Bitcoin Halving Cycle Math: The Wallets Behind the Chart

In June 2026, Bitcoin’s two most closely watched buyer classes moved in opposite directions at the same time. Spot Bitcoin ETFs listed in the United States shed a record $4.06 billion over the month, the worst outflow stretch since the funds launched in January 2024 and well past the previous record of $3.56 billion set in February 2025, according to CoinDesk. Over roughly the same stretch, wallets large enough to qualify as whales added an estimated 270,000 BTC, worth about $16.7 billion at the time, much of it bought near the cycle low around $59,000, per data Bitfinex analysts shared with the same outlet.

That is not really a contradiction. It is a demonstration that “Bitcoin demand” is not one thing. It is several distinct cohorts of wallets, each with its own incentives, tax exposure, time horizon and access to liquidity, and they do not always move together. The halving itself doesn’t touch any of this directly: it is a fixed, code-enforced cut to new supply that arrives on schedule regardless of who is buying or selling that week. What actually sets price is the netting of decisions made by miners, long-term holders, short-term traders, ETF allocators and exchanges, and a meaningful share of those decisions leave a visible trail on a public ledger.

This piece sets aside the forecasting models this series has already tested against each other (see our look at how stock-to-flow, the power law and the global-liquidity thesis have held up at the cycle’s midpoint) and asks a narrower, more falsifiable question instead: what do on-chain wallet cohorts actually do around each halving, and does that behavior repeat in a way that’s useful, or is it just the price chart redescribed in wallet-age units?

The reason this question is worth a dedicated look is that it sidesteps a lot of the argument that surrounds price-target models. Nobody seriously disputes that a wallet either moved a coin on a given date or didn’t; that part of the record is just arithmetic on a public ledger. The disagreement starts one step later, over what that movement, or lack of it, actually implies about conviction, causation or what happens next, which is a more honest place to locate the uncertainty than arguing over whose price model is right.

The Fixed Part: What a Halving Actually Does

Bitcoin’s issuance schedule is enforced directly in the reference client’s code, in the GetBlockSubsidy function inside Bitcoin Core’s validation.cpp, not by any company or foundation vote. Every 210,000 blocks, roughly four years, the block reward halves: 50 BTC at genesis in 2009, 25 from November 2012, 12.5 from July 2016, 6.25 from May 2020, and 3.125 since April 2024. The fifth halving lands at block 1,050,000. As of this writing Bitcoin sits at block 960,380, with about 89,620 blocks, roughly two years, left to go; CoinGecko’s live countdown currently estimates the date at April 13, 2028, though that estimate drifts with hashrate and shifts slightly on every check.

Bitcoin’s circulating supply is about 20.064 million coins against the 21 million cap, with the asset trading near $63,644 and a market capitalization around $1.277 trillion at time of writing, per CoinGecko. None of that changes because of anything a wallet does. What changes, cycle to cycle, is how existing holders respond to it, and that response is where the more interesting, less mechanical story lives.

Reading the Ledger: HODL Waves, Cost Basis and SOPR

On-chain analytics firms such as Glassnode and CryptoQuant built an entire vocabulary around watching how coins age. A HODL wave chart buckets the whole circulating supply by how long each coin has sat unmoved since its last on-chain transaction, then stacks those bands over time so you can watch old coins pile up or get spent. The most commonly cited cutoff is 155 days: Glassnode classifies any coin that hasn’t moved in that window as belonging to a long-term holder (LTH), based on the empirical finding that coins tend to stop correlating with short-term price swings once they cross that age. Anything younger belongs to short-term holders (STH), a cohort dominated by recent buyers, active traders and exchange-routed flow.

It’s worth being precise about what that 155-day line is and isn’t. It’s a statistical convention describing on-chain behavior, not a legal or tax category. A country’s capital-gains holding period for crypto is a completely separate question with its own rules and its own paperwork; readers who actually need to track that for a specific exchange can see our comparison of how Coinbase, Binance, Kraken and OKX handle tax reporting, which is a different problem from the one this piece is describing.

Two derived metrics matter for what follows. Realized price treats each coin as if it were bought at the price it last moved on-chain, then averages that across a cohort, giving a rough estimate of that group’s aggregate cost basis. The short-term holder realized price is watched closely because it tends to act as support in bull phases and resistance in bear phases, since it marks where the average recent buyer sits at a profit or a loss. As of early July 2026, on-chain trackers put that short-term cost basis in the high $60,000s, above spot price at points during the month, meaning a meaningful share of recent buyers were sitting underwater. SOPR, the Spent Output Profit Ratio, measures whether coins moving on-chain right now are being sold at a profit (above 1.0) or a loss (below 1.0); a SOPR reading that falls and holds under 1.0 during a drawdown is a rough proxy for capitulation, forced or panicked selling rather than orderly profit-taking.

None of these metrics are exotic anymore. They show up as standard chart types across Glassnode Studio, CoinGlass and Bitcoin Magazine Pro, and they get cited constantly in cycle commentary. Whether the patterns they reveal are genuinely new information, or just the price chart restated in wallet-age units, is a question worth holding onto; it comes back later in this piece.

Halving One and Two: A Blind Spot, Not an Absence

The 2012 and 2016 halvings predate almost all of the tooling described above. Glassnode wasn’t founded until 2018, and while UTXO age-band analysis is technically possible to reconstruct retroactively for any point in Bitcoin’s history (the ledger is a complete public record back to the genesis block), granular, cohort-level reporting on holder behavior around those first two halvings is thin compared to what exists for 2020 and 2024. What is well documented is the price outcome: Bitcoin ran from roughly $12 at the November 2012 halving to about $1,150 by its November 2013 peak, and from about $650 at the July 2016 halving to roughly $19,700 by December 2017.

Whether those runs were preceded by the same accumulate-then-distribute wallet pattern visible in later cycles, or reflected a different dynamic in a market with a fraction of today’s user base and almost no institutional participation, is harder to say with confidence than for the two most recent cycles. Treat any precise holder-cohort percentage claimed for this era with real skepticism. The instruments to measure it properly, at scale, with any consensus methodology, simply weren’t built yet.

That gap matters for how much weight to put on any four-cycle pattern claimed later in this piece. A dataset with two data points that are effectively qualitative and two that are rigorously quantified isn’t the same as four equally weighted observations. Readers should discount the older half of the comparison table further down accordingly, rather than treat it as four equal trials of the same experiment; it’s really closer to two solid trials plus two rough historical sketches.

Halving Three: The Cohort Record That Preceded the Parabola

The May 2020 halving is the first cycle with genuinely solid cohort data behind it. Long-term holder supply climbed through 2020 and hit what was then an all-time high of 12,656,092 BTC on October 19, 2020, according to on-chain data compiled by CryptoBriefing and echoed in Bitcoin Magazine Pro’s own long-term holder supply chart. At the time, that figure represented roughly two-thirds of Bitcoin’s then-circulating supply of about 18.5 million coins, an unusually high share of the ledger sitting completely still. Over the following four months, as fresh capital (much of it institutional, arriving ahead of Bitcoin’s run past $60,000 and eventually to nearly $69,000 by April 2021) flowed in, that long-term holder supply figure drew down as long-time holders sold into the rally.

That sequence, quiet accumulation into a supply peak followed by distribution as price discovers new highs, is the closest thing this dataset offers to a repeatable, falsifiable on-chain pattern tied to a halving cycle. It’s worth being careful about what it does and doesn’t show. It doesn’t prove the halving caused the accumulation; issuance dropping by half a year earlier is one plausible motive among several, alongside 2020’s extraordinary monetary and fiscal response to the pandemic, which pushed capital toward scarce assets generally, Bitcoin included. What it does show is that, empirically, this specific cycle’s price discovery phase was preceded by a genuine, multi-month build-up of dormant supply, not a sudden appearance of new buyers with no on-chain footprint beforehand.

Halving Four: Two Accumulation Waves in One Cycle

The current cycle complicates the tidy one-peak story above, because it has produced at least two distinct on-chain accumulation waves rather than one. In the run-up to the April 2024 halving, Glassnode data cited by The Block showed long-term holder profit-taking cooling off in the weeks after Bitcoin’s break above $73,000 in March 2024, even as that same all-time high triggered a wave of early distribution. Exchange reserves were already falling sharply through that period: CryptoQuant data reported the same month showed roughly 90,700 BTC withdrawn from exchanges in a single month, part of a decline of about 900,000 coins from the 2.8 million held on exchanges back in July 2021, per a separate Block report citing the same data provider.

The more striking pattern shows up after Bitcoin’s October 2025 all-time high of $126,198 and the drawdown that followed. A research note on 2026 on-chain signals, published in April 2026 and summarized by on-chain analytics blog BGeometrics, described a first wave of whale accumulation partway into that decline: roughly 270,000 BTC added to large-wallet balances over 30 days into February 2026, including a single-day record of 66,940 BTC moved into accumulation addresses on February 6, alongside exchange net outflows of about 48,500 BTC over the same window. The note flagged its own limitation, worth repeating here: wallets holding 1,000 or more BTC mix genuine whale investors together with exchange and custodian wallets, so headline accumulation figures in that bucket can overstate how much is truly independent buying rather than internal reshuffling.

Then, months later, a second wave. As Bitcoin bottomed near $59,375 on June 5, 2026, whale wallets added roughly another 270,000 BTC over two weeks, worth about $16.7 billion, even as ETFs recorded their worst outflow month on record. Two accumulation waves of similar size, five months apart, both clustered around local lows within the same cycle, is a considerably stronger pattern than either wave taken alone. Standard Chartered’s Geoffrey Kendrick called the June low the cycle’s bottom that same week, saying “Winter is over. Welcome back to crypto Spring,” according to CoinDesk, a call that lines up with, though it doesn’t prove causation from, the on-chain accumulation data around the same date.

By May 21, 2026, long-term holder supply had climbed to about 14.83 million BTC, or 76.09% of circulating supply, itself roughly 17% above the October 2020 peak described earlier, per CryptoBriefing’s tracking. Separately, Fidelity Digital Assets research has tracked an even slower-moving cohort: bitcoin dormant for ten years or more, so-called ancient supply, which crossed a notable threshold in April 2024, the same month as the halving itself. More on that below.

Four Halvings, Four Different Cohort Stories

Laid side by side, the four cycles look less like one repeating pattern and more like a story that gets richer, and more testable, as the tooling to observe it improves.

The Ancient Supply Queue: A Slower Clock Than the Halving

Long-term holder supply (155 days and up) is the fast-moving version of this story. Ancient supply, coins that haven’t moved in ten years or more, is the slow-moving one, and it has its own crossover point worth flagging. According to Fidelity Digital Assets research, ancient supply now accounts for more than 17% of all bitcoin ever issued, with roughly 3.4 million BTC having entered that category since January 2019. The critical threshold was crossed in April 2024, the same month as the fourth halving: the average daily rate of bitcoin aging into the ancient-supply bucket (about 566 BTC per day) began exceeding the daily rate of new issuance (450 BTC per day post-halving), the first time in the asset’s history that coins permanently leaving circulation outpaced the rate of new coins entering it.

The same research found that the ancient-supply share increases on about 97% of days, a sign of minimal selling pressure from this specific cohort under normal conditions. It also found a notable exception: in the volatile period around the November 2024 US election, ancient supply declined on roughly 10% of days, well above its historical 3% baseline, meaning even genuinely ultra-long-term holders sold during that stretch of stress. Fidelity’s own projections put ancient supply at roughly 20% of total issuance by 2028, around the time of the fifth halving, and 25% by 2034, purely as a function of the fixed issuance schedule continuing to slow down.

Exchange Reserves: Coins Leaving the Order Book

A third lens, distinct from holder-age cohorts, is simply where coins sit. CryptoQuant’s exchange reserve tracking has shown a long, fairly steady decline: from about 2.8 million BTC held on exchanges in July 2021 down to a range of roughly 2.4 to 2.7 million BTC by early 2026, according to figures reported by Benzinga, which cited CryptoQuant, Glassnode, Bitwise and Fidelity Digital Assets research in aggregate. The same reporting put illiquid supply, coins effectively removed from active circulation and unlikely to be sold soon, at about 14.37 million BTC, or 72% of all mined bitcoin, up from 13.9 million at the start of 2025.

Two forces get most of the credit for that shift. One is self-custody: hardware wallet adoption reportedly reached record levels through early 2026, an aftereffect of the 2022 FTX collapse that has proven durable rather than a short-lived reaction. The other is the growth of spot Bitcoin ETFs, whose custodians (a small number of firms, Coinbase Custody chief among them) hold coins in cold storage that rarely returns to exchange order books at all. ETF custodians collectively hold roughly 1.3 to 1.5 million BTC, on the order of 6.5% to 7% of total supply, functioning as what Benzinga’s sourcing called “one-way vaults.” Falling exchange reserves are frequently read as a bullish supply signal on their own, but it’s worth noting the two drivers cut differently: self-custodied coins can return to an exchange in minutes if their owner decides to sell, while ETF-custodied coins require someone to redeem fund shares first, a slower and more visible process.

