Bitcoin Halving Cycle Math: The Road to Halving Five

Why Halving Math Anchors Every Bitcoin Prediction

Bitcoin’s fourth halving cut the block subsidy from 6.25 BTC to 3.125 BTC on April 20, 2024, at block height 840,000. In the roughly two years since, the pattern has looked familiar in outline and unusual in detail: a rally to a new all-time high, a punishing correction, and a running argument about whether the math that produced three prior four-year cycles still holds now that a large share of Bitcoin sits in ETFs, corporate treasuries and sovereign allocations rather than retail wallets.

That argument keeps circling back to two things that can actually be measured: the supply schedule that halving events change directly, and the price and capital-flow data that halving events are supposed to influence indirectly. The fifth halving is expected around April 2028, roughly six hundred days from today, which makes this a reasonable point to work through the mechanics rather than the headlines. This piece works through what the supply code actually does, how the four completed cycles compare once market size is accounted for, and what would have to be true, mathematically, for a fifth parabolic run to happen at all.

The Supply Mechanics: How the Subsidy Actually Falls

Bitcoin’s issuance schedule is not a policy that can be revised; it is a rule enforced by every node running the reference client. The subsidy halves every 210,000 blocks, a calculation called GetBlockSubsidy in Bitcoin Core’s validation.cpp, which right-shifts the original 50 BTC reward by the number of halvings that have occurred. At an average ten-minute block interval, 210,000 blocks works out to roughly four years, though the actual gap between halvings has varied by several weeks each time depending on how much hashrate was pointed at the network.

The schedule terminates exactly rather than asymptotically: the reward rounds down to zero at block 6,930,000, expected around the year 2140, after 33 halvings in total. Four have happened so far, and the table below lays out the full history alongside the fifth halving’s current estimate.

Two consequences follow from that table. First, each halving’s effect on new supply shrinks in absolute terms even though it is proportionally identical: the move from 50 to 25 BTC per block removed 3,600 BTC a day from potential sell pressure, while the 2024 halving removed only 450 BTC a day, a much smaller absolute change despite being the same 50 percent cut. Second, annual inflation keeps falling along a fixed glide path. Post-2024 issuance runs at roughly 0.83 percent a year against a circulating supply of about 20.05 million BTC, or around 95.5 percent of the eventual 21 million cap, according to CoinGecko. The stock-to-flow ratio (circulating supply divided by annual issuance) sits near 122 today and is set to roughly double to about 245 once the fifth halving lands, comparable to the range of 60 to 90 typically cited for above-ground gold stocks.

This is also where Bitcoin’s design diverges sharply from proof-of-stake issuance models. A staking chain’s reward rate is usually a function of the active validator set and a protocol-set target, and it can be re-tuned through governance; Bitcoin’s schedule was fixed at genesis and cannot change without a hard fork the entire network would have to adopt. For a look at how issuance and rewards work on the other side of that divide, see our explainer on how stakers actually get paid.

Four Halvings, Four Diminishing Peaks

Every completed halving cycle has produced a new all-time high, and every completed cycle has produced a smaller peak multiple than the one before it. The pattern is clean enough to be the single most-cited chart on both sides of the “cycle is dead” debate: for the diminishing-returns camp it is the clearest evidence that returns keep shrinking on schedule, and for the “of course it’s shrinking” camp it is proof the trend was always going to keep bending toward smaller numbers as the asset grew, which says nothing about whether a cycle still exists at all.

Read across the multiple column and the shape is hard to miss: roughly 95x, then 30x, then 8x, then about 2x. Anyone arguing for a fifth cycle that resembles the first two has to explain why that curve would flatten out or reverse rather than continuing to compress toward 1x, which is arguably the more natural extrapolation of four consecutive data points. The counterargument, made by Bitwise CIO Matt Hougan among others, is that the mechanism that produced the pattern, a supply shock large enough to move price against a comparatively small market, has itself weakened, so the pattern it produced should not be expected to repeat even in decayed form. Whether that argument holds is really a question about capital, which the next section works through directly.

