The question price cannot answer
Bitcoin has held near $65,000 for seven consecutive sessions. On its own, a flat tape tells a long-term holder nothing at all.
What matters is which cohort sits on each side of that flat tape. On-chain data published this week by CryptoQuant offers the clearest answer the market has produced in months.
Coins are moving out of small wallets and into large ones. The transfer began in December, and it accelerated sharply when price broke below $60,000 in June.
What the cohort data shows
Entity balances excluding exchanges and mining pools now stand at 3.06 million BTC. The same cohort bottomed near 2.87 million BTC in December 2025.
That works out to roughly 190,000 coins absorbed over eight months by wallets that are neither trading venues nor miners. These are holders taking delivery, not intermediaries warehousing supply for someone else.
The number still sits below the 2025 cycle peak of about 3.23 million BTC. Large holders have not rebuilt their full position. What they have done is reverse the direction of travel.
Small wallets are the sellers
Data from Santiment covering the past week shows larger wallets adding 19,610 BTC while balances held in small retail addresses declined over the same stretch.
This is the least glamorous pattern in Bitcoin and historically one of the more informative. Supply migrates from short-horizon holders to long-horizon ones at prices the sellers tend to regret, and the process runs in months rather than days.
It is worth being precise about what this does and does not mean. Cohort accumulation describes who owns the supply. It says nothing whatsoever about when price responds to that ownership.
The sell-side pressure gauge
A second metric points the same direction. Axel Adler Jr., an analyst whose work CryptoQuant publishes, notes that the Adaptive Sell-side Risk Ratio has fallen to 0.031.
That places the reading in the third percentile of the current halving cycle. The ratio weighs realised profit and loss against the size of the realised cap, which in plain terms measures how much value is actually being cashed out relative to the stored cost basis of the network.
Readings this low mean holders have stopped selling. Not that they are buying aggressively, but that coins are no longer being moved to market in size.
CryptoQuant frames the combination as consistent with the final phase of a bear market. The firm is careful to add that a bottom is not confirmed, and that caveat deserves considerably more weight than the headline it sits under.
Reading it against the last two cycles
The useful comparison is not to the 2021 top but to the accumulation stretches that preceded the two prior recoveries.
In both 2015 and 2019, the cohort of large non-exchange entities began growing well before price found a durable floor. In both cases the growth was gradual rather than dramatic. The roughly 190,000 coins added since December fit that shape closely.
What differs now is the presence of the spot ETF complex, which existed in neither prior cycle. Roughly 1.26 million BTC also sits on the balance sheets of about 200 public companies, a category that barely registered before 2020.
The practical effect is that the free float, meaning coins realistically available to be sold at short notice, is smaller than the raw supply figure suggests. That cuts both ways. A thinner float amplifies moves in both directions, and nobody should assume it only amplifies the pleasant one.
The second bid
The cohort data does not exist in isolation. US spot Bitcoin ETFs took in roughly $755 million over the past week, their strongest five sessions since April, with BlackRock accounting for the clear majority of it.
Two independent pools of capital are therefore absorbing supply at the same time: large on-chain entities and regulated fund wrappers. Neither group is a leveraged trading desk, and neither turns over its position weekly.
That is a structurally different market from one propped up by borrowed money. It is also a slower one. Steady supply absorption without a demand shock produces exactly what the past week has looked like, which is a chart that goes nowhere.
What this does not tell you
Three cautions are worth stating plainly before anyone treats the above as a signal.
First, cohort labels are heuristics. A whale address is simply an address above a size threshold, and clustering real entities out of chain data involves assumptions that are sometimes wrong.
Second, accumulation during a downtrend has preceded further downside before. Large holders bought steadily through 2022 and sat underwater for well over a year afterwards.
Third, none of this addresses the macro calendar. A US CPI print lands next week, and the September Federal Reserve meeting remains genuinely uncertain, with markets pricing close to even odds on a hold.
Bitcoin Gate Take
The signal here is not the direction of the arrow but the identity of the buyer. Supply migrating from wallets that sell on a six-week horizon to wallets that sell on a six-year one shrinks the float available to drive the next move, whichever way that move goes.
Watch whether the whale cohort keeps climbing through a genuinely bad macro print rather than a quiet one. Accumulation into calm is ordinary; accumulation into fear is the part that has historically mattered. None of this makes a bottom certain, and anyone using a third-percentile risk ratio as a timing tool has misread what it measures.
If you accumulate on a fixed schedule instead of trying to identify the floor, the DCA calculator lets you model what eight months of sideways price actually did to a recurring buy.