Bitcoin Halving Cycle Math: What ETF Flows Actually Change

Bitcoin’s halving is still just arithmetic. Every 210,000 blocks, one line of code cuts the coinbase reward in half, and it has done exactly that four times since 2009 without missing a beat. What has changed is who sits on the other side of the trade when that arithmetic tightens the supply. For the first three halvings, the marginal buyer was mostly retail, often leveraged, and almost always exchange-based. Heading into the fourth halving and now the fifth, a meaningful share of demand arrives instead through a regulated wrapper that creates and redeems shares against custodied coins, reports its flows every trading day, and answers to allocators who have never touched a self-custody wallet. Spot Bitcoin ETFs are barely two and a half years old, yet they already hold something in the neighborhood of 1.3 million BTC, and in mid-2026 they just finished the longest losing streak of their short existence. That makes this a good moment to ask a narrower, more useful question than whether the four-year cycle is dead: does the arrival of ETF flows actually change the math behind the halving cycle, or does it just relabel the same boom-and-bust mechanics under a different ticker. Bitcoin is trading in the low-to-mid $60,000s as of mid-July 2026, per CoinGecko, roughly halfway between Halving Four and the currently estimated date for Halving Five. That midpoint, with real flow data from a full ETF cycle now in hand, is a reasonable place to take stock of what has actually changed and what has not.

The Halving Clock: What Actually Happens Every Four Years

Bitcoin’s issuance schedule is not tied to a calendar date. It is tied to block height, and it has been enforced by the same subsidy function since the genesis block: the reward halves every 210,000 blocks, a rule coded directly into Bitcoin Core’s GetBlockSubsidy function. Because blocks average roughly ten minutes, 210,000 of them work out to close to four years, but the exact date always drifts a little depending on how much hashpower is pointed at the network at any given time. That is why halving-date estimates move around in the months before the event and only firm up once the target block is close. The schedule runs all the way out to a 33rd and final halving, expected to round the subsidy down to zero near block 6,930,000, sometime around the year 2140.

Four halvings are in the books. The fifth is coming into view, though not yet close enough to pin to a single day.

As of this writing, Bitcoin sits at a block height near 957,985, leaving roughly 92,015 blocks, or a little under two years at the current pace, before the network reaches block 1,050,000 and the reward is cut again. CoinGecko’s halving countdown currently points to on or around April 13, 2028, a date that will keep sliding by a few days in either direction as block production runs faster or slower than the ten-minute target. Circulating supply sits at 20,056,171 BTC, or about 95.5% of the 21 million cap, worth close to $1.26 trillion at current prices, per CoinGecko. Bitcoin’s stock-to-flow ratio (the number of years of current production it would take to reproduce the existing supply) is close to 122 today; Halving Five will push that figure to roughly 245, putting Bitcoin’s new annual supply growth in the range commonly cited for above-ground gold stocks.

Four Cycles, Two Ways to Measure Shrinking Multiples

The four-year cycle became a market narrative because the first three halvings were each followed by a run to a new all-time high, and each of those highs represented an enormous multiple of the price on halving day itself.

Measured from halving-day price to the following cycle peak, the multiple has fallen from roughly 95x in the 2012-to-2013 run, to roughly 30x in 2016-to-2017, to roughly 8x in 2020-to-2021, to only around 2x from the April 2024 halving to October 2025’s all-time high of $126,198. A separate way of slicing the same pattern, comparing cycle peak to cycle peak rather than halving price to peak, produces a similar story: 2013’s peak near $266 gave way to a 2017 peak near $20,000, about a 75x jump; 2017’s peak led to 2021’s roughly $69,000 top, about 3.5x; and 2021’s peak led to 2025’s $126,000 top, about 1.8x, according to a CoinDesk analysis of the pattern. Both framings point at the same mechanical explanation: doubling a $10 billion asset and doubling a $1 trillion asset require entirely different amounts of new capital. One estimate puts it starkly: the 2011 cycle rode roughly $2.8 billion of net inflows to a 55,000% gain, the 2015 cycle needed about $69 billion for a 10,000% gain, the 2018 cycle needed about $365 billion for a 2,000% gain, and the cycle since 2022 has absorbed around $697 billion for a 689% gain. Repeating anything like a classic parabolic run from today’s roughly $1.26 trillion market cap would likely require well over $1 trillion of fresh capital.

That has not stopped forecasters from calling for another outsized move. Veteran trader Peter Brandt has pointed to a peak somewhere between $300,000 and $500,000, and Bernstein’s Gautam Chhugani and Mahika Sapra have a standing $500,000 target for 2029. Reaching even the low end of that range would require more than double the entire 2025 rally, which is exactly the kind of jump the last four cycles have been failing to repeat.

