Bitcoin Below Mining Cost: The Floor Stock-to-Flow Missed
PlanB's stock-to-flow model said scarcity would push Bitcoin toward $500,000. In November 2026, with BTC near $63,800 and below the cost to mine it, a different floor is doing the real work.
Scarcity Promised $500,000; the Market Is Testing the Mining Bill
On 19 November 2026, Bitcoin changed hands near $63,800, down about 13% over seven days and roughly 49% below the record just above $126,000 it set on 6 October 2025, according to CoinGecko. That price matters for one specific reason: it sits below almost every serious estimate of what it costs to mine a single coin. It also sits a very long way below $500,000, the figure the most famous Bitcoin valuation model of the last decade has pointed at for this cycle.
The $500,000 comes from stock-to-flow, the scarcity model built by the pseudonymous analyst PlanB. Its logic is that Bitcoin’s hard-coded supply schedule, tightened again by the April 2024 halving, should drag price toward the hundreds of thousands. The mining bill comes from a different and much older idea: that the price of a produced commodity gravitates toward the cost of producing it. In 2026 the two views could hardly be further apart, and the market keeps settling near the second one, not the first.
This is not another victory lap over a model that has missed. It is an argument about floors. Stock-to-flow treats scarcity as a destiny that sets price from the top down. Production cost describes a feedback loop that works from the bottom up, enforced by real machines, real electricity bills, and a difficulty adjustment no one controls. Both get called a floor. Only one of them has a mechanism behind it, and in November 2026 you can watch that mechanism operate in close to real time, as miners switch off and the network recalibrates around them.
What Stock-to-Flow Actually Claims
Stock-to-flow is borrowed from commodities. Divide the existing above-ground stock of an asset by the new flow produced each year and you get a ratio that climbs as fresh supply slows. Gold scores roughly 62 on this measure, silver around 22. Bitcoin sat near 27 before the 2020 halving and jumped to about 119 after the 2024 halving, per CoinGecko’s own explainer. The higher the ratio, the scarcer the flow relative to the stockpile, and the theory says the scarcer the flow, the higher the price.
PlanB’s 2019 paper, Modeling Bitcoin’s Value with Scarcity, fit a regression between that ratio and Bitcoin’s market value and reported an R-squared near 95%, a correlation so tight it looked like a law of nature. A cross-asset version in April 2020 put Bitcoin’s average value for 2020 to 2024 near $288,000. The current, widely quoted target is an average of about $500,000 for the 2024 to 2028 cycle, with a stated range of $250,000 to $1 million, as summarized in coverage such as this Yahoo Finance report.
Crucially, the model has been refit more than once. The original 2019 version implied an average near $55,000 for the 2020 to 2024 cycle; the cross-asset upgrade a year later lifted that toward $288,000 with an even higher reported R-squared, and the current cycle target settled at $500,000. Each revision arrived with a fresh chart that fit the past beautifully, which is exactly what a model does when it is describing history rather than predicting the future. A framework that keeps redrawing its own target to match where the price has already been is not forecasting; it is tracing.
The appeal is obvious. Scarcity is measurable, the halving dates are known years ahead, and a single chart seemed to turn Bitcoin into a stopwatch for wealth. The trouble is that a correlation is not a mechanism. Nothing in the stock-to-flow ratio explains why a buyer would pay more; it simply assumes they will, because the supply got tighter. That assumption is the hinge the whole model swings on, and it is the assumption 2026 has been quietly prying loose.
The One Thing the Model Gets Right
Give stock-to-flow its due: the supply half of the equation is flawless. Bitcoin’s issuance is not a forecast, it is code. The subsidy halves every 210,000 blocks, and the April 2024 halving cut the block reward from 6.25 to 3.125 BTC, dropping daily issuance to roughly 450 coins and annual supply growth to under 1%. About 20.1 million BTC now exist, close to 95.7% of the 21 million cap. The flow tightened exactly on schedule, exactly as the model assumed it would.
And that is precisely the problem. The scarcity event the model is built around happened, on time, and the price went the other way. Bitcoin set its record five months after the 2024 halving, then fell. By November 2026 it is down roughly 49% from that high while the stock-to-flow ratio sits at the highest level in its history. When the one input a model controls fires perfectly and the output still misses by a factor of eight, the model is not early. It is wrong about what drives the output.
