The level that finally went

$62,500 has been the number in these pages since Tuesday. It was the line whose loss we said would open the way back toward $60,000. It survived Tuesday's selloff, survived the Fed decision, and survived Thursday's reversal. It went today at 14:10 UTC, with the low printing $62,466.

The hour that did it was the largest of the entire day: $227.6 million at a 44.4% buy share, closing 0.93% lower. The 13:00 hour ahead of it carried $96.1 million at 46.8% and fell 0.76%. Neither was a panic. Both were persistent, heavy, seller led grinding, which is the same texture as Wednesday evening's 19:00 hour and a very different thing from a liquidation cascade.

What makes today unusual is the width. Bitcoin printed $65,409 overnight in the 01:00 hour, which is a new high for the week and above Thursday's $65,176. From that high to today's low is roughly $2,940, the widest single session range of the week. A market that makes a weekly high and then loses a three day floor inside fourteen hours is not a market with a settled view.

The reason was oil, not risk appetite

Here is where the day becomes readable. Stock futures rallied while Bitcoin fell, which looks like a contradiction if you think of Bitcoin as a high beta risk asset. It is not a contradiction once you see what changed overnight.

The United States paused its airstrikes in Iran, and with the break in fighting holding, Brent crude fell below $90 a barrel. Lower oil means lower headline inflation pressure, which means less need for a hawkish Fed, which is straightforwardly good for equities. Nasdaq 100 futures rose about 1.23% and S&P 500 futures were higher.

But that same de-escalation takes a premium out of anything that was being held as an inflation or geopolitical hedge. Recall that the FOMC statement on Wednesday explicitly flagged elevated uncertainty owing in part to the conflict in the Middle East. That uncertainty just got smaller. And the cleanest evidence that this is the correct reading is not in Bitcoin at all, it is in gold.

Over the last twenty hours the taker buy share is 40.3% for Tether Gold and 42.2% for Paxos Gold, with both tokens down on the day. Two independent tokenised claims on the same metal, both sold, at the same time as Bitcoin, while equities rose. That is not a risk-off session and it is not a rotation. It is a hedge unwind.

Notice how thoroughly this inverts Wednesday. Then, the Dow fell 2.1% in its worst session since April 2025 and Bitcoin rose. Today equity futures rallied and Bitcoin fell 3%. The sign of the correlation flipped twice in three sessions, and both times the actual driver was the same variable: the market's expectation for inflation. Bitcoin has been trading off that, not off the appetite for risk.

Nothing was bought, which is the whole breadth story

The fastest way to tell a rotation from a de-risking is to check whether anything was being accumulated. On Tuesday, when Bitcoin sold off, Tether Gold printed a 67.9% buy share. Something was being bought. Today, nothing was.

PairTaker buy share, 20h24h change
Tether Gold40.3%1.39% lower
Paxos Gold42.2%1.48% lower
Ethereum44.4%2.87% lower
Solana45.2%1.72% lower
BNB45.2%0.99% lower
XRP46.8%2.47% lower
Bitcoin46.6%3.04% lower

Every single one below 50%. Not one asset on the board attracted aggressive buying, including the two that exist specifically to be defensive. Bitcoin's own aggregate is 46.6% on $1.27 billion, the first sub 50% daily reading of the week after 52.8% on Wednesday and 52.7% on Thursday.

Sentiment finally moved too, and downward. The Fear and Greed index reads 25, the lowest of the week, with the last six readings at 26, 30, 29, 29, 28 and 25. Total crypto market capitalisation is $2.243 trillion, down 2.27%, and Bitcoin dominance has slipped to 56.2% from 56.6%.

Our two conditions contradicted each other in 65 minutes

This is the part of today that needs writing plainly, because it is a genuine flaw in something we published rather than a market surprise.

Yesterday's article named two conditions. First, that losing $64,744 on a buy share under 45% would mean the whole recovery had been a squeeze. Second, that clearing $65,176 on a buy share above 60% would extend the reclaim rather than cap it. Both conditions were met. They pointed in opposite directions and they fired about 65 minutes apart.

Time (UTC)What happenedVolumeBuy share
00:00 hourClosed below $64,744 at $64,725$42.8M37.2%
01:05Cleared $65,176$19.6M68.6%
01:10Printed the high, $65,409$13.2M60.1%
01:30Reversal underway$5.0M30.9%
14:10Broke $62,500, low $62,466$36.1M46.7%

The bearish condition fired first, at 00:00, and it was correct: a 3% decline followed. The bullish condition fired second, at 01:05, on a genuinely strong 68.6% buy share with real size behind it, and it was wrong. The move did not extend by a single dollar beyond 01:10. Twenty minutes later the buy share was 30.9%.

So the lesson is not that the flow lied. It is that a five minute buy share at the moment of a level break is not a confirmation of anything. It is one of the noisiest readings available, because a level break is precisely where a few large aggressive orders can dominate a small sample. The hourly reading, on a much larger sample, said sellers were in control before the spike and after it. We should not have offered a five minute breakout condition as a confirmation signal, and we will use sustained hourly readings for that purpose going forward.

The 07:00 hour, where someone very large was buying quietly

Before the afternoon break there was an hour that deserves its own paragraph. The 07:00 UTC hour carried $197.7 million at a 30.3% buy share, which means roughly seventy percent of the aggressive flow inside it was selling. Look inside it and the readings are extreme:

  • 07:20, a 6.0% buy share, and price moved 0.13% lower.
  • 07:30, $60.1 million at an 11.5% buy share, and price closed exactly flat.
  • 07:40, $37.2 million at 23.6%, and price moved 0.11% lower.

