Flows & data
Bitcoin ETF flows explained
What an inflow actually is, how trackers calculate the daily numbers, the four different reasons money arrives, and the three questions flow data cannot answer.
Daily Bitcoin ETF flow figures are quoted more widely than almost any other number in crypto and understood by a much smaller group. Headlines treat them as a sentiment index. They are a bookkeeping measurement of one specific thing, and knowing exactly what that thing is changes how much weight you should give it.
What a flow actually is
A flow is not trading volume. When you buy IBIT shares from another investor, money changes hands between two people and the fund is unaffected — no bitcoin is bought, no shares are created, and no flow is recorded.
A flow happens only when the share count changes. An authorised participant delivers bitcoin or cash to the trust and receives newly created shares: that is an inflow, and the trust's bitcoin holdings genuinely rise. Run it in reverse and shares are retired, bitcoin leaves, and it is recorded as an outflow. The mechanics are in creation and redemption.
So flow measures demand strong enough that the existing share supply could not absorb it. That is a genuinely interesting signal. It is just a narrower one than "investors are bullish".
How the daily numbers are produced
- Take the change in shares outstanding
Each issuer publishes shares outstanding daily. The day-on-day change is the number of shares created or redeemed.
- Multiply by net asset value per share
Using that day's struck NAV converts a share count into a dollar figure.
- Sum across every fund
Fund-level figures are aggregated into a category total. This is what "Bitcoin ETFs saw $242 million of inflows" refers to.
- Publish after the close
NAV is struck once daily, so flow data is inherently backward-looking. There is no live flow feed, whatever a dashboard implies.
Reading a daily flow table
A typical published line looks like the Farside Investors table for 27 August 2026: total net flow $242.3 million, with IBIT +$277.6m, FBTC −$83.6m, ARKB +$29.7m, BITB +$21.7m, BTC +$11.7m, MSBT +$6.7m, HODL +$5.7m, GBTC −$27.2m, and several funds flat at zero.
| Observation | What it means |
|---|---|
| Positive total, mixed funds | Money entered the category while rotating between funds. Common and unremarkable. |
| IBIT taking most of the inflow | The persistent pattern — 83% of Q1 2026 inflows went to IBIT or FBTC. Liquidity attracts liquidity. |
| GBTC negative again | A years-long structural trend, not news. Holders leaving a 1.50% fee — see GBTC. |
| Several funds at exactly zero | No baskets created or redeemed. Normal for small funds; it does not mean nobody traded them. |
| One fund out of step with the rest | Usually a single large allocation or a rebalance, not a view on bitcoin. |
Single-day figures are noisy. Weekly and monthly aggregates carry far more information than any one session.
Four reasons money arrives, only one of them a bet
1. Directional conviction. Someone thinks bitcoin goes up and buys. This is the interpretation headlines assume, and it is a real component.
2. Basis trading. When CME futures trade above spot, an institution buys spot exposure through the fund and simultaneously shorts the futures, locking the premium with no view on direction whatsoever. These creations look identical in the data and mean something entirely different. See arbitrage.
3. Allocation mechanics. Model portfolios rebalancing, advisers implementing a new target weight, retirement contributions landing on schedule. Flows that reflect a calendar rather than an opinion.
4. Migration between funds. A holder leaving GBTC for the Mini Trust generates an outflow and an inflow on the same day. Category totals net this out; fund-level data does not, which is why GBTC's persistent outflows are not a bitcoin signal.
Three things flows cannot tell you
Who bought. Flow data has no attribution. Retail, institutional, adviser-directed and hedge-fund basis trades are indistinguishable. Quarterly 13F filings offer a partial, delayed view of large holders and nothing about the rest.
Why they bought. Covered above — conviction, hedging, rebalancing and migration produce the same data point.
What happens next. This is the important one. The category attracted roughly $52.8 billion in cumulative net inflows and bitcoin still traded near $77,000 in September 2026, well below its highs. Flow and forward return are not reliably linked, and the evidence is examined properly in do ETF flows move the price.
Watch the asset, not the wrapper
Flow data is a measurement of fund plumbing. If you want exposure to bitcoin without a once-daily NAV and a 4pm close, a licensed exchange trades it continuously — CEX.IO is registered with FinCEN as a money services business, licensed for money transmission across US states, and authorised in Gibraltar as a DLT provider.
The record so far
Since launching on 11 January 2024, US spot Bitcoin ETPs have accumulated approximately $52.8 billion in cumulative net inflows, supporting roughly $84.3 billion in net assets as of mid-August 2026. The gap between those two figures is price: assets are higher than cumulative inflows because bitcoin appreciated over parts of the period, and the gap narrows when it falls.
Ether funds tell a similar story on a smaller scale — about $13.0 billion of cumulative inflows against roughly $15.2 billion in assets. August 2026 was the strongest month for ether fund demand since the products found their footing, with a nine-to-ten session streak adding $1.42 billion from 17 August and a single-day figure of $225.8 million on 27 August — the best since October 2025. ETHA alone led one week with $567 million.
Two patterns are stable enough to rely on. Concentration: IBIT and FBTC take the overwhelming majority of new bitcoin money. And GBTC's persistent bleed, which has continued for years and reflects fee arbitrage rather than sentiment.
Our take from the desk
We look at flows weekly rather than daily, and we look at them to understand fund competition rather than to predict price. As a read on which wrapper is winning, the data is excellent — it explains the fee war and the concentration precisely. As a price signal it has disappointed everyone who has leaned on it, including in 2026, when inflows continued and the price did not cooperate.