Skip to content
← All posts

What does the average cost basis of Bitcoin ETF holders mean, and why is it not guaranteed support?

In September 2026 bitcoin moved above the estimated average cost basis of US spot ETF holders, roughly $81,722. How the level is calculated, why breakeven selling can cluster around it, and where the metric misleads.

Candlestick chart crossing a dashed level marking the Bitcoin ETF holders' average cost basis near .7K

The average cost basis of Bitcoin ETF holders is an estimate of the price at which the bitcoin now held by US spot funds was bought on average. Bloomberg Intelligence put it at roughly $81,722 as of September 21, 2026. Above that level the typical fund investor is in profit; below it, underwater. It is a sentiment reference point, not guaranteed support.

What happened to Bitcoin ETF holders in September 2026?

In the week of September 21–25, US spot Bitcoin ETFs took in $2.4 billion in net inflows, the biggest week since October 2025. September 21 alone brought about $999 million. That flipped the year back to positive: +$934 million year to date, after the funds had been around $5.8 billion in the red in mid-July.

At the same time, bitcoin climbed to $86,000 and briefly above $87,000. According to Bloomberg Intelligence ETF analyst James Seyffart, this put the average ETF holder back in profit for the first time since January. By September 25 the price had pulled back to $84,063, still about 3% above the estimated cost basis.

How is ETF cost basis calculated?

Funds do not publish the average purchase price of their investors, so it has to be estimated from public data. A typical approach: take daily dollar flows (published, for example, by Farside Investors), divide each day's inflow by that day's bitcoin price to get the number of coins bought, then compute a weighted average across all purchases.

One detail matters here. The $81,722 figure is a buys-only cost basis: it counts inflow days and does not subtract outflow days. If redemptions are included, the net cost basis, per CryptoSlate, sits closer to the current market price. The same metric can produce different numbers depending on the method.

The average also keeps moving. Every large inflow at a price above the average pulls it higher. So the $2.4 billion that entered at $84,000–87,000 shifts the reference level up rather than simply "testing" the old one.

Why does the market care about this level?

The main reason is behaviour around breakeven. An investor who has been underwater for months may want to exit "at even" as soon as the position stops losing money. CryptoSlate flags exactly this risk: holders can reduce exposure near their average entry without realizing a loss.

The scale matters. Unrealized gains of ETF holders peaked at $86.32 billion on October 6, 2025 and turned into an unrealized loss of about $780 million on September 18, 2026. The funds hold about $108.4 billion in net assets. When a pool that size sits near zero profit, even a small share of breakeven selling becomes a visible flow.

The second effect works the other way. While price holds above the cost basis, most holders are not under loss pressure, and panic is less likely. That is why the level is often called "support". But it describes behaviour, not a law of the market.

What do the flows show beyond the average price?

It is more useful to watch the flow trend than a single level. Day by day last week, inflows shrank: $999 million, $714.7 million, $347 million, $190.6 million and $134.5 million. Still positive — seven straight days of inflows totalling about $2.98 billion — but demand was weakening while price slipped from $87,000 to $84,000.

Context matters too. A week earlier, on September 15–16, the funds lost $746.3 million in two days after a failed Senate cloture vote on the Clarity Act. ETF flows reverse quickly, so cost basis is a snapshot, not a fixed number.

What are the limits and risks of this metric?

  • It is an estimate, not a filing. Analysts do not fully disclose the methodology, and different approaches (buys-only vs. net) give different levels.
  • An average hides the distribution. Some bought earlier and cheaper, others near the October 2025 peak, when bitcoin traded at $126,296. For the latter group, $81,722 means nothing: they are still underwater.
  • Not every ETF holder cares about entry price. Part of the holdings belongs to basis traders who own the ETF and short futures at the same time. They sell based on the spot–futures spread, not on breakeven.
  • ETFs are only part of the market. Exchanges, miners, long-term holders and derivatives also set the price. Strong fund inflows do not cancel selling elsewhere.
  • The level can fail. If price drops below the average cost basis, "support" can turn into a selling zone: holders are underwater again, and some will choose to cut risk.

How to use ETF cost basis without misreading it?

Treat it as a map of one investor group's sentiment. Compare three things: where price sits relative to the estimated cost basis, whether daily inflows are growing or fading, and how the funds' net assets change. Agreement between these signals tells you more than any single level.

Do not treat $81,722 as a reversal point. In June 2026 fund net assets fell to $71 billion, and the cycle's price low was around $58,642 — no "average level" stopped those moves. The metric explains why selling pressure or confidence may appear, but it does not predict price.

Sources

  • The Block — https://www.theblock.co/news/markets/2026-09-26-bitcoin-etfs-turn-positive-for-2026-with-2-4-billion-weekly-inflow-their-largest-since-october-416944
  • Decrypt — https://decrypt.co/379391/bitcoin-etfs-notch-seven-day-winning-streak-as-2026-flows-turn-green
  • Cointelegraph — https://cointelegraph.com/markets/bitcoin-etf-inflow-2-days-holder-profitable
  • CryptoSlate — https://cryptoslate.com/bitcoin-hits-86000-putting-etf-investors-back-in-profit-after-86-billion-wipeout/
  • Farside Investors — https://farside.co.uk/btc/

This article is for information only and is not individual investment advice. Trading crypto carries the risk of losing your funds; results on historical data do not guarantee future results.

Read this next

Why Is Bitcoin's Four-Year Cycle Called a Self-Fulfilling Prophecy?

Check the strategy against your own data

ACARO is a terminal that executes a strategy on your own exchange account. Parameter search and backtesting on history are part of the subscription.

See pricing