Home › Market › Daily brief

Daily Brief — September 6, 2026: $906 million bought 2.85%

By TradingPrimer · Published 6 Sep 2026

Two things happened this week that usually point in opposite directions, and the market ended up more or less splitting the difference. The interesting part is not the flat tape. It is the arithmetic that connects a large, widely reported inflow number to a small price change — because that arithmetic is the single most reliable way a beginner gets the size of a position wrong.

Where the market actually is

Read at 23:18 UTC on 5 September, from the Binance spot and perpetual APIs:

Measured against each asset's own opening price, Bitcoin travelled 0.95% between its high and its low over the full 24 hours, Ether 2.00% and Solana 2.72%. Bitcoin's entire day fitted inside a $758 band. For comparison, the same measurement in Wednesday's brief was 3.57% — the daily range has compressed to roughly a quarter of what it was four sessions ago, and it is compressing into a scheduled event, which is the ordinary shape of the week before a central bank meeting.

Set against the 2 September readings, Bitcoin is up 2.85% from $77,576, Ether up 2.54% from $2,418.51 and Solana up 2.74% from $100.52. Four sessions, three near-identical gains, no drama.

Why it moved: money in, odds up

The week delivered one strongly positive input for crypto and one strongly negative one, in that order.

On 3 September, US spot Bitcoin ETFs recorded roughly $731 million in net inflows — the largest single day since 14 January. BlackRock's IBIT took about $453.96 million of that, roughly 62% of the total, with ARK and 21Shares' ARKB at $137.74 million and Fidelity's FBTC at $74.45 million. Reporting attributed the surge to dovish remarks from Fed Governor Christopher Waller.

On 4 September, the August employment report landed at 162,000 jobs added against expectations in the region of 55,000, with unemployment steady at 4.1% and the gains concentrated in a rebound in leisure and hospitality and local government. A labour market that strong, with inflation still above target, pushed futures-implied odds of a rate hike at the 15–16 September FOMC from roughly 55% to somewhere around 60–62% on the CME's FedWatch tool. Spot Bitcoin ETFs still took a further $175 million that day, and Ether ETFs about $27 million.

So: at least $906 million of ETF demand across two sessions, and a meaningful step up in the probability that the cost of holding dollars is about to rise. Bitcoin's answer to both was 2.85%. US inflation data lands next week and is the next input the same committee will read.

The number most people skipped

Two of them, and they say the same thing.

Binance perpetual funding at the time of reading: BTC +0.003679% and ETH +0.006482%. Those decimals are almost impossible to feel. Funding is paid three times a day, so multiply by 3 and then by 365: Bitcoin longs are paying about 4.0% a year to hold the position, Ether longs about 7.1%. Positive, so longs are paying shorts — but only barely, and well below the levels that mark a crowded leveraged long.

The second number is open interest. Binance carried about 106,369 BTC in perpetual open interest, against roughly 109,224 BTC on 2 September. That is a fall of 2.6% in contracts over four sessions in which price rose 2.85%.

Now watch what happens if you measure the same thing in dollars. At today's price that open interest is worth about $8.49 billion; on 2 September it was worth about $8.47 billion. In dollar terms it looks essentially unchanged — up 0.2% — and the entire story disappears. The dollar figure moved because price moved, not because anybody opened or closed a contract. Counted in coins, which is what a contract is actually denominated in, positions were being closed while spot buyers were arriving. Two different crowds, moving in opposite directions. If the term is new, open interest explains it from zero, funding rate covers the other one, and live readings for all three majors sit on Market Pulse.

So what

The durable lesson is that a flow number is not a multiplier. It is tempting to read "$906 million of ETF buying" and expect a price response scaled to it, and that intuition is where a lot of oversized positions come from. But the Binance BTC/USDT pair alone turned over about $693 million in the last 24 hours, and that is one trading pair on one venue among dozens. Against the total volume Bitcoin transacts in a week, $906 million is a real bid but a modest one, and it is competing with everyone who wanted to sell into it. Flow tells you that demand existed. It does not tell you what price paid for it.

The second lesson is about what a quiet tape ahead of a scheduled event is worth. Funding near neutral and open interest falling means the market is not carrying a crowded leveraged position into the FOMC decision. That is a description of the present, not a forecast — a market can be unleveraged the day before it falls 8%. But it does change what a violent move would mean if one arrives: with less leverage stacked up, a sharp move is less likely to be amplified by forced closing, and more likely to be genuine repricing. The mechanics of the alternative are in anatomy of a liquidation cascade.

The practical version, for anyone holding a position into next week: cheap funding is not a reason to use leverage. 4% a year is a low rent, and a low rent on a position that is too big for your account is still a position that is too big for your account. The size question is settled by the stop distance and the account, not by the carrying cost — position sizing works through the arithmetic, leverage and margin covers why the multiplier is the dangerous part, and the event calendar has the FOMC and inflation dates in your local time.

← Market analysis