Enter the ETF: A Demand-Side Variable With No Precedent

None of the first three halvings happened alongside a regulated, exchange-listed product that let a retirement account or a registered investment advisor buy Bitcoin exposure without touching a crypto exchange. That changed on January 10, 2024, when the SEC approved 11 spot Bitcoin ETPs at once. Then-Chair Gary Gensler was careful to frame it as a legal formality rather than an endorsement: “we did not approve or endorse Bitcoin,” he said in the agency’s own statement, calling it a “speculative, volatile asset” also used for illicit activity. The approval came only after the SEC lost in court to Grayscale, and it landed inside a much larger, ongoing argument between the crypto industry and the agency over where enforcement ends and rulemaking begins, a fight that kept reshaping the SEC’s posture well beyond the ETF decision itself, as HOGE Wire has tracked in its explainer on SEC enforcement and rulemaking.

Two and a half years later, the scale is no longer trivial. Spot Bitcoin ETFs hold roughly 1.3 million BTC between them, something like 6.5% of the entire 21 million supply. That is a structural holder class that simply did not exist for three of Bitcoin’s four halvings, and it sits on top of, not instead of, the miners, long-term holders, and exchanges that made up the old supply-and-demand picture.

How Creation and Redemption Actually Move the Market

The mechanism matters because it determines whether ETF flows are a real proxy for spot buying and selling, or just a wrapper that trades on its own supply and demand. Spot Bitcoin ETFs work through authorized participants who create new shares by delivering cash (and, depending on the fund, sometimes Bitcoin itself) to the issuer, which then has the custodian acquire the underlying coins; redemptions run the same process in reverse. Because the funds are required to hold actual Bitcoin rather than futures or synthetic exposure, sustained inflows translate into real purchases of spot BTC, and sustained outflows translate into real sales, unlike the CME futures-based products that preceded them. That is precisely why the weekly and daily flow tables published by trackers such as Farside and SoSoValue get parsed by traders the way earnings reports get parsed by equity analysts: a green week is read as new capital actually entering the base layer, not just repositioning within it.

It also means the ETF complex concentrates execution in a small number of authorized participants and an even smaller number of custodians, a structural detail that becomes important later in this piece.

The Supply Squeeze Thesis: Bernstein’s Case for Half a Million Dollars

The most fully worked-out version of “ETFs change the math” comes from Bernstein. Analysts Gautam Chhugani and Mahika Sapra built a research note, originally published alongside their coverage of Strategy, around a straightforward collision: fixed 21 million supply, a subsidy that keeps cutting itself in half, and ETF assets under management they projected would grow from roughly $60 billion to around $190 billion in short order, alongside continued corporate accumulation. Their base case called for $200,000 by the end of 2025, $500,000 by 2029, and $1 million by 2033. The 2025 leg of that call did not land. Bitcoin’s actual all-time high came in at $126,198 on October 6, 2025, well short of $200,000, though still a genuine record.

Bernstein has not abandoned the longer-dated targets, and the logic behind them is not unreasonable on its face: a persistent, price-insensitive-ish buyer layered onto a supply curve that halves every four years is, mechanically, a squeeze. The open question is whether ETF demand is actually persistent in the way that logic requires, or whether it behaves like every other pool of capital that shows up during a rally and heads for the exit during a drawdown, a question the next section takes head on.

The Skeptic’s Rebuttal: Bitcoin Is Getting Harder to Move

The rebuttal, laid out in the same CoinDesk analysis that tracked the shrinking cycle multiples, does not really dispute that ETFs bring in new capital. It argues instead that the multiples were always going to shrink regardless of who the marginal buyer is, because a bigger asset is mechanically harder to move. Institutional participation, ETFs, and a deeper derivatives market have made Bitcoin, in the piece’s framing, “less volatile and more Wall Street-like,” which cuts against the odds of another 30x or 95x run no matter how much money funnels in through IBIT or FBTC.

Bitwise CIO Matt Hougan has made a related but distinct argument: that the four-year cycle itself, not just the size of its multiples, is breaking down. In a memo published in December 2025, he pointed to weaker leverage in the system after the October 2025 liquidations, falling interest rates heading into 2026, and the same institutional adoption wave (he cited Morgan Stanley, Wells Fargo, and Merrill opening allocation channels) as reasons the old halving-timed boom-bust rhythm no longer applies. There is a wrinkle worth flagging: by February 2026, with Bitcoin deep into its correction, Hougan was quoted citing the four-year cycle itself as a reason for the market’s losses, alongside gold and AI competing for capital and fears around a Federal Reserve nominee. The two positions are not strictly contradictory (a structural view about what drives multi-year trends differs from a tactical read on one drawdown), but the overlap is a useful reminder that even the analysts arguing the cycle is dead keep reaching for cycle language whenever price actually falls.