Whales and ETFs Are Not the Same Cohort

It’s tempting to lump “big buyers” into a single category, but the June 2026 divergence is a useful reminder that whale wallets and ETF flows are structurally different animals. A whale wallet is simply an address holding a large balance; nothing about the label says whether the underlying owner is a long-term conviction holder, a market maker, an exchange cold wallet, or a fund’s own custody address, which is exactly the ambiguity the BGeometrics-summarized research note flagged above. ETF flows, by contrast, are reported daily as fund creations and redemptions, a regulated, disclosed process with far less ambiguity about what’s actually happening, even if the identity of the underlying end investor is opaque.

CryptoQuant founder Ki Young Ju has argued the market still leans too hard on the ETF wrapper as a proxy for genuine demand, saying Bitcoin “needs to be a core macro asset, not just a retail-driven ETF trade,” in comments reported by CoinDesk. The June divergence is a real-world test of that framing: when the ETF wrapper went into reverse, on-chain wallets didn’t follow it down, they moved the opposite way. Whether that’s evidence of a maturing, diversified base of demand or simply two different investor types reacting to the same drawdown with opposite risk tolerances is a genuinely open question. Large trades on either side still have to clear somewhere, whether through an ETF’s authorized participants or directly on exchange order books and OTC desks; for a sense of how differently major venues actually handle size, see our test of how Coinbase, Binance, Kraken and OKX handle institutional-size trading.

Miners: The Cohort That Has to Sell

Every cohort discussed so far chooses when to transact. Miners are the one cohort in this story that frequently doesn’t have that luxury. Unlike a long-term holder sitting on appreciated coins with no bills to pay, miners carry real, recurring costs, electricity, hosting and hardware amortization chief among them, denominated in fiat, which means a share of newly mined supply tends to move toward exchanges regardless of what price is doing, simply to cover operating expenses. That dynamic doesn’t disappear at a halving; if anything it sharpens, since the same fixed costs now have to be covered by half as many new coins per block.

Miner-to-exchange flow is tracked as its own on-chain category, distinct from whale or long-term-holder flow, precisely because the selling motive is structurally different: forced, cost-driven liquidation rather than a voluntary portfolio decision. In practice this shows up as a fairly steady background drip rather than the sharp, clustered waves seen in the whale-accumulation data above, though it can spike sharply during periods of miner financial stress, the kind this series has covered separately in its look at the sector’s post-halving shakeout. A sudden jump in miner-to-exchange flow, especially alongside falling hashprice, tends to say more about miner balance sheets than about the broader market’s conviction, and shouldn’t be read the same way as a spike in whale or ETF flow.

Sizing this cohort correctly matters for reading the rest of the data in this piece. New issuance since April 2024 runs at 450 BTC per day; even if every single newly mined coin were sold immediately, that’s a small fraction of the roughly 48,500 BTC in monthly exchange net outflows described above. Day-to-day exchange balance changes are overwhelmingly a story about existing holders repositioning, not new supply hitting the market, which is exactly what the halving’s shrinking daily issuance would predict as its share of total flow keeps getting smaller.

The On-Chain Dashboard Heading Into the Back Half of 2026

Pulling the threads above together into one snapshot, here is where the main on-chain cohort metrics stood as this cycle’s most recent data points were reported.

The Circularity Problem: Is This Actually Predictive?

Before treating any of the above as a trading signal, it’s worth applying the same scrutiny this series has applied to price-based forecasting models. On-chain cohort metrics have a structural weakness that’s easy to miss: they are themselves derived from price. Realized price only exists because a price was recorded at the moment each coin last moved. LTH and STH supply splits are a function of when holders chose to transact, and those choices are influenced by, among other things, the price at the time. That doesn’t make the metrics worthless, but it does mean statements like “long-term holders are accumulating, which is bullish” can slide into circular reasoning if you’re not careful: coins classified as long-term-held are, almost by definition, coins whose owners weren’t selling, so noting that they didn’t sell is close to restating the definition rather than revealing new information.

The more defensible use of this data is descriptive rather than predictive: it tells you what already happened to a reasonable degree of confidence (coins really did stop moving for a period, wallets really did add balance over a specific window), even if it can’t cleanly tell you what happens next. The two 2026 whale-accumulation waves are a good test case. Both are real, well-documented events. Neither one, on its own, guaranteed the price would recover afterward, and the research note that flagged the February wave rated its own predictive conviction as only “medium,” citing exactly this kind of measurement ambiguity around wallet classification. Treat on-chain cohort data the way a careful reader treats any single indicator: useful context for what’s already happened, not a clean crystal ball for what happens next.