This is the cleanest possible falsification, and it needs no statistics to see. The halving is the test stock-to-flow set for itself. It ran in 2024, the supply shock arrived precisely as coded, and the price did not follow the curve. Scarcity, it turns out, is a necessary ingredient in Bitcoin’s story and a hopeless predictor of its price. A model that is right about the one thing that cannot fail and wrong about the thing that matters has told you almost nothing you can trade on.
The Floor Stock-to-Flow Never Modeled
If scarcity does not set the floor, what does? The oldest answer in the Bitcoin canon is the cost of production. Satoshi Nakamoto put it plainly in a 2010 forum post, writing that the price of any commodity tends to gravitate toward the production cost. If price sits below cost, production slows; if it sits above cost, miners pile in, difficulty rises, and the cost of winning the same reward climbs back toward the price. It is a self-balancing loop, and it is exactly the mechanism stock-to-flow is missing.
Charles Edwards, founder of Capriole Investments, turned that principle into a measurable model in 2019. His Energy Value and Bitcoin Production Cost framework prices a coin from the joules of energy burned to mine it, using network hashrate, hardware efficiency, and an electricity cost drawn from Cambridge University data, then grossed up for the hardware, rent, and wages that sit on top of the power bill. Energy alone, Edwards argues, maps the fair value of Bitcoin from inception, with no wacky formulas and no power laws. Over five years, he notes, Bitcoin’s electrical cost of production acted as a floor the market price kept bouncing off, which is the data-backed version of Satoshi’s commodity principle.
Edwards draws a useful distinction between two lines. The total production cost, which includes every expense, is a soft floor; price can slip under it for a while as higher-cost miners bleed. The electrical cost alone, the pure power bill, is the harder floor, because a miner will keep running as long as each coin clears the marginal cost of the electricity to make it, even at a loss on paper. Bitcoin’s price has historically come close to that electrical floor and rarely broken it. Where the market sits between those two lines, Edwards says, has tended to mark the best long-term entries. That is a floor with a reason, not a floor by assertion.
What It Costs to Mine a Coin in 2026
The estimates cluster in a band well above today’s price. JPMorgan’s team, led by managing director Nikolaos Panigirtzoglou, pegs the production cost for listed miners at about $78,000 a coin, with private operators closer to $88,000, as reported by 24-7 Wall St. CoinShares put the industry’s weighted cash cost at roughly $75,500 in the second quarter. One independent miner-tracking model, highlighted in a KuCoin market insight, placed the figure near $88,800 in mid-November, with Bitcoin trading about 15% below it.
Put those side by side with a spot price in the low $60,000s and the picture is stark: the asset is changing hands for less than it costs the median public company to produce. That is the condition stock-to-flow says is impossible, because in a pure scarcity frame price only climbs as the ratio rises. It is also a condition the production-cost view expects to be temporary, resolved not by a scarcity rally but by miners switching off until the economics heal and the cost of production itself comes down to meet the price.
Why do the estimates disagree at all? Because the cost to mine a coin is not one number; it is a distribution that spans an order of magnitude. It turns on the price a miner pays for power, which runs from a few cents per kilowatt-hour in subsidized jurisdictions to several times that on an open grid, on the efficiency of the hardware measured in joules per terahash, and on whether the operator owns its machines or rents hosting. Add depreciation on fast-aging rigs and the all-in figure can sit $10,000 or more above the bare cash cost. The headline number is really a weighted average of a very wide field.
| Estimate | Cost to mine 1 BTC | Source |
|---|---|---|
| JPMorgan, listed miners | about $78,000 | Panigirtzoglou via 24-7 Wall St, Sep 2026 |
| JPMorgan, private operators (all-in) | about $88,000 | Panigirtzoglou via 24-7 Wall St |
| CoinShares weighted cash cost (Q2) | about $75,500 | CoinShares Q2 2026 report |
| Independent model (mid-Nov) | about $88,800 | On-Chain Mind via KuCoin |
| Spot price, 19 Nov 2026 | about $63,800 | CoinGecko |
The Difficulty Thermostat: the Loop Scarcity Models Lack
Here is the part stock-to-flow has no equivalent for. Roughly every two weeks, the Bitcoin network retargets its mining difficulty so that blocks keep arriving about every ten minutes. When price falls and the least efficient miners power down, blocks slow, and the next adjustment lowers difficulty, which lowers the cost of producing a coin for everyone still running. The cost of production is not a fixed line under the price; it is a thermostat that chases the price down, then warms back up when the price recovers and machines switch on again.