A 5 minute bar carrying sixty million dollars of which nearly ninety percent is aggressive selling, and the price does not move at all. That only happens when somebody has a very large passive bid sitting in the book absorbing everything that hits it. Whoever that was did not chase, did not advertise, and did not move the price. They simply took the other side.

The honest coda is that the bid was finite. It held that hour, and it held the next several, and by 14:00 it was gone or filled and $62,500 broke. But that hour is the single clearest example this week of the thing the monthly number is describing, so hold onto it for the next section.

The month in one number: 49.5%

Now zoom out, because the daily noise has been hiding something clean.

Across the whole of July, on about $31.4 billion of Binance spot volume, Bitcoin's aggregate taker buy share was 49.5%. Aggressive sellers held a slight majority of the flow for the entire month. And Bitcoin rose from $58,625 to about $62,867, a gain of 7.24%, with a July low of $57,800 on the first and a July high of $66,956 on the twenty first.

Those two facts cannot both be true unless the buying was passive. If price rises 7% while the aggressive side is net selling, the gains did not come from buyers lifting offers. They came from buyers sitting in the book, letting sellers come to them, and absorbing supply without paying up for it. The 07:00 hour described above is that behaviour caught in a single frame.

There is independent corroboration. Bitcoin's open interest is reported to have stayed static at around 750,000 all month, which says the move was not built on new leveraged positioning. A 7.24% monthly gain with flat open interest and a sub 50% buy share is the signature of spot accumulation, and it is a healthier structure than the same gain driven by chasing, even though it feels far less exciting while it is happening.

For context on the wider market, the CoinDesk 20 index gained 8.7% over the month, reported as its biggest monthly advance since July of last year, even after falling 2.34% since Monday. So the month was strong and the week that ended it was not.

The week, seven days out of seven

One more pattern before we close the month. Take the last seven daily candles and put the direction next to whether the day's buy share was above or below 50%.

DayCloseChangeBuy share
Jul 25$64,3750.37% higher52.1%
Jul 26$65,4001.59% higher51.6%
Jul 27$63,7562.51% lower45.3%
Jul 28$63,9150.25% higher50.9%
Jul 29$63,9840.11% higher53.2%
Jul 30$64,7801.24% higher52.5%
Jul 31$62,8702.95% lower45.9%

Seven days, seven agreements. Every session that closed higher had a buy share above 50%, and both sessions that closed lower came in below 46%. There is no ambiguous case in the set, and the two down days are not marginally below the line, they are clearly below it.

Do not over read this. Seven observations is a week, not evidence of a law, and the daily buy share and the daily direction are partly measuring the same underlying thing rather than one predicting the other. But it does establish the scale at which this metric is stable. At the daily level it is coherent. At the five minute level, as the 01:05 spike demonstrated today, it is not.

What to watch next

  • Whether $62,500 gets reclaimed or becomes resistance. A three day floor that breaks usually gets retested from below. Reclaiming it on an hourly buy share above 55% would say today was the hedge unwind and nothing more. Failing there puts $60,000 in play, which is the level we have said all week sits behind it.
  • $62,466 as the marker. Today's low. A lower low next session on a sub 45% hourly reading is the continuation case.
  • Whether the ceasefire holds. This is the actual driver, uncomfortable as that is for a crypto site to write. If fighting resumes, the hedge premium comes back into both Bitcoin and gold and today reverses. If the pause holds and Brent keeps falling, the inflation trade keeps unwinding.
  • The gold pairs as the tell. Tether Gold at 40.3% and Paxos Gold at 42.2% are the cleanest read on whether this is hedge unwinding or genuine risk aversion. Both climbing back above 55% while Bitcoin stays weak would mean the story has changed to something worse.
  • July's final ETF number. Spot Bitcoin ETFs stood near $205 million net for the month, the weakest positive month on record, against $342.85 million for Ether funds. Today closes the book on it.
  • The next scheduled release, for the third pre-release test. Two for two this week on the aggressive buying arriving before the number rather than after it. That observation is still live and still untested a third time.

How to read this as a trader

The week produced three lessons and they are worth separating because they sit at different time scales.

At the level of a session, the correlation everyone assumes is unstable. Bitcoin rose on a day equities lost 2.1% and fell on a day equity futures gained 1.23%. If you were trading it as a Nasdaq proxy you were wrong twice in three days. The variable that actually explained both sessions was the market's inflation expectation, and the tell was gold moving with Bitcoin rather than against it.

At the level of a minute, today was a caution against our own enthusiasm. A 68.6% buy share on nearly twenty million dollars looks like conviction and it marked the exact high of the week. Small windows at level breaks attract the kind of order that is designed to be seen. The fix is not to abandon the metric but to stop asking it questions at a resolution where it cannot answer.

At the level of a month, the most useful number of all was the dullest: 49.5%. Sellers were marginally in charge of the aggressive flow for thirty one days and Bitcoin still gained 7.24%. That describes patient bids absorbing supply, which is what accumulation actually looks like from the inside, and it is invisible on a price chart. It also means today's 3% decline arrives on top of a month of that behaviour, not in place of it. Our live dashboard separates aggressive buying from aggressive selling minute by minute across the most liquid pairs, which is how the 07:00 absorption was visible while it was happening. For the week's full arc, see yesterday's reclaim and the second pre-release bid, the Fed's 9 to 3 hold, the morning the sellers stopped, Tuesday's broken bid and the gold rotation, and Monday's reclaim of $65,000. If the buy and sell split is new to you, the volume guide covers the mechanics.