The Capital Efficiency Problem

The diminishing multiples above are, underneath everything else, a market-size problem. Every dollar of new capital moves a smaller asset further than it moves a larger one, and Bitcoin has been getting larger throughout its entire trading history. A CoinDesk analysis published July 4, 2026, built on CryptoQuant research, quantified exactly how much larger: in the 2011 cycle, roughly $2.8 billion of net new capital was associated with a gain of about 55,000 percent; the 2015 cycle needed about $69 billion for a 10,000 percent gain; the 2018 cycle needed roughly $365 billion for a 2,000 percent gain; and the cycle that began around 2022 has absorbed about $697 billion in net inflows for a gain of roughly 689 percent.

The same analysis makes the point a second way: in 2011, an estimated $5 million in fresh buying was enough to double Bitcoin’s price; doing the same in the current cycle required on the order of $101 billion, a difference of roughly four orders of magnitude in under fifteen years. CryptoQuant founder Ki Young Ju, whose data underpinned the analysis, aimed the finding at the next leg up rather than the last one: another parabolic run, he argued, would likely require more than $1 trillion in fresh institutional capital, a threshold that ETF outflows through mid-2026 suggest is not close to being met. “Bitcoin needs to be a core macro asset, not just a retail-driven ETF trade,” Ki Young Ju wrote, framing the finding as a case for patience rather than evidence of a market top. With Bitcoin’s total market capitalization above $1.2 trillion, a trillion dollars of fresh capital is not a small ask; it is roughly comparable to adding another Bitcoin’s worth of money to the asset from entirely new sources.

Where the 2026 Cycle Stands Right Now

The fourth cycle’s shape so far tracks the historical pattern more closely than the “cycle is dead” framing might suggest, even where the underlying drivers have clearly shifted. Bitcoin hit its cycle peak of $126,198 on October 6, 2025, a stretch traders nicknamed “Uptober.” What followed was the fastest three-year correction on record: a drop of roughly 36 percent to about $80,000 within six weeks, a further slide to around $87,000 by the end of December 2025, and continued weakness into the new year.

On February 12, 2026, Standard Chartered’s Geoffrey Kendrick cut his year-end Bitcoin target for the second time in three months, from $150,000 down to $100,000, and warned that a drop toward $50,000 was possible before any recovery. Bitcoin fell to about $71,000 by March 10 and kept grinding lower, eventually touching an intraday low of $59,375 on June 5, 2026, a drawdown of roughly 53 percent from the October high. A week later, on June 12, Kendrick called the bottom: “Winter is over. Welcome back to crypto Spring,” he said, while keeping his $100,000 year-end target intact. Bitcoin spent the rest of June in a range between the high $50,000s and low $60,000s before recovering to roughly $63,000 to $64,300 by July 9 and 10, a move CoinDesk attributed to a rally in Asian semiconductor stocks and a weaker dollar rather than to anything crypto-specific.

MEXC Research’s Shawn Young offered a caution about reading too much into the bounce: “once liquidations begin to drive price action, the market can move faster than real demand would justify,” pointing to leveraged positioning rather than fresh conviction as the proximate driver of both the crash and the recovery. Our earlier piece on what the 2026 drawdown actually proved goes deeper into that stretch of price action. The short version: a roughly 53 percent peak-to-trough drawdown inside a halving cycle is not new, 2021 to 2022 delivered a similar move, which is one reason the “cycle is dead” claim stays contested even among people who agree institutional ownership has changed the market’s composition.

Is the Four-Year Cycle Actually Dead?