Realized Cap and the Strongest Evidence Something Changed

If there is one on-chain metric that supports the “something is different this time” camp, it is realized capitalization, which values each coin at the price it last moved on-chain rather than at the current market price. It is a rough proxy for the aggregate cost basis of everyone still holding, and unlike simple market cap, it does not move just because the last trade did.

Bitcoin’s realized cap crossed $1 trillion for the first time in July 2025. What is more notable is what happened next: through the fastest correction of any cycle, the roughly 36% drop from October’s all-time high inside six weeks, realized cap held above that $1 trillion mark instead of collapsing the way it did in 2018 and 2022. That is the clearest fingerprint of ETF-era buy-and-hold behavior in the data: capital that entered through IBIT, FBTC, and ARKB has not, so far, flinched the way exchange-based retail capital did in prior drawdowns. HOGE Wire covered what that drawdown actually demonstrated about the broader cycle debate in a companion piece on what the 2026 drawdown proved.

Not everyone reads persistent realized cap as evidence that ETFs have fixed the cycle’s old failure mode. CryptoQuant founder Ki Young Ju has pushed back on leaning too heavily on ETF flow as Bitcoin’s new engine, arguing in July 2026 that “Bitcoin needs to be a core macro asset, not just a retail-driven ETF trade.” It is a useful check on the supply-squeeze narrative from someone who is not a Bitcoin skeptic: a single, concentrated flow channel is not obviously more stable than the leveraged retail flow it replaced, just different.

When the New Buyer Panics: Inside the Eight-Week Outflow Streak

The clearest stress test of the ETF-as-stabilizer thesis happened in real time this year, and the timeline is worth laying out in full. Bitcoin hit its all-time high of $126,198 on October 6, 2025, during what traders nicknamed “Uptober.” What followed was the fastest three-year-cycle correction on record: a roughly 36% drop to around $80,000 within six weeks, a further slide to about $87,000 by the end of December 2025, and then, on February 12, 2026, Standard Chartered’s Geoffrey Kendrick cut his Bitcoin target from $150,000 to $100,000, his second downgrade in three months, warning of a possible drop toward $50,000. Bitcoin kept falling: around $71,000 by March 10, and an intraday cycle low of $59,375 on June 5, 2026, roughly 53% below the October high.

Spot ETFs mirrored the mood. They logged eight consecutive weeks of net outflows heading into early July 2026, the longest streak since the products launched and well past the previous record of five weeks, draining more than $8.2 billion in aggregate. BlackRock’s IBIT alone shed roughly $300 million in a single day on June 29. Kendrick called the bottom on June 12 (“Winter is over. Welcome back to crypto Spring,” he told CoinDesk) while keeping a $100,000 year-end target, but price stayed pinned in the high-$50,000s to low-$60,000s through most of June regardless.

The streak broke in stages rather than all at once: a single strong day around July 2 and 3 pulled in roughly $221 million and snapped the immediate losing run, and a full green week followed from July 6 to July 10, with about $197.4 million in net inflows led by IBIT’s $291.9 million and a further $95.1 million into Grayscale’s Bitcoin Mini Trust, the funds’ first clean positive week since May. By mid-July, Bitcoin had climbed back into the low-to-mid $60,000s.

Zoom out from that one good week, though, and the 2026 picture looks considerably less like a supply squeeze in progress. Bernstein’s Chhugani wrote to clients on July 6 that spot ETFs had actually shed a net $5.5 billion over the year against a roughly $74 billion asset base, and that combined ETF and corporate treasury inflows had slowed to about $10 billion in 2026, down sharply from $60 billion in 2025. The offsetting demand came mostly from treasury companies rather than ETFs: Strategy alone added roughly 175,000 BTC for about $14 billion during 2026, taking its total holdings to 847,363 BTC. Bernstein still called its own $150,000 year-end target “ambitious,” noting the current 54% drawdown is milder than the 75% to 90% declines that ended prior cycles, and shorter so far than the 12 to 15 months those corrections typically ran. If ETFs are going to dampen the four-year cycle the way their proponents expect, 2026 was the first real test, and the honest answer is: partially. Flows did reverse and price did recover off the low, but only after the longest, and on a net annual basis among the weakest, stretches of ETF demand since the products existed.

