It cannot include “all sellers” because every executed trade has both buyer and seller and many off-chain holdings are not observable.

Use it for a defined dataset, such as fund holdings increase versus miner-company disclosed sales, not as a market-wide law.

Worked Multi-Dataset Reconciliation

Suppose one week shows:

Estimated ETF net flow: +$600m

BTC average valuation price: $60,000

Approximate flow equivalent: +10,000 BTC

Issuer-reported aggregate trust holdings change: +8,500 BTC

ETF share trading volume: $8bn

Labeled custodian gross inflows: 30,000 BTC

Exchange-cluster net outflow: 12,000 BTC

CME reportable futures shorts increase by 4,000 BTC equivalent

Incorrect conclusion:

“ETFs bought 10,000 BTC, whales removed 12,000, custodians absorbed 30,000, so 52,000 BTC left supply while shorts added 4,000.”

Correct audit questions:

1.Why does estimated flow imply 10,000 BTC while trust holdings rose 8,500? Check timing, price, fees, excluded funds, and in-kind activity.

2.How much of 30,000 custodian inflow is internal movement or the same 8,500 fund BTC?

3.Does the 12,000 exchange outflow feed the custodian wallets, making it another view of the same transfer?

4.Are futures shorts hedging the created ETF shares or unrelated positions?

5.Did share volume reflect turnover rather than creations?

The only strong initial conclusion is that trust holdings rose 8,500 BTC under the issuer dataset. Other figures describe overlapping plumbing until reconciled.

The CryptosEyes ETF-Whale Audit Table

For every claim, fill one row.

Field	Required entry
Claim	Exact quantity, asset, and direction
Period	Start/end with time zone
Entity scope	Funds, exchanges, wallets, futures contracts included
Primary source	Filing, issuer, CFTC, or blockchain transaction IDs
Method	Shares × NAV, holdings delta, cluster query, contract conversion
Revisions	Whether historical values can change
Overlap	Which other metrics may count the same BTC
Attribution	Observed fact versus inferred owner or intent
Alternative explanations	At least two plausible mechanisms
Confidence	Transaction alert, hypothesis, strong attribution, verified event

If the table cannot be completed, narrow the headline.

A Daily ETF Flow Checklist

Fund layer

Compare total and fund-level flows.

Check shares outstanding and issuer BTC quantity.

Identify cash versus in-kind capability and any disclosed order type.

Account for sponsor-fee BTC reductions.

Separate AUM change from net creation.

Market layer

Measure ETF premium/discount and spread.

Compare spot depth and volume across benchmark venues.

Review futures basis, funding, and liquidations.

Check options open interest and expiry concentration.

Note macro events that can dominate fund plumbing.

Chain layer

Use transaction IDs and dated labels.

Remove internal custodian and exchange reshuffles.

identify change outputs and chained transfers.

Do not combine gross and net flows.

Follow outputs before attributing sale or accumulation.

Publication layer

State source and cutoff.

Mark estimated versus reported values.

Describe trust-held Bitcoin precisely as a digital asset rather than borrowing commodity-delivery terminology.

Do not describe futures as fake or paper supply without explaining the contract.

Do not claim permanence from one month.

Frequently Asked Questions

Does an ETF inflow mean the fund bought Bitcoin that day?

Not necessarily. Cash creations can lead to Bitcoin purchases, while in-kind creations can deliver existing BTC. Flow estimates may also use share changes and NAV rather than trade records. Check fund process and holdings.

Does ETF share volume measure inflow?

No. Most share volume is secondary trading between investors. Net creation changes shares outstanding; turnover alone does not.

Can ETF inflows offset futures shorts?

Spot demand can influence price while futures positioning influences basis and hedging, but there is no fixed one-for-one cancellation. A futures short may be directional, hedged, or part of arbitrage.

Do authorized participants buy Bitcoin whenever shorts sell futures?

No. AP creation and redemption economics respond primarily to ETF share price versus NAV and order economics. Futures can affect those prices indirectly, but no automatic rule links a short sale to a fund creation.

Are ETF custodian wallets public?

Some addresses may be disclosed or attributed, but custodians serve multiple clients and can reorganize wallets. Use issuer filings and first-party evidence before assigning an address to a fund.

Does exchange outflow mean Bitcoin is unavailable for sale?

No. It means a provider believes BTC left its labeled exchange cluster. The destination may be custody, another venue, collateral, OTC settlement, or an internal wallet, and the holder can move it again.

What is the best measure of ETF Bitcoin demand?

Use a combination of issuer-reported BTC quantity, shares outstanding, net creations/redemptions, and a consistent flow dataset. No single measure reveals ultimate seller identity or execution venue.

Can ETF outflows force Bitcoin sales?

Cash redemptions can lead to sales. In-kind redemptions can deliver BTC without an immediate open-market sale. Fund documents and order type matter.

Do SEC 13F filings show daily ETF inflows?

No. Form 13F reports certain institutional investment-manager holdings quarterly and with a delay. It can show reported ETF share ownership for covered managers, not daily fund creations or the fund's BTC sourcing.

Conclusion

ETF flows, wallet movements, exchange reserves, dormant coins, miner transfers, and futures positions are valuable datasets. They are not additive by default and do not reveal motive automatically.

The April 2026 one-way supply story failed because it treated estimates as verified holdings, custody transfers as new demand, exchange outflows as removed supply, and futures shorts as an automatic ETF trigger. A proper audit reconciles shares, NAV, BTC quantity, creation method, fees, and filings first. Wallet and derivatives data then explain possible plumbing, with overlap and uncertainty stated.

The strongest market conclusion is often narrower than the loudest one. That restraint is what makes the result reusable.

What to Read Next

Continue with <a href="/insights/bitcoin-etf-flow-impact-analysis-2026">Bitcoin ETF Flow Impact Analysis</a>. It extends this reconciliation into price-impact windows, cumulative demand, GBTC offsets, shares outstanding, and market-depth tests.

CryptosEyes publishes general market research, not individualized investment advice. ETF structures, holdings, custodians, authorized participants, and reporting methods can change.