The clearest single data point in this debate is also the most ironic. In a December 16, 2025 memo titled “The Four-Year Cycle Is Dead,” Bitwise CIO Matt Hougan argued that the forces behind prior cycles, halving-driven supply shocks, a rising-rate environment and leverage-fueled boom and bust swings, had all weakened at once: the halving is mechanically half as important every four years by definition, interest rates were headed down rather than up heading into 2026, and there had been no comparable leverage blowup in 2025. He predicted new all-time highs for 2026.

Two months later, with Bitcoin down sharply, CNBC reported that Hougan pointed to the four-year cycle as a leading explanation for the losses, alongside competition from gold and AI stocks for investor attention and lingering worries about a more hawkish Federal Reserve leadership pick. The two positions are not strictly contradictory: the December memo was about structural, multi-year drivers weakening, not about Bitcoin becoming immune to near-term drawdowns. But the timing made for an awkward headline, and it shows how quickly the “is it dead” framing can flip depending on which week’s price chart happens to be on screen.

Morgan Stanley’s Denny Galindo offered a different lens in late 2025, describing Bitcoin’s history as a “three-up, one-down” rhythm rather than a strict four-year cycle, and characterizing the period around Bitcoin’s correction as the market’s “fall season.” “We are in the fall season right now,” Galindo said, “it’s the time you want to take your gains,” treating the pattern as a seasonal guide for position-sizing rather than a hard price model. Fidelity’s Jurrien Timmer took the most literal view among widely quoted analysts, arguing the 2025 to 2026 stretch lined up “about perfectly” with the prior three cycles and that the bearish phase should be expected to run deep into 2026.

We covered this debate in full, including the on-chain evidence on both sides, in Is the Four-Year Cycle Dead? The short version: nobody with a public track record has actually called the cycle dead and been proven right over the following twelve months yet, which is itself informative.

Stock-to-Flow’s Broken Promise

PlanB’s stock-to-flow model, published in 2019, was the first widely popularized attempt to turn halving math directly into a price forecast. It regressed Bitcoin’s price against its stock-to-flow ratio (circulating supply divided by annual issuance) and extrapolated a “worst case” price of about $98,000 by November 2021; the actual price that month was roughly $57,000 to $58,000. The model fared worse from there: Bitcoin fell to around $15,500 in 2022, more than 80 percent below even the model’s stated floor, and the band widely cited for 2024 through 2028 averaged somewhere around $500,000, with a quoted range of roughly $250,000 to $1 million, against actual prices that ran between about $59,000 and $71,000 through the first half of 2026.

The model’s problems turned out to be structural rather than a matter of bad luck. A widely cited Bitcoin Magazine analysis argues the regression is tautological: market value decomposes to stock multiplied by price, while stock-to-flow places stock on the other side of that same equation, meaning the model effectively tests stock against a transformation of itself rather than against an independent variable. The same critique found the model’s statistical fit collapses toward an R-squared near zero once the data is corrected for autocorrelation, a basic requirement for time-series regression that the original model skipped. PlanB has since offered restated versions with different parameters, but critics note those parameters shift enough between iterations to look more like curve-fitting after the fact than a stable law.

None of this means halving-driven scarcity is irrelevant to price, only that turning “supply issuance falls on a known schedule” into “price will therefore reach a specific number on a specific date” required statistical assumptions that did not hold up. That distinction matters for the rest of this piece, because it is exactly the line between mechanics, which are knowable, and forecasts, which are not.

Realized Cap and ETF Flows: The New Cycle Variables

If stock-to-flow models the supply side and mostly ignores demand, realized cap and ETF flow data are an attempt to measure demand directly, and they are the two data series bulls point to most often as evidence that this cycle behaves differently from the first three.

Realized capitalization values each coin at the price it last moved on-chain rather than at the current spot price, making it a rough proxy for the aggregate cost basis of the entire holder base. It first crossed $1 trillion in July 2025 and, unlike in the 2018 and 2022 bear markets, it did not fall alongside price during the correction that followed the October 2025 peak. CoinDesk reported it holding at a record high above $1 trillion even as spot price fell nearly 40 percent, based on Glassnode data. By mid-2026 it had climbed further, to around $1.125 trillion. Capital that does not exit during a correction is a meaningfully different signal than capital that does, and it is the strongest single data point for the argument that this cycle’s composition, more ETF and treasury holders, fewer short-term speculators, has genuinely changed.