Stock-to-Flow, Power Law, and Why the Supply-Only Models Broke First

Long before ETFs entered the picture, the dominant halving-cycle framework was PlanB’s stock-to-flow model, published in 2019, which treated scarcity, the ratio of existing supply to new annual production, as close to the entire story. It called for a “worst case” of roughly $98,000 by November 2021 (actual: around $57,000 to $58,000) and implied a 2024-to-2028 price band widely cited as averaging around $500,000, with a range spanning roughly $250,000 to $1 million. Actual prices ran between about $59,000 and $71,000 through the first half of 2026, nowhere near that band. A widely cited structural critique of the model points out that since market value is just stock multiplied by price, regressing price against a stock-based ratio is partly regressing price against a function of itself, and its statistical fit largely disappears once the analysis properly accounts for autocorrelation.

Stock-to-flow’s core flaw was treating supply as sufficient on its own, with no explicit demand term. That is, in a sense, exactly the gap ETF flow data now fills, whether or not it fixes the forecasting problem. Other frameworks try to bridge the same gap differently. The power law model, associated with physicist Giovanni Santostasi, fits Bitcoin’s price to a function of time since genesis rather than the halving clock specifically, and has tracked reasonably well within wide bands, though its long-run fair-value estimates vary enough between sources that they are best treated as a rough corridor rather than a precise target. A fuller side-by-side comparison of these competing frameworks, including how each is actually performing against 2026 prices, is in HOGE Wire’s earlier piece on the road to Halving Five.

At least four distinct camps are now competing to explain what happens next, and the split increasingly runs along supply-side versus demand-side lines rather than pure technical disagreement:

• Stock-to-flow (supply only): missed its 2021 and 2024-2028 targets badly enough that most serious analysts have dropped it
• Power law (time and supply): fits Bitcoin’s growth curve to time since the 2009 genesis block rather than the halving clock alone, still widely cited but with wide error bands
• ETF supply-squeeze (Bernstein): treats ETF creation pace against fixed supply as the dominant price driver, underpinning the $500,000-by-2029 case
• Maturation or compression (the CoinDesk synthesis of the historical data): expects every future multiple to keep shrinking simply because the asset’s market cap keeps growing

The Miner Side of the Equation Still Matters

Halving math is not only a demand-side story. Every halving also cuts what miners physically receive, and often have to sell, to cover power and operating costs. Post-Halving-Four daily issuance runs around 450 BTC, half of what it was before April 2024 and a small fraction of the 7,200 BTC per day miners received in Bitcoin’s first years. That is structural sell pressure quietly cut in half again, on the same schedule that is supposedly squeezing supply for ETF buyers.

Miners are still overwhelmingly dependent on that subsidy rather than transaction fees, which have stayed under 5% of block reward through most of 2026. Hashprice, the dollar revenue a miner earns per unit of hashpower per day, sat around $29 per petahash per second in mid-2026, levels last seen around the 2020 post-COVID crash, even with network hashrate running in a roughly 780-to-900-plus exahash range depending on the snapshot. That combination, thin margins plus a reward that halves again in 2028, is why consolidation among public miners has become its own storyline heading into Halving Five; HOGE Wire covered the hashprice mechanics, breakeven math, and the resulting shakeout among miners in a dedicated piece. The long-run version of the same question, what replaces subsidy revenue as issuance keeps halving toward zero around the year 2140, gets sharper with every halving rather than easier, and it is a question no amount of ETF demand on the price side actually answers.

Custody Concentration: A New Single Point of Stress

The ETF era has introduced a dependency the four-year cycle never had to price before. Coinbase Custody Trust Company serves as custodian for the majority of U.S. spot Bitcoin ETFs, including BlackRock’s IBIT, Grayscale’s GBTC and Bitcoin Mini Trust, ARK 21Shares’ ARKB, and Bitwise’s BITB. BlackRock added Anchorage Digital as a second, backup custodian for IBIT in April 2025, but its own prospectus notes no current plan to actually move assets there, meaning the practical concentration has not changed much. Morgan Stanley picked the same combination, Coinbase for crypto custody and BNY Mellon for cash, when it built its own planned Bitcoin ETF, extending the pattern rather than breaking it.

None of this is a halving-schedule risk in the way a difficulty adjustment or a subsidy cut is. It is a new, ETF-era risk: an operational failure, security incident, or regulatory action at a single custodian now has more leverage over cycle-scale price action than any individual miner or exchange carried in 2013 or 2017, simply because so much of the float that used to sit across many exchange order books now sits behind one company’s cold storage. For context on how the custody and compliance posture of the major venues compares heading into this cycle, see HOGE Wire’s post-MiCA scorecard of Coinbase, Binance, Kraken, and OKX.

