比特币牛市指数触及90分，但需求去哪儿了？

CryptoQuant’s trend gauge registered an emphatic reading just as its estimate of new spot demand deteriorated. The apparent contradiction tells a more useful story about who bought the breakout, who supplied the coins and why a strong trend can lose its marginal buyer.

Summary CryptoQuant put its Bitcoin Bull Score at 90 out of 100 after the September breakout.

Its apparent demand measure contracted by about 170,000 BTC over the preceding 30 days.

US spot Bitcoin ETFs took in roughly $2.39 billion during the week ended September 25.

Strategy disclosed a purchase of 1,665 BTC in its latest weekly update.

Bitcoin traded near $83,300 on September 30 after reaching roughly $87,400.

Bitcoin’s Bull Score reached 90 out of 100 after the coin crossed its 365-day moving average. CryptoQuant’s separate estimate of apparent spot demand, meanwhile, contracted by roughly 170,000 BTC over 30 days. One number describes the state of a trend. The other asks whether the market is absorbing new supply at the same pace. They can disagree without either being wrong.

That disagreement matters after a fast advance. Bitcoin traded near $83,300 during Asian hours on September 30, below an eight-month high around $87,400. A buyer who saw the score alone might read a near-perfect verdict on the next move. The demand measure offers a different, narrower warning: the market’s net absorption has weakened even while price and several trend indicators remain strong.

The question is not whether somebody bought every coin sold. Every executed trade has a buyer and a seller. It is whether new holders are taking enough coins out of the liquid supply to absorb miners, profit takers and other distributors at current prices. Exchange-traded funds plainly attracted cash during the run. That does not automatically mean the entire market gained demand on a net basis.

The 90 score measures a condition already reached

CryptoQuant’s Bull Score combines onchain and market indicators into a composite reading. It rose after Bitcoin broke above the 365-day moving average last week, a threshold the firm treats as confirmation of a bullish market regime. A score of 90 says that most inputs currently meet its bullish criteria. It is not a forecast that Bitcoin has a 90% probability of rising, and it is not a count of coins available for purchase.

Momentum indicators routinely strengthen after a rally. The price crossing a long moving average records that the recent market is stronger than the preceding year by that particular test. It does not identify who will buy the next sale. The score can stay high while the marginal buyer retreats because many of its ingredients reflect moves that have already happened or conditions that take time to reverse.

The September 30 market report attributed the 90 reading and the 170,000 BTC contraction to CryptoQuant. Readers should treat the two as different measurements from the same research provider, not competing opinions about an observable quantity. A composite score has its own weighting and lookback choices. Apparent demand is an estimate constructed from supply behavior. Neither is a complete tape of named buyers.

The price path has already tested the distinction. Bitcoin moved above $87,000 before retreating toward the low $83,000s. A strong score was compatible with that pullback. It may remain compatible with a renewed rise if fresh purchases arrive. Calling the number a buy signal would assign it a job its construction does not support.

JUST IN: Bitwise research shows major institutions are holding Bitcoin alongside gold

Bitwise research head Ryan Rasmussen said none of the 15 institutions surveyed sold their crypto during Bitcoin’s drop from about $125,000 to $60,000, with several increasing their positions. pic.twitter.com/4yXFMbH4nV — crypto.news (@cryptodotnews) September 30, 2026

What does a contraction of 170,000 BTC actually mean?

CryptoQuant’s apparent demand measure aims to capture changes in the supply held for longer periods relative to newly issued coins. The firm’s estimate of a roughly 170,000 BTC contraction over the past 30 days is a change in that demand measure, not a report that exactly 170,000 BTC were dumped on exchanges in a single month. It should not be added mechanically to exchange deposits, ETF flows or miner sales. Those series use different definitions, dates and sometimes overlapping coins.

For scale, Bitcoin’s current block subsidy is 3.125 BTC. At an average of 144 blocks a day, that implies roughly 450 newly issued BTC daily, or about 13,500 BTC over 30 days before variation in actual block production. The reported 170,000 BTC change is around 12.6 times that illustrative monthly issuance. That comparison does not imply miners sold 170,000 BTC. It shows why a change in investor holding behavior can overwhelm the flow of freshly mined coins.

Nor is the figure a balance sheet of the entire Bitcoin market. A coin moving from an exchange to a wallet can affect one dataset differently from another. Custodial addresses may combine thousands of investors. The same economic owner can move a coin without selling it, and an ETF can change its holdings through a creation or redemption without revealing the ultimate person on the other side. Apparent demand is useful as a consistent time series, particularly when its direction changes, but its label should not be read literally as a census of buyers.

The temporal mismatch matters. The 90 score reacts to a breakout and a set of current bullish inputs; the demand contraction covers a rolling 30 days. If the market spent much of that interval distributing coins and only recently attracted a wave of ETF inflows, both results can hold. The next several updates, computed on the same method, would tell whether the ETF wave changed the 30-day measure as weaker days leave the window.