ETF flows tell a messier story. Spot Bitcoin ETFs strung together eight consecutive weeks of net outflows heading into July 2026, the longest streak since the products launched in January 2024 (the previous record was five weeks), pulling more than $8 billion out of the funds in total, according to crypto.news. BlackRock’s IBIT, the largest fund by assets, lost roughly $300 million in a single day on June 29. The streak briefly interrupted itself on July 3, when the funds pulled in a net $221 million, snapping a ten-trading-day daily losing streak, led by Fidelity’s FBTC fund, which accounted for roughly three-quarters of that day’s inflow even as IBIT posted another outflow the same session, according to CoinDesk. The eight-week streak on a weekly basis broke fully the following week. US spot Bitcoin ETFs collectively hold around 1.3 million BTC, or roughly 6.5 percent of circulating supply, meaning flows into or out of these funds now move a meaningful share of tradable supply in a way no single product could before 2024.

Miners Are Already Pricing In Halving Five

Miners feel halving math before anyone else does, since it cuts their primary revenue source overnight. Hashprice, the dollar value a miner earns per petahash of computing power per day, fell to roughly $29 per PH per second per day in mid-2026, a level last seen during the post-COVID crash of 2020, even as network hashrate simultaneously set records above 900 exahashes per second. That combination, falling revenue per unit of hashpower alongside record total hashpower, is survivable mainly because of efficiency gains in newer mining hardware and a wave of diversification into AI and high-performance computing hosting by public miners looking to offset thinning margins with a second revenue line. Transaction fees, which many expected to become a larger share of miner revenue as the subsidy shrank, still make up less than 5 percent of the block reward, leaving miners more exposed to the next halving than the protocol’s original design assumptions anticipated.

The pattern from prior halvings is that inefficient miners capitulate in the months around the event, hashrate dips briefly, and the difficulty adjustment (Bitcoin’s built-in two-week retargeting mechanism) brings block production back toward the ten-minute target. Whether that plays out the same way in 2028 depends heavily on how much of today’s hashrate belongs to operators who no longer need Bitcoin mining revenue alone to stay solvent. We covered the mechanics of that squeeze, including breakeven math for major public miners, in Bitcoin Mining Margins Explained.

On-Chain Signals: What MVRV Says About Cycle Position

Market Value to Realized Value, or MVRV, compares Bitcoin’s spot market capitalization to its realized capitalization, and the Z-Score version of the metric is a commonly cited tool for judging how stretched a given price level is relative to the aggregate cost basis of holders. At the 2017 top, the MVRV Z-Score reached roughly 10; at the 2021 top, it peaked around 7. Through the run into the October 2025 high, it only reached the mid-single digits, well below either prior peak, before falling back sharply during the correction that followed (different data providers report somewhat different exact readings for early 2026, so any single decimal figure should be treated as approximate rather than as an agreed-upon precise number).

The directional read matters more than the exact figure: on this metric, the market did not get anywhere near as overheated at the October 2025 peak as it did at the 2017 or 2021 tops, which lines up with the realized-cap story above (steadier, less speculative capital) but cuts against a simple mean-reversion argument that Bitcoin was “due” for a correction of the size it actually got. That mismatch, a moderate on-chain reading followed by a historically fast drawdown, is part of why MEXC’s Shawn Young pointed to leverage and liquidations rather than valuation as the proximate trigger for the crash, a dynamic that spot-based on-chain metrics like MVRV are not well suited to capture on their own, since they say nothing about derivatives positioning.