What to Watch Heading Into Halving Five

Skybridge Capital’s Anthony Scaramucci, reflecting on the compressing pattern across Bitcoin’s four completed drawdowns (roughly 86%, 84%, 77.5% and 53% peak-to-trough in turn), put it simply: “Every bottom looked like this.” Whether or not that holds a fifth time, the on-chain cohort data gives readers something more specific to track than price alone between now and the fifth halving, still roughly two years out. A short list worth bookmarking:

• Long-term holder supply as a share of circulating supply, watching for the point it stops climbing and starts drawing down, historically a distribution signal rather than an accumulation one
• Short-term holder realized price relative to spot, since spot trading persistently below it has marked recent-buyer capitulation in past cycles
• Exchange reserves and illiquid supply, distinguishing self-custody outflows (reversible quickly) from ETF-custodian inflows (slower to reverse)
• Whale-wallet accumulation or distribution, while remembering the 1,000+ BTC bucket mixes real investors with custodians and exchanges
• ETF flow direction versus whale-wallet direction, specifically whether they continue to diverge or start moving together again

None of this replaces position sizing or risk management, which this series has covered separately in a dedicated positioning playbook. What it offers instead is a way to read the market’s plumbing rather than just its price tag, and enough historical grounding across two well-documented cycles to know roughly what a genuine accumulation or distribution signal has looked like before.

The honest summary going into Halving Five is that the wallets are telling a more nuanced story than either the pure bull or pure bear case would like. Long-term holders keep aging their coins to record highs, exchange balances keep thinning, and miners keep quietly selling what they must regardless of sentiment, while an entirely new institutional wrapper adds a fourth variable this dataset has fewer than two full cycles to judge. That is not a reason to ignore on-chain data. It is a reason to read it the way this piece has tried to: as a record of what happened, checked against its own limits, rather than a verdict on what happens next.

Frequently Asked Questions

What counts as a Bitcoin long-term holder on-chain?

On-chain analytics firms such as Glassnode classify any bitcoin that hasn’t moved in an on-chain transaction for at least 155 days as belonging to a long-term holder. It’s a statistical threshold based on when coins empirically stop correlating with short-term price swings, not a legal or tax definition; a country’s capital-gains holding period is a separate question entirely, with its own rules depending on where an investor is based and which platform they use.

Does the Bitcoin halving directly cause long-term holders to accumulate?

Not provably. The data shows correlation: long-term holder supply climbed heading into and following past halvings, hitting a then-record 12.66 million BTC in October 2020 and roughly 14.83 million BTC, about 76% of supply, by May 2026. But other factors, monetary policy, ETF availability and broader risk appetite among them, move at the same time, so attributing the behavior specifically to the halving rather than to the wider conditions surrounding it isn’t something on-chain data alone can prove. The most honest framing is that halvings and holder accumulation have coincided repeatedly, which is worth tracking, without treating that coincidence as a proven causal mechanism.

Why are Bitcoin exchange reserves falling?

A mix of self-custody adoption, accelerated by the 2022 FTX collapse and sustained hardware wallet demand, and the growth of spot Bitcoin ETFs, whose custodians hold coins in cold storage that rarely returns to exchange order books. CryptoQuant data has tracked exchange balances falling from roughly 2.8 million BTC in July 2021 to a 2.4 to 2.7 million BTC range by 2026, with illiquid supply covering around 72% of all mined bitcoin.

What is the difference between whale accumulation and ETF inflows?

They are different cohorts that behave differently. Whale wallets are simply on-chain addresses holding large balances, some of which are exchanges or custodians rather than individual investors, while ETF flows are regulated fund creations and redemptions reported daily with far less ambiguity about the mechanism, if not the end owner. In June 2026 the two diverged sharply: ETFs posted a record monthly outflow of $4.06 billion while whale wallets added an estimated 270,000 BTC over two weeks.

When is the next Bitcoin halving (Halving Five)?

Halving Five is expected at block 1,050,000. As of this writing Bitcoin has reached block 960,380, leaving about 89,620 blocks to go, and CoinGecko’s live countdown currently estimates the date at April 13, 2028. That estimate shifts slightly with network hashrate and should be treated as an approximation rather than a fixed date.

Written by the HOGE Wire markets desk.