On-Chain Metrics ETF Era: What Weakened, What Survives
Lesson 9 of 10 · Advanced
What does a Bitcoin ETF look like on-chain?
A spot Bitcoin ETF is a fund whose shares trade on a stock exchange and which holds real bitcoin with one or more custodians. To understand its effect on on-chain data, you have to separate two markets:
- The secondary market: daily buying and selling of ETF shares between investors on the exchange. These trades move no coins on-chain.
- The primary market: when demand for, or supply of, shares exceeds a threshold, "authorized participants" create new shares or redeem existing ones. Only at this stage does bitcoin move into or out of the custodian's wallet, in cash or in kind depending on the fund's structure.
The scale is not small. According to Farside data cited in a ryder.id report, net assets of US spot Bitcoin ETFs were about $106 billion as of October 9, 2026. The same source reports roughly $5.4 billion of net outflows in the first half of 2026 and about $2.4 billion of inflows in one week in late September 2026. A large share of institutional demand now passes through a channel that on-chain data sees only indirectly and with a delay.
Indicators that weakened
| Indicator | Why it weakened |
| Exchange reserves and netflow | Coins move from exchanges to custodians, and this "outflow" is mistaken for holders withdrawing. Many trades never pass through public exchanges at all. |
| Active addresses | Millions of ETF shareholders do not have even one address; activity has moved to custody, exchanges and layer 2. |
| Whale count and balances | Custodian wallets are large and busy, and the smallest reshuffle changes the whale count. |
| Coin age and cohorts | Moving to custody resets coin age to zero; coins held in ETFs count as "short-term" for months. |
| Label-dependent indicators | Any error in a custodian label passes directly into the result. According to CryptoDaily, Coin Metrics' 19-month revision of ETF wallet data in September 2026 is a clear example. |
"Weakened" does not mean useless; it means these can no longer be read without adjustment and without checking custody, and pre-ETF thresholds are probably not valid.
Indicators that have held up better
- Realized cap and realized price: because they do not depend on exchange labels and simply sum the value of coins at their last movement. When a coin is bought for an ETF and moved, it is recorded at the day's price, which is a fair approximation of a new buyer's cost basis. The remaining weakness: internal custodian reshuffles also revalue the coin.
- MVRV as a state variable: its logic holds, but its thresholds must be read with caution and relative to the new era.
- Entity-adjusted SOPR: to the extent custodians are labelled correctly, it still describes profit-taking behaviour.
- Supply in profit and loss: the share of supply above or below the price of its last movement, regardless of where the coin is held.
- Stablecoin supply and miner flows: less affected by ETF custody, though they have their own limits.
The common pattern: indicators built on valuing coins are more robust; indicators built on counting addresses and labelling location are more vulnerable.
Data to add to the toolkit
- Daily ETF flows: net creations and redemptions for each fund. Read them next to price: an inflow on falling days means something different from an inflow on rising days. This data is usually compiled after the US trading day ends, is sometimes completed late for some funds and may be corrected afterwards, so do not treat a single day's figure as final; look at the multi-day trend.
- CME futures and the basis: part of ETF inflow is "cash-and-carry" trading, meaning buying the ETF and simultaneously selling CME futures to capture the price gap. That inflow is not directional demand and can reverse when the basis narrows. So do not read every ETF inflow as optimistic institutional buying.
- The Coinbase premium: the price difference on a US exchange versus global exchanges, which gives a regional picture of spot demand. This too is market sentiment, not a forecast.
- Open interest and funding: to understand how much of a price move was built with leverage.
A combined framework in the custody era
A logical order for reading the market today:
- Long-term state: MVRV, realized price and supply in profit, understanding that old thresholds may not repeat.
- Capital flow: change in realized cap, stablecoin supply and net ETF flows, all as multi-week trends.
- Quality of flow: does ETF inflow come with a high basis (possibly a hedged trade) or with a normal basis?
- Holder behaviour: SOPR and cohorts, checking spikes by hand to remove custody moves.
- Leverage and execution: funding, open interest and CVD, to understand how this overall picture turns into trading in the short term.
None of these layers is a signal. The goal is to build a picture in which you know which parts are reliable and which are ambiguous.
Common mistakes
- Reading the fall in exchange reserves as a supply shock without allowing for transfers to custodians.
- Applying pre-ETF thresholds to MVRV, cohorts and whale counts.
- Reading every ETF inflow as directional demand without checking for basis trades.
- Assuming daily ETF share trading appears in on-chain data.
- Trusting historical label data without noting provider revisions.
What do volume and data say here?
In the ETF era, part of bitcoin's "real volume" happens on the stock exchange and in CME futures, not on-chain. Practical use:
- Put daily ETF flow next to volume and spot CVD on the major exchanges. A strong ETF inflow together with negative CVD on global exchanges means the two groups' demand is not pointing the same way.
- CME Bitcoin futures, which according to a CME announcement have traded 24 hours a day, 7 days a week since May 29, 2026, are the closest regulated-market equivalent to gold futures (GC), and their volume and order flow can be analysed with the same logic.
- Watch changes in the CME basis alongside ETF flow to estimate the share of hedged trades.
On-chain data still shows the ownership layer, but without this traditional-market data the picture of today's demand is incomplete.
📊 See real gold futures volume behind these ideas: free order flow course · today's gold volume map
Frequently asked questions
How did Bitcoin ETFs change on-chain analysis?
A large share of demand now goes through ETF shares and custodians, which show up on-chain only at creation and redemption. Indicators that depend on address counts and exchange labels have weakened.
Which on-chain indicators still work?
Indicators based on valuing coins, such as realized cap and price, MVRV and supply in profit, are more robust, though their historical thresholds should be read with caution.
Does money flowing into an ETF mean optimistic institutional buying?
Not always. Part of it can be a basis trade hedged by selling CME futures. Read ETF flow together with the basis and price.
Education only, not a recommendation to buy or sell; we never give signals. Price examples are illustrative. Test any method on a demo account first, with strict risk management.