Counting Down to Halving Five, and the Halvings After It

Strip away the ETF debate and the underlying schedule has not moved an inch. Halving Five is due at block 1,050,000, currently estimated around April 13, 2028, cutting the reward from 3.125 BTC to 1.5625 BTC and daily new supply to roughly 225 BTC. Halving Six would follow on the same four-year rhythm, around 2032, cutting the reward again to 0.78125 BTC, and the schedule keeps halving every 210,000 blocks until the subsidy rounds down to zero near block 6,930,000. None of that code has changed once since 2009, and nothing in the ETF debate is going to change it before 2028 either.

What is genuinely new for this cycle, unlike the previous four, is the size and behavior of the buyer sitting on top of that fixed schedule. By Halving Five, stock-to-flow roughly doubles from about 122 to about 245, pushing Bitcoin’s marginal annual supply growth into gold-like territory by traditional commodity standards regardless of what ETFs do. The open question is whether a several-hundred-billion-dollar ETF and treasury-company complex responds to that shrinking supply the way Bernstein’s supply-squeeze case assumes, with price rising to clear a tighter market, or the way mid-2026 showed it can also behave: an air pocket that has very little to do with the halving clock and a great deal to do with rates, equities, or a handful of large holders’ decisions in a given quarter. Both outcomes are consistent with the same 210,000-block schedule; they just describe very different markets sitting on top of it.

What Would Actually Break the Playbook

A handful of scenarios would matter more to the next few years of Bitcoin’s price than anything in the halving code itself:

• A redemption wave longer or larger than mid-2026’s eight-week streak, which would test whether institutional capital is actually stickier than the retail leverage it partly replaced
• An operational or security failure at a concentrated custodian, given how much ETF-held Bitcoin now sits with one company
• A macro shock, rates, the dollar, or an equity drawdown, that pushes Bitcoin to trade purely as a risk asset; July 2026’s rebound was already driven more by a weaker dollar and an Asian semiconductor rally than by anything crypto-specific
• A reversal in SEC posture or a comparable regulatory shock in another major jurisdiction, reopening questions the 2024 approval was supposed to have settled
• Miners capitulating faster than the difficulty adjustment can smooth out, briefly straining network security economics around a halving

Every one of those is a demand-side or execution-side risk. The subsidy itself will keep cutting in half every 210,000 blocks with or without an ETF bid behind it; the argument that actually matters now is entirely about who shows up to buy what the code takes away, and how they behave the next time price falls rather than rises.

Frequently Asked Questions

What is Bitcoin’s halving cycle and how does the math work?

Bitcoin’s block reward halves every 210,000 blocks, roughly every four years given average ten-minute block times, a rule enforced directly in Bitcoin Core’s code. Four halvings have happened so far, in 2012, 2016, 2020 and 2024, cutting the reward from 50 BTC to the current 3.125 BTC per block. Each of the first three halvings was followed by a run to a new all-time high, though the size of those rallies has shrunk every cycle.

When is the next Bitcoin halving?

Halving Five is expected at block 1,050,000, currently estimated around April 13, 2028, though the date will shift slightly as actual block production runs faster or slower than the ten-minute average. It will cut the mining reward from 3.125 BTC to 1.5625 BTC per block, reducing new daily supply from roughly 450 BTC to about 225 BTC.

Do Bitcoin ETFs change the four-year halving cycle?

The evidence is mixed. Bitcoin’s realized cap held above $1 trillion through the fastest correction of any cycle instead of collapsing like in 2018 and 2022, and ETFs now hold roughly 6.5% of total supply. But spot ETFs also logged their longest-ever outflow streak, eight straight weeks into July 2026, and net ETF flows for all of 2026 turned negative even after a strong reversal week.

Why do Bitcoin’s cycle price multiples get smaller with every halving?

Mostly base effects. The multiple from halving-day price to the following cycle peak has fallen from roughly 95x in 2012-2013 to roughly 30x in 2016-2017, to about 8x in 2020-2021, to around 2x from 2024 into 2025’s all-time high of $126,198, because moving a much larger asset the same percentage requires far more capital.

What happened to the stock-to-flow Bitcoin price model?

It missed badly. PlanB’s stock-to-flow model projected a worst case of about $98,000 by November 2021, when the actual price was roughly $57,000 to $58,000, and implied a 2024-to-2028 price band averaging around $500,000, far above the $59,000-to-$71,000 range Bitcoin actually traded in through the first half of 2026.

Reporting and analysis by the HOGE Wire markets desk.