That loop was visible to the hour in November 2026. Blocks had been arriving slowly, averaging close to eleven minutes, because machines were going dark. On 13 November the network delivered its steepest difficulty cut since June, a drop analysts pegged at roughly 7%, pulling difficulty from about 148 trillion toward the mid-130-trillions, per Crypto Briefing. Hashprice, the daily revenue per unit of mining power, hovered near $33 per petahash, close to its low for the year. The adjustment eased the pressure, but as one analysis put it, that is a relief, not a turning point.
Panigirtzoglou’s team describes the sequence in a single sentence: when Bitcoin trades below its production cost, higher-cost miners power down, the hashrate declines, and difficulty adjusts lower. It is the opposite of a scarcity spiral. It is a cost curve bending to meet a falling price. Readers watching the security budget shrink can trace the same story in our coverage of how hashrate growth reversed once Bitcoin fell below cost.
| Adjustment | Difficulty change | Context |
|---|---|---|
| 14 Jun 2026 | about -10.1% | Price under cost; hashrate fell roughly 12% in June |
| Oct 2026 (two steps) | about +11% combined | Seasonal hashrate rebound as machines returned |
| 13 Nov 2026 | about -7% | Steepest cut since June; blocks near 11 minutes |
Below Cost, Twice in One Year
November is not the first time in 2026 the market dipped under the cost line. Bitcoin traded below the roughly $78,000 production-cost estimate for five straight months through June, the longest such stretch of the cycle. In early June the difficulty adjustment fell about 10%, the network’s hashrate dropped roughly 12%, and by late summer hashrate was sitting around 23% below its October 2025 peak. Miners did what the model predicts: they shut off machines and, where they could, sold coins to cover bills that arrive in dollars every month.
The selling was heavy. In the first quarter alone, six public miners, among them Marathon, CleanSpark, Riot, Cango, Core Scientific, and Bitdeer, sold a combined 32,000 BTC, exceeding their total 2025 sales and setting a quarterly record. By the autumn, Riot had sold the large majority of the Bitcoin it produced in the quarter, and Hut 8 offloaded more than 3,000 coins, worth roughly $233 million, in a single reporting period. None of this is scarcity at work. It is distressed supply hitting the market precisely because production was underwater.
That is the behavior a stock-to-flow chart cannot contain. The model counts the shrinking issuance and concludes the float is tightening. On the ground, the opposite happened: a wave of coins came loose from the balance sheets of companies that could no longer afford to hold them. The flow the model watches, roughly 450 freshly minted coins a day, is small next to the flow the model ignores, the treasuries of miners and other holders who sell when the price of making a coin exceeds the price of selling one.
Why “Floor” Is the Wrong Word for Both
Honesty requires a caveat. Production cost is not a magic line either. Bitcoin is below it right now, which is exactly why the word floor flatters both models. The difference is what each theory expects to happen next. Under stock-to-flow, a price below $500,000 is simply an error waiting to be corrected by scarcity, with no deadline and no mechanism to force the correction. Under the production-cost view, a price below cost triggers a visible, dated response: shutdowns, a difficulty cut, a lower cost of production, and eventually a cleared market.
Edwards’ two-line framing captures why. The total production cost is a soft floor that price can pierce during capitulation, and it is being pierced now. The electrical cost, the bare power bill, is the harder floor, and it sits well below the all-in number because so much of a miner’s cost is sunk hardware and fixed contracts rather than this month’s electricity. A rational miner keeps hashing as long as each coin clears the marginal cost of the power to make it. That is the line that has historically held, and it is far lower than the headline $78,000.
So the honest claim is not that production cost is an unbreakable floor. It is that production cost is a floor with a motor: when price drops through it, the network mechanically works to pull the cost back down to meet the price, and distressed miners stop selling once they are no longer losing money on every coin. Stock-to-flow has nothing like this. Its $500,000 is a destination with no road and no engine, a price with no account of how the market would ever get there.
Scarcity Without Demand: the Gold Problem
The deepest flaw in stock-to-flow is that it is a supply model in a market set by demand. Gold is the tell. Gold carries the highest stock-to-flow ratio of any real commodity, roughly 62, precisely because almost all the gold ever mined still exists and annual production barely moves the pile. Yet gold’s price is not set by that ratio. It is set by real interest rates, central-bank buying, and the dollar. The single commodity stock-to-flow leans on for credibility is the clearest proof that a high ratio does not dictate a price.
