On-chain weekly: 1.15 million BTC cluster at $63,000 as redistribution risk builds

Core call from this week’s on-chain report

An on-chain weekly report from Dashu Finance for week 31 of 2026 says the most striking number in the current Bitcoin market sits at one price: $63,000. According to the report, 1.15 million BTC are now concentrated at that single level, equal to roughly 5% of circulating supply. The authors argue that when position clusters reach that scale, the market has historically gone on to experience a major move, making the current setup one that deserves close attention.

The report also says concentration within 5% of spot price has climbed to 13.5%, pushing the market into what it describes as a warning zone and leaving it just short of the 15% high-risk threshold. The longer price stays flat, the more inventory piles up, and the more likely the next redistribution will arrive through a violent swing rather than a gradual adjustment.

Its summary line is blunt: the upstream source of liquidity has been shut off, while the downstream market has filled with powder and is now waiting for a spark.

Upstream conditions: cash is paying again

The report says the reason for this week’s quiet tape is not mainly on-chain. It starts further upstream in the macro funding environment.

First, the direction of monetary pricing has shifted. Bond markets are no longer pricing rate cuts; they have started pricing rate hikes. The 2-year U.S. Treasury yield has stayed above the federal funds rate since April, and the spread between the two has widened to its largest level since November 2022. At the latest policy meeting, the vote split 9:3, the highest number of dissenting votes since September 2016, which the report reads as a sign of unusually deep internal disagreement.

Second, the mechanical incentive to stay active in crypto has weakened. Since February, the 3-month futures basis, used here as the return available from institutional cash-and-carry trades, has remained below the 2-year Treasury yield. The report says only one other period on record lasted this long: August 2022 through January 2023, a stretch that ended at the previous cycle low.

That matters because institutional trading desks provide leverage, order-book depth, and a large share of market turnover. If U.S. Treasuries offer a higher and more certain return than crypto arbitrage, the report argues, those desks have little reason to keep meaningful capital in the market.

It uses that framework to explain a set of weak downstream readings. Spot trading volume has dropped to its lowest level since 2019; measured in BTC terms and adjusted to remove the appearance created by falling prices, the conclusion still holds, and the same is true even when Binance is excluded. Exchange inflows and outflows have both stayed muted, leaving combined flow among the weakest readings of the past three years. ETF flows briefly turned positive in mid-July, then slipped back into negative territory less than a week later.

In the report’s framing, this is not fear. It is absence. If cash itself pays investors to wait, some of that capital will keep waiting. Any real improvement, it says, would have to begin at the policy level before it shows up on-chain.

Market structure: a single vertical stack at $63,000

If the upstream picture is cold, the structure on-chain is getting hotter.

The headline figure is the 1.15 million BTC concentrated at $63,000. The report notes that 1 million BTC is roughly 5% of total circulating supply, and says single-price clusters above that scale have, in past cases, been followed by major dislocations.

It adds that the concentration at $63,000 could have been even larger. Around 550,000 BTC are locked in Coinbase wallets in the $83,000 to $84,000 range, described in the report as institutional inventory bought near the highs. Without that overhang, the pile at $63,000 would look more extreme. Taken together, the $62,000 and $63,000 levels already account for 8% of circulating supply.

The report points to a ready comparison. In late October 2022, just before the FTX collapse, around 1 million BTC were stacked at $19,000 and 870,000 BTC at $18,000, for a combined 9.7% of circulating supply. What followed is well known: a sudden trigger collided with an already fragile cost-basis structure and produced a violent repricing.

Supporting indicators are moving in the same direction. Coin concentration within 5% of spot has risen to 13.5%, which the report calls a warning area, with the 15% high-risk line now close by. As reference points, it cites readings of 18% in November 2025 and 16% in January 2026, both of which were followed by large swings.

The mechanism is straightforward. The more concentrated the inventory, the easier it is for a modest price move to trigger chain-reaction repositioning among sensitive holders. That magnifies volatility. The report’s view is that inventory cannot keep stacking forever. Once the standoff between bulls and bears reaches a critical point, the market has to choose a direction and redistribute.

The order book tells a similar story. Since early June, a large body of bids has remained parked 2% to 20% below spot and is repeatedly replenished. On the other side, asks above the market have thinned out sharply, with static sell-side depth near the lowest level of the past month.

That suggests buyers are still there, but they are not willing to lift offers at current levels. At the same time, there is not much visible supply overhead to slow a push higher. Thin books work both ways, and the report says quiet markets often turn into violent ones under exactly this setup.

Cycle position: the shallowest drawdown on record, but not finished in time terms

Before arguing over whether Bitcoin has bottomed, the report steps back and recalibrates the cycle.

By drawdown, it says this is the shallowest bear market on record. Relative to the 200-day moving average, no previous bear phase kept price this close to the long-term trendline. Even the deepest discount in the current cycle remains clearly smaller than the low seen in the mildest bear market of earlier cycles. The same conclusion holds when measured by drawdown from the all-time high: previous bear-market lows all sat materially below the lowest levels seen so far in this cycle.

By duration, the report says the market is still not done. Bitcoin has spent roughly three-quarters of the average historical bear-market duration below its 200-day moving average, and many bear markets lasted longer than that average.

That leads to a cautious interpretation. A cycle with a relatively mild decline that has not yet matched the historical time test should not be treated as complete too quickly. The report says patience remains the more reasonable posture, especially for investors who use a four-year cycle framework.

At the same time, it says this does not contradict the idea that bottom evidence has been building. Both can be true at once: the evidence chain has thickened, and the market still has not fully matched past bears in either time or depth.

A methodology warning: the “never failed” bottom signal may never show up

One of the report’s most pointed sections focuses on a common analytical trap around an on-chain bottom indicator: STH-RP < LTH-RP, meaning the average cost basis of short-term holders falls below that of long-term holders. The signal appeared at the bottom of the last three cycles and has been treated by many analysts as a core bear-market bottom feature.

The report separates two claims. The first is: if STH-RP < LTH-RP appears again in this cycle, it would be a bottom signal. The second is: because every previous bear-market bottom had it, this cycle must have it too. The report accepts the first statement and rejects the second.

Its reasoning is that a cost inversion between short-term and long-term holders is the result of a market spending enough time in a loss-dominated regime to complete a full transfer from trapped high-cost buyers to lower-cost buyers. It is the outcome of sufficient redistribution, not the cause of it.

This cycle, however, has produced something that did not appear in the prior three. On July 28, the 7-day rate of change in LTH-RP was -0.54%, while STH-RP was -0.25%. In other words, the long-term holder cost basis had turned from rising to falling, and it was falling faster than the short-term holder cost basis. The report says that condition has already lasted seven days.

The implication is simple. If that relationship persists, the two lines may never cross. The famous signal that “never failed” in prior cycles may simply never appear in this one.

The broader point is methodological. Historical signals only work when the conditions that produced them still exist. The report warns against turning a historical pattern into an assumed certainty. If the premise changes, the outcome may never arrive.

Event watch: unusual transfers by long-term holders

The other major variable this week was abnormal transfer activity by long-term holders.

According to the report, long-term holders moved more than 65,000 BTC per day on Aug. 3 and Aug. 4 after same-entity internal transfers were excluded, producing a sharp drop in net LTH holdings. Before that, net LTH holdings had already stopped their steady climb from May onward and stayed flat through July, something the report calls rare over the past year.

The first market reaction was concern: did these holders know something in advance? After two days of tracking, the report leans to a less alarming answer.

• Transfer volume has already eased. A peak of about 65,000 BTC seen around July 31 to Aug. 1 fell to 38,000 BTC on Aug. 3 to Aug. 4, though that is still above the usual 10,000 to 15,000 BTC range.
• The share sent to exchanges did not expand in a notable way and was around 14,000 BTC.

Taking those pieces together, the report says the move looks more like forced migration triggered by an unexpected event such as a hardware wallet security flaw than by defensive selling. Most of the coins that moved did not go to exchanges, so the transfers do not appear to translate into immediate sell pressure. A smaller share may have been tied to selling by a listed company.

On that point, the report cites a newly filed 8-K. It says a well-known public company sold 1,638 BTC last week at an average price of $63,957, raising about $105 million. That was clearly below its average holding cost of $75,419, making it a loss-taking sale. The proceeds were mainly used to pay preferred dividends and repurchase notes.

The report says the more important detail came from the company’s earnings call, where it stated that under its current capital management plan it could sell as much as $5 billion in BTC, four times the $1.25 billion board authorization set at the end of June.

Because firms like this often act as psychological anchors for long-term holders, their behavior can be interpreted far beyond its direct size. For now, the report says the market is still absorbing that potential overhang and the fragile balance has not broken. But risk can accumulate for a long time. The key question is what external force, and when, might set it off.

The report lists several possible triggers: the possibility of a Federal Reserve rate hike; Middle East conflict and oil prices as an upstream inflation variable; concentrated valuations in U.S. AI equities and capital spending that increasingly depends on debt and private-credit financing, which could lead to systemic deleveraging if revenue delivery disappoints; and renewed crowding in yen carry trades, where net short positioning has again approached historical extremes.

None of those outcomes is presented as inevitable. The report’s point is narrower: in a market with this kind of sensitive inventory structure, any one of them could amplify volatility if it becomes the spark.

The overlooked constructive signal: bear-market supply is weakening cycle by cycle

After the more defensive sections, the report ends with a longer-term data set that helps explain why this drawdown has been shallower than earlier ones.

Its starting principle is that the main source of supply in bear markets is not profit-taking but loss-taking. In each bear cycle, the dominant selling pressure usually comes from coins bought in the previous year, which tends to line up with late-stage bull-market demand.

The report says the ratio has fallen from cycle to cycle. Its explanation is that as more long-term investors enter the market, Bitcoin’s ownership structure matures, and a smaller share of bull-market-top inventory gets forced out during the bear phase.

In the current cycle, a $64,000 Bitcoin price leaves all BTC bought in 2025 under water. About 4.94 million of those coins remain, and the report says they are now all in long-term holder hands. Total underwater supply held by long-term holders stands at about 5.69 million BTC, which means roughly 87% of that loss-making LTH inventory was bought in 2025. Those coins make up the main source of sell pressure on down moves.

That supply has already dropped nearly 40% from a peak of 8.14 million BTC, with most of the decline occurring in late December 2025. The report says that cliff-like reduction coincided with a sharp Bitcoin decline.

It also notes that coins absorbed in 2025 by ETFs and corporate treasuries, roughly 750,000 to 850,000 BTC on a net accumulation basis, have mostly not moved during this bear market. If those coins had been bought instead by short-term traders and later sold in the drawdown, the reduction in 2025-vintage supply would have reached 50%, according to the report.

The takeaway is that high-cost distributable supply has already been cut by nearly half, while demand has not fallen to zero. When one side weakens and the other side still exists, markets often become less responsive to bad news. The report presents that as the clearest explanation for the relative resilience of this cycle.

Technical indicators, it says, are moving in the same direction. The trend-following CCI flashed a weekly oversold signal this week. Over the past five years, the same signal appeared only four times: November 2018, March 2020, June 2022, and November 2025. The report adds an important qualifier: oversold only describes intensity at a moment in time. What matters more is whether the curve then keeps narrowing and diverges from price, a setup it sees as conceptually aligned with the on-chain signal of contracting realized net losses diverging from price.

Key levels and what would confirm improvement

The report closes by pinning the current map to three key zones.

• $62,000 to $68,000: the heaviest cost-basis cluster and the market’s core support band. Roughly half belongs to short-term holders who bought during this year’s decline and are mostly under water, making them likely to react first on rebounds. The other half belongs to long-term holders who have held through the slide and typically act as patient supply near bottoms.
• $69,000: the short-term holder cost basis and the first break-even resistance level. The report calls it the key price for the next phase.
• $83,000 to $86,000: the long-term holder supply wall and the next major ceiling once price clears $69,000.

From there, the report defines its observation points. Genuine improvement, it says, would likely start at the policy level and then show up as stronger trading volume, a move back above $69,000, and ETF channels shifting from stagnation to sustained buying. On the other hand, if the $62,000 to $68,000 support band breaks while exchange inflows become active again, that constructive thesis would be invalidated.

It also highlights one final line on the chart. For the past six months, BTC has traded back and forth around what the report calls the true average cost basis after excluding coins that have not moved for more than 10 years. In the previous three cycles, that line acted as strong resistance. In this cycle, it has turned into support.

One possible reading is that sellers stopped pressing below average cost, while buyers began to see opportunity there. If supply and demand had not found some form of balance at that level, the report argues, price would not have stayed there this long.

The report says on-chain data can show where the cycle stands and where the key coordinates lie. It cannot decide when to enter, how much risk to take, or where to admit a trade is wrong. Those answers sit outside the chain and inside a trader’s own plan.

The data and some of the views in the article were compiled from public analysis published by on-chain analyst Murphy (@Murphychen888) between July 30 and Aug. 7, 2026, and from Glassnode’s week 30, 2026 on-chain report, “There Is a Return in Waiting,” dated Aug. 1. The publication said its conclusions were for information sharing only and do not constitute investment advice.