Modeling the Road to Halving Five

Extrapolating from four data points is a risky habit, and this piece has spent most of its length explaining why the two most popular attempts to do exactly that, stock-to-flow’s price regression and a simple “multiply the last cycle’s return” heuristic, have already failed or are actively failing. What can be stated with more confidence is the supply side alone, since that part is code, not sentiment.

Between now and the fifth halving, expected around block 1,050,000 in April 2028 (halving countdown trackers put the date roughly 600 to 650 days out as of early July 2026, moving slightly as block times fluctuate), the network keeps issuing new Bitcoin at a fixed, known rate of 450 BTC a day, or 3.125 BTC per block. Across the full interval between the fourth and fifth halvings, that adds up to a maximum of 210,000 blocks multiplied by 3.125 BTC, or 656,250 new BTC, entering circulation over roughly four years. Circulating supply, about 20.05 million BTC as of mid-2026, keeps climbing toward the 21 million cap at that fixed, slowing rate until the fifth halving cuts issuance again, to 225 BTC a day and a stock-to-flow ratio near 245.

That is the entire supply-side forecast anyone can make with real confidence: a known, shrinking, disinflationary issuance curve that does not depend on price, sentiment, regulation or ETF flows. Everything else, whether continued scarcity translates into a fifth cycle peak, a slow grind, or something that no longer resembles a “cycle” in the traditional sense at all, depends on the demand side covered above: whether ETF and treasury capital resumes sustained net inflows, whether a genuinely new buyer base (sovereign funds, insurers, pension allocators) enters at the scale Ki Young Ju’s trillion-dollar threshold implies, and whether interest rate policy stays supportive through 2027 and 2028 the way Hougan’s original thesis assumed it would.

The SEC’s approval of the first eleven spot Bitcoin ETPs in January 2024 remains the single regulatory event that most changed how that demand side can express itself. Then-Chair Gary Gensler framed the approval narrowly, stating at the time that “we did not approve or endorse Bitcoin,” and describing it as a “speculative, volatile asset.” For more on how that kind of regulatory line-drawing has evolved since, see our explainer on SEC rulemaking in 2026.

Three Forecasting Frameworks, Compared

Stock-to-flow is not the only attempt to turn halving math into a forecast, and it is worth being specific about how the alternatives differ rather than lumping them together as one big category of “price prediction models.”

The power law model, most associated with independent researcher Giovanni Santostasi, treats Bitcoin’s price as a function of time since the genesis block raised to a fixed exponent, rather than as a function of the halving schedule directly. Because it does not reference issuance or scarcity at all, a halving shows up in the power law framework only indirectly, through whatever effect it has on adoption and price over time, not as an input to the formula itself. Proponents argue this makes it more robust than stock-to-flow, since it cannot suffer the same tautology critique; skeptics note that a curve fit to Bitcoin’s entire trading history will always describe that history reasonably well, and the real test, how tightly future prices track the band, is one the model is still in the middle of as of mid-2026.

Morgan Stanley’s seasonal framing, the “three-up, one-down” rhythm Denny Galindo described, is explicitly not a price model at all. It makes no numeric prediction and instead offers a qualitative rule for position-sizing: take profits when the pattern suggests a correction is due, add exposure when it suggests accumulation. Its strength is also its weakness. It is nearly impossible to falsify with a single bad quarter, but it also cannot tell an investor how far a correction will run or how high a rally will reach, only roughly when to pay attention.

A third camp, associated with Hougan’s original December 2025 thesis and Ki Young Ju’s macro-asset framing, argues that no calendar-based model should be expected to work going forward at all, because the buyer base has changed from retail speculators trading on sentiment to institutions trading on capital allocation rules, macro conditions and balance-sheet policy. This is less a forecasting framework than an argument for retiring cyclical forecasting altogether, in favor of treating Bitcoin like any other macro asset: correlated with rates, dollar strength and risk appetite rather than with its own halving calendar.