Another caution follows from arithmetic. Multiplying 170,000 BTC by a current price near $83,300 produces about $14.2 billion. That is a scale illustration, not $14.2 billion in measured withdrawals. The BTC figure is a change in an estimated demand series, and valuing every unit at one end-of-period price does not turn it into a cash-flow statement. This is precisely why comparisons with ETF dollars need explicit units and dates.

The ETF bid was real, but its time window was shorter

US spot Bitcoin funds drew approximately $2.39 billion in the five trading sessions through September 25, according to Farside daily fund-flow table. The week included about $999 million on September 21 and $714.7 million on September 22. Those are large inflows into a defined investment wrapper. They establish that one identifiable class of buyers added exposure during the breakout.

At an illustrative $84,000 per BTC, $2.39 billion would equal about 28,450 BTC of purchasing power. Actual fund acquisitions occur at the prices and mechanisms prevailing on each day, so 28,450 is a conversion for scale, not a reconstruction of daily custody changes. It is also about one-sixth of the 170,000 BTC change cited for a different 30-day measurement. Subtracting one from the other as if both were cash trades would be invalid. Placing them in the same units, with those caveats, shows why five good ETF sessions need not reverse a month of deteriorating apparent demand.

JUST IN: BlackRock’s Bitcoin ETF is among the top 8 most traded stocks today

IBIT has done $2.3 billion in volume so far pic.twitter.com/n0fIDus5tT — crypto.news (@cryptodotnews) August 28, 2026

ETF net flow itself deserves care. A positive daily net subscription generally leads the fund structure to acquire or hold more Bitcoin, but a fund share can trade between investors all day without any new coins entering custody. Secondary-market volume is not the same as creations. Net inflow is the number relevant to aggregate fund exposure, and even that cannot disclose whether the ultimate purchaser was a pension, hedge fund, adviser or individual investor.

The timing of the inflows is instructive. The biggest sessions clustered around the breakout. By September 26, the seven-session streak included a much smaller $134.5 million Friday intake. A positive but decelerating stream can still support price if the sell side recedes faster. It can fail to hold a breakout if profit taking accelerates. A flow headline without the supply response is an incomplete market story.

Earlier in September the same wrapper was capable of moving the other way. Funds lost approximately $746.3 million across September 15 and 16 before demand returned, as the Fed hike feature documented. The sequence argues against treating ETF participation as a permanently open tap. Fund investors buy, pause and redeem. A rolling 30-day measure includes both weak and strong stretches.

The fund flow becomes a Bitcoin purchase through several steps

An ETF inflow begins with money entering a fund share structure, not with a named investor taking delivery of a particular coin. Authorized participants create or redeem blocks of shares through the fund’s prescribed process. The trust and its trading counterparties arrange Bitcoin exposure and custody according to its documents. The investor usually ends with a brokerage position in shares, while the custodian controls the keys for the trust’s underlying coins. Each layer answers a different question about demand.

The daily net creation number is the clearest public indication that the wrapper’s aggregate size grew. Even there, timing can matter. A market maker may have acquired coins before shares were created, or hedge inventory while an order is processed. A dealer that already owns Bitcoin may transfer inventory to meet demand. A publicly reported creation is a useful dated signal, but it is not a timestamp for the first instant someone decided to buy. It cannot prove that its full dollar amount hit a spot order book on the same trading day.

JUST IN: Strategy adds 1,666 $BTC to its holdings

Today's purchase brings the company’s total Bitcoin stash to 847,666 $BTC. pic.twitter.com/VicQoVnpO9 — crypto.news (@cryptodotnews) September 28, 2026

That distinction is particularly relevant when headlines attribute a one-hour price move to that day’s fund flow. Fund totals are normally confirmed after the US trading session. Price responds continuously to orders, dealer positioning and expectations. An observed $999 million session can validate that significant exposure was added; it cannot allocate precise causation to every earlier candle. A trader who sold futures into an anticipated fund order and covered later may appear nowhere in the net fund statistic.

Net flows conceal dispersion across products too. A positive sector total can contain redemptions from one fund and creations in another. Some of that activity may be investors changing vehicles for fee, tax or access reasons while maintaining much the same Bitcoin exposure. The sector’s net result is still meaningful, but gross flows should not be counted as new market demand when they partly reflect a transfer between wrappers. Farside’s product columns allow a reader to inspect those offsets rather than relying solely on the final total.

You might also like: OKX Bitcoin user holdings rise 4% in latest reserve report

Neither the fund reports nor the blockchain supplies the beneficial owner of a brokerage account. A transfer into a custodian address does not reveal whether the owner of the new fund shares plans to hold for a decade, rebalance next month or hedge the entire exposure through futures. That holding period is central to CryptoQuant’s apparent-demand concept. Two investors can buy the same number of ETF shares today and have opposite implications for the market’s future supply.

The same caution applies to a week of ETF inflows quoted in dollars against a month of BTC-denominated apparent demand. Converting the dollars to an approximate BTC count is helpful, as the 28,450 BTC illustration above shows. It does not change the coverage of either series. A properly paired comparison would show daily fund creations in coins where available, the method used to convert dollar flow, the 30-day demand update on each date, and the rest of the market’s holder distribution. Without all four, the visible fund bid is evidence of a buyer class, not a reconciliation of the whole market.