Bitcoin now scores about 119, nearly double gold, and has fallen by roughly half from its record. If scarcity set price, that could not happen. What actually moved Bitcoin from 2024 to 2026 was a parade of demand-side forces the model has no slot for: spot ETF flows and their reversal, a hawkish Federal Reserve, swings in dollar liquidity, forced miner selling, and a broad risk-asset drawdown. The halving is one input among many, and in a year like this one, not the loudest. A ratio that only measures supply is deaf to every voice on the other side of the trade.
The point is not that scarcity is meaningless. A fixed supply is what makes Bitcoin worth modeling as a monetary good at all. The point is that scarcity is a constant, baked in from the start, while price is a variable that moves every day. A constant cannot explain a variable. Through 2026 Bitcoin traded less like a scarce rock and more like a long-duration risk asset, falling when liquidity tightened and the Federal Reserve leaned hawkish, exactly the behavior you would expect from an asset priced by demand rather than by the slope of its issuance curve.
The Drivers Stock-to-Flow Cannot See
Strip the year down to what actually moved the tape and stock-to-flow accounts for a single line of it. The spot ETFs changed the marginal buyer from retail to institutions that price policy and liquidity months ahead, a shift laid bare when 2026 became the ETF era’s first genuine bear market. The Federal Reserve’s rate path and the regulatory calendar moved prices more than any halving math, as our rundown of the lame-duck rulebook details. Even the yield that institutions chase has migrated on-chain into tokenized Treasuries and stocks, a demand channel invisible to a supply ratio.
The table below is the whole case in one frame. Of the forces that set Bitcoin’s price across the last cycle, stock-to-flow can see exactly one, and it is the one that was never in doubt.
| Price driver, 2024 to 2026 | Does stock-to-flow see it? |
|---|---|
| The 2024 halving (supply) | Yes, this is the whole model |
| Spot ETF flows and outflows | No |
| Federal Reserve policy and rates | No |
| Dollar liquidity conditions | No |
| Forced miner selling below cost | No |
| Broad risk-asset drawdown | No |
Even PlanB Watches the Mining Bill Now
The most telling endorsement of the production-cost view comes from the author of stock-to-flow. PlanB still reiterates the $500,000 target and insists the October 2025 record was not the cycle top. But when he talks about reading the market day to day, he no longer reaches for his own ratio. In September 2026 he told followers that mining difficulty “better reflects market price than models such as S2F or power laws,” as reported by KuCoin, framing a rebound in difficulty as a sign miners were reinvesting and a bottom was firming.
Read that twice. The creator of the scarcity model, asked where the price is, points at the mining economics, not the stock-to-flow curve. It is the clearest possible admission that the useful signal lives in the cost of production, while the $500,000 lives on the wall as a slogan. His track record reinforces the split: the 2021 floor model that promised a $98,000 close in November 2021 instead watched Bitcoin close near $57,000, a miss PlanB himself called a “big miss and the first in 10 years,” per CryptoPotato. The flexible target survived; the datable prediction did not.
It is worth being fair to him. PlanB has never formally retired stock-to-flow, and he frames the $500,000 as a rough cycle average rather than a timing tool. But the tools he actually cites now tell the story: moving averages, realized price, the relative strength index, and above all mining difficulty, all of which are measures of market behavior and production economics, not of scarcity. When the author of the scarcity model reaches for the cost curve to find the bottom, the debate is effectively over.
What the Critics Said, and What the Data Shows
The academic case against stock-to-flow has been on the record for years, and 2026 keeps confirming it. The statistician writing as Level39 showed in Bitcoin Magazine that the model’s tight fit is mostly autocorrelation: market value is stock times price, so regressing it against stock over flow comes close to testing stock against itself. Swan Bitcoin chief executive Cory Klippsten put the same point bluntly, calling the model “autocorrelated” and arguing it “does not need S2F at all. Just price vs time.” Strip the artifact out, the critics say, and the R-squared collapses toward zero.
There is a second statistical problem beneath the first. Both sides of the regression trend upward over time, which can manufacture a strong fit between series that have no real causal link, a trap known as spurious regression. And the model rests on a sample of just three completed halving epochs, four counting the current one, which is far too thin to establish a durable law. A relationship fit to a handful of data points, using variables that both rise with time, will look convincing on a log chart and tell you almost nothing about tomorrow.
Nico Cordeiro of Strix Leviathan warned in CoinDesk back in 2020 that stock-to-flow was a “chameleon” model, refitted whenever reality diverged, and about as reliable as the astrological models of the past. Ethereum co-founder Vitalik Buterin was harsher still after the 2022 crash, writing on The Block that financial models offering “a false sense of certainty and predestination that number-will-go-up are harmful and deserve all the mockery they get.” The production-cost view, whatever its limits, makes no such promise. It tells you where the pain is, not where the jackpot is, and that modesty is a feature.
How to Read a Bitcoin Price Floor
For a reader trying to separate a useful floor from a comforting story, the test is whether the floor has a mechanism you can check. A model that only ever points up, with no date and no feedback loop, is a narrative. A floor tied to production cost comes with instruments you can watch: the difficulty adjustment every two weeks, hashprice, the all-in cost disclosures in miners’ quarterly filings, and whether miner treasuries are growing or being sold off to pay the power bill.
A few practical signals follow from the production-cost frame:
- When price is below all-in cost but above the electrical floor, expect shutdowns and a difficulty cut, not an instant bottom.
- A falling difficulty is not bearish on its own; it lowers the cost of production and relieves the miners still running.
- Record miner selling signals stress, not scarcity; the float can swell even as fresh issuance shrinks.
- No model, scarcity or cost, hands you a date. Treat any target that quotes a price but no deadline as a slogan.
None of this makes production cost a crystal ball. It makes it accountable. You can be wrong about it in a way you can measure, test, and update, which is more than stock-to-flow has ever offered. The right lesson is not to swap one magic number for another, but to prefer the model that shows its work.
What It Means for the Rest of 2026
Heading into year-end, Bitcoin is on track for a second consecutive annual loss, something it had never done before this cycle, and it is doing so in a market that long ago stopped trading on halving math. The production-cost frame suggests the near-term path is a grind, not a scarcity-driven moonshot: with price under cost, weaker miners keep capitulating, difficulty keeps easing, and the cost of production drifts down to meet the price rather than the price leaping up to meet $500,000.
That is a less thrilling story than the one stock-to-flow tells, and a more honest one. It also carries practical weight for anyone managing a position into a down tape; our year-end tax playbook for a down market covers the loss-harvesting and record-keeping side. The lesson of 2026 is not that Bitcoin is doomed. It is that the forces setting its price are demand, liquidity, and the brute economics of mining, and that a model built on scarcity alone was always going to be a passenger, never the driver.
Frequently Asked Questions
Is the stock-to-flow model wrong?
By its own test, yes. The 2024 halving tightened supply exactly as the model assumed, yet Bitcoin fell about 49% from its record rather than climbing toward the model’s $500,000 cycle target. A model whose key input fired perfectly while its output missed by a wide margin is not early; it is mis-specified, because price is set by demand and cost, not scarcity alone.
What is Bitcoin’s production cost in 2026?
Estimates cluster in the high $70,000s to high $80,000s. JPMorgan puts listed miners near $78,000 a coin and private operators near $88,000, CoinShares estimated a weighted cash cost around $75,500, and one independent model placed the figure near $88,800 in mid-November. With Bitcoin near $63,800 on 19 November 2026, the asset traded below every one of those estimates.
Why does mining difficulty fall when the price drops?
Difficulty retargets about every two weeks to keep blocks near ten minutes apart. When price falls below the cost of production, higher-cost miners switch off, blocks slow, and the next adjustment lowers difficulty. That reduces the cost of mining each coin for everyone still running, which is why the 13 November 2026 cut of roughly 7%, the steepest since June, eased pressure on surviving miners.
Does PlanB still predict $500,000?
Yes. PlanB continues to reiterate a $500,000 average for the 2024 to 2028 cycle, with a $250,000 to $1 million range, and argues the October 2025 high was not the top. Notably, he has said mining difficulty reflects market price better than stock-to-flow or power-law models, pointing to mining economics rather than his own scarcity ratio for day-to-day signals.
Is production cost a reliable price floor for Bitcoin?
It is a floor with a mechanism rather than a guarantee. Total production cost can be pierced during capitulation, as it was in November 2026, but the lower electrical-cost floor has rarely broken, and the difficulty adjustment pulls the cost of production down toward a falling price. That feedback loop, not an unfalsifiable $500,000 target, is what has historically shaped Bitcoin’s downside.
By Priya Reddy, predictions desk, HOGE Wire.