What Would Have to Be True for Another Parabolic Run

Pulling the threads together, a fifth cycle that looks like the first two (a 30x or 95x move) is arithmetically difficult to construct at Bitcoin’s current size. Even matching the fourth cycle’s already-diminished roughly 2x would require, per the capital-efficiency math above, inflows on an order not yet seen from any single source, ETF, sovereign fund or corporate treasury, in the asset’s history. A more modest outcome, a new all-time high without a parabola like 2013 or 2017, is far more consistent with both the diminishing-multiple trend and the capital-efficiency data, and it is roughly what Kendrick’s $100,000 year-end target and Hougan’s original December 2025 thesis both implied before the year’s drawdown made either look optimistic in the short term.

• Sustained ETF and institutional inflows resuming, rather than a repeat of retail-driven leverage
• Realized cap holding its floor near $1.125 trillion through further price swings, not just during this one drawdown
• A genuinely new buyer base, sovereign funds, insurers, pension allocators, entering at a scale that can plausibly approach Ki Young Ju’s $1 trillion threshold
• Interest rate policy staying supportive into 2027 and 2028 rather than reversing

The eight-week ETF outflow streak into July 2026 and its uneven reversal are arguably the single most important data series to watch between now and 2028, more so than any price level in isolation, because every framework covered in this piece, realized cap holding steady, the capital-efficiency threshold, the macro-asset argument itself, depends on new capital actually showing up rather than existing holders simply declining to sell. Realized cap holding near $1.125 trillion through a 53 percent price drawdown has no clean precedent in the 2018 or 2022 cycles. Whether that represents a permanently higher floor or just a slower-moving version of the same cycle is likely to stay unresolved until well after the fifth halving actually happens, at which point there will be a fifth data point in the diminishing-multiples table above, and a considerably stronger basis for saying whether the four-year pattern actually broke or just became harder to see in real time.

Frequently Asked Questions

When is the next Bitcoin halving?

The fifth halving is expected around April 2028, when the block reward falls from 3.125 BTC to 1.5625 BTC at block height 1,050,000. The exact date depends on average block times over the next couple of years, which is why countdown estimates shift by a few weeks in either direction as the date gets closer.

Does a Bitcoin halving guarantee the price will go up?

No. A halving mechanically cuts new supply issuance, but it does not create new demand on its own. Prices rose sharply after all four halvings to date, but each cycle’s peak to halving day multiple has shrunk, roughly 95x, then 30x, then 8x, then about 2x, and the 2026 drawdown showed that a halving cycle can still include a correction of more than 50 percent from its high.

Is the Bitcoin four-year cycle dead in 2026?

Analysts are split. Bitwise CIO Matt Hougan argued in December 2025 that heavier institutional ownership and ETF flows had weakened the old boom and bust pattern, then pointed to that same cycle as a leading cause of 2026’s losses two months later. Standard Chartered’s Geoffrey Kendrick called a cycle bottom in June 2026 while keeping a $100,000 year-end target, and Morgan Stanley researchers described 2026 as the fall phase of the familiar rhythm rather than proof the pattern had broken.

Why did the stock-to-flow model fail?

PlanB’s stock-to-flow model projected a worst case price near $98,000 for November 2021 and an average price around $500,000 for 2024 through 2028; actual prices came in far below both figures. Critics, including a widely cited Bitcoin Magazine analysis, argue the regression is tautological because it tests a supply-based ratio against a price series that is itself partly a function of that same supply figure, and that its statistical fit largely disappears once the data is corrected for autocorrelation.

How many Bitcoin halvings are left before new supply stops?

Bitcoin’s block subsidy keeps halving roughly every four years until it rounds down to zero at block 6,930,000, expected around the year 2140, for 33 halvings in total. That leaves 29 halvings after the fifth one in 2028, though about 95.5 percent of the eventual 21 million supply is already in circulation today.

Written by the HOGE Wire markets desk.