Why the Security Model Cannot Stabilize and Must Collapse to Functional Zero
An incentive-based analysis of the post-subsidy security cliff, the compounding fragility trap, and the absence of a stable low-equilibrium state.
Note: This article is attracting readership well outside the usual BiggerPockets Money audience. If you understand Bitcoin’s security model and mining economics, and take the core mechanism here seriously, I’d welcome a call. Reach me at Scott@biggerpocketsmoney.com. A longer 42-page version is available as a free PDF in our resource library – no email required.
Prologue
I spent several years trying to understand something I initially dismissed: Bitcoin. Advocates kept telling me to do more homework, so I did. I reread the whitepaper, studied the protocol rules, tracked mining economics and fee trends, and modeled long-run security incentives. I expected to find what I was missing.
What I found instead was a structural vulnerability that gets more binding the larger and more successful the system becomes.
This is an incentive and capital-structure analysis. The conclusion follows from the math of subsidy halving, the economics of stranded mining capital, and the game theory of defection when honest mining loses money. Those rules force a binary long-term outcome: Bitcoin either compounds at a real rate no major asset class has ever sustained, or it decays to negligible economic relevance and security.
I’m publishing this for the record. If Bitcoin defies these dynamics and achieves durable, self-sustaining monetary relevance, I’ll be the first to say I was wrong. The argument needs to be falsifiable and mechanically sound, popular or not.
1. The Core Claim: There Is No Middle Path
Bitcoin’s long-term fate is binary. Either it succeeds as a durable monetary system – meaningful purchasing power, network security, and economic relevance for decades – or it collapses to functional zero: a chain with negligible hashrate, no transaction activity, no reliable price discovery, and no realistic path back.
The popular intuition that Bitcoin can muddle along indefinitely as a niche digital collectible is comforting and wrong. The protocol’s incentive structure doesn’t permit a stable low equilibrium as block subsidies decline. Once honest mining revenue falls below a critical threshold for long enough, the system enters a self-reinforcing defection dynamic that runs to collapse rather than settling into dormancy.
“Winning” means lasting monetary relevance: a security budget sufficient to deter rational attacks for decades, reliable block production, sustained economic activity, and purchasing power that compounds faster than any other large asset class for generations. Something like the “digital gold” vision.
“Losing” means hashrate falling below 1-5% of peak and never recovering, block intervals stretching to hours or days, fees near zero, exchanges freezing or delisting, and price discovery disappearing into sporadic illiquid trades. Bitcoin wouldn’t literally hit $0. It would become a ghost chain – alive in name only.
The middle path fails for three reasons, each developed below:
- Exponential subsidy decay. The block reward halves every four years, forever. After 2032 and 2036, the subsidy becomes trivial unless price compounds at roughly 17-22% real CAGR – a rate no large asset has ever sustained.
- Hyper-specialized sunk capital. Nearly all hashrate comes from ASICs, remote data centers, and power contracts with almost no alternative use. Every bull cycle commits more of this stranded capital, making the next downturn more violent.
- Persistent defection incentives in the tail. Once revenue gets low enough, even the cheapest remaining miners face a choice between mining for pennies and defecting once for a large, immediate payoff. That gradient is structural, so any low-hashrate equilibrium is unstable.
There’s a perverse corollary: temporary success makes permanent failure more likely. Every price run-up deploys more non-redeployable capital, raising both the probability and severity of the eventual defection cascade. Bull markets make it worse.
2. The Security Budget Cliff
Miners must earn enough real revenue to keep majority hashrate honest. That revenue comes from two sources: the block subsidy and transaction fees. The subsidy halves every ~210,000 blocks until it reaches zero in the 22nd century – exponential decay in the dominant component of the security budget.
Fees are real but limited. Since 2021 they’ve occasionally hit 20-40% of miner income during congestion spikes (Ordinals in 2023, Runes in 2024), then collapsed back to 5-15% for long stretches. No multi-year period has ever sustained fees above ~10-15% of total block reward value. Block space is fixed, users are price-sensitive above $10-20 per transaction, and Layer-2 networks keep pulling activity off-chain. Fees might matter in the distant future. Over the next several halving cycles, even bulls should be cautious about leaning on them.
Where things stand in early 2026. Bitcoin trades around $75,000. The subsidy is 3.125 BTC per block, ~450 BTC per day, worth ~$33.75 million. Fees add $1-4 million. Total daily miner revenue: $35-38 million. Global all-in mining cost (power, operations, depreciation, debt service) is an estimated $40-55 million per day, based on public miner filings, Luxor Hashrate Index data, and CoinShares reports.
The network is already underwater on average. Many miners are cash-flow negative. Some are covenant-tripped. This isn’t a hypothetical future problem.
The required price path. The table below shows the BTC price needed at each halving to hold today’s real security budget (~$45 million/day in 2026 dollars), under zero fees and under a generous assumption that fees cover 40% of revenue – far above any sustained historical level.
| Halving | Daily Subsidy (BTC) | Required Price (Zero Fees) | Required Price (40% Fees) | Multiple from $75k (40% Fees) |
| 2028 | 225 | $200k – $220k | $120k – $132k | 1.6× – 1.8× |
| 2032 | 112.5 | $400k – $440k | $240k – $264k | 3.2× – 3.5× |
| 2036 | 56.25 | $800k – $880k | $480k – $528k | 6.4× – 7.0× |
| 2040 | 28.125 | $1.6M – $1.76M | $960k – $1.06M | 12.8× – 14.1× |
| 2048 | 7.03 | $6.4M – $7.0M | $3.84M – $4.2M | 51× – 56× |
| 2056 | 1.76 | $25.6M – $28.0M | $15.4M – $16.8M | 205× – 224× |
| 2064 | 0.44 | $102M – $112M | $61.2M – $67.2M | 816× – 896× |
Even with fees at 40% of revenue forever, Bitcoin needs $61-67 million per coin by 2064 just to hold today’s real security budget. That’s roughly 17-17.5% real CAGR for 38 years (19.5-20% nominal). With zero fees, 18.5-19% real.
A bet on 17% real returns for four decades is absurd. Gold has done 1-2% real over centuries. Global equities, 6-7% real long-term. Top venture funds rarely exceed 15% real over 40 years at any meaningful size. No asset that reached multi-trillion-dollar scale has ever then compounded at 17%+ real for decades. Bitcoin would need venture-capital returns at reserve-asset scale, indefinitely, with its security mechanism entirely dependent on hitting exactly that performance.
Why use today’s budget as the baseline? Because nobody has demonstrated a lower bound at which proof-of-work security self-stabilizes under persistent negative revenue. Security is adversarial: what matters is whether attack cost reliably exceeds attack payoff for all rational actors. At a sub-$5 million/day budget, modestly capitalized actors can dominate block production while the honest majority has no surplus to respond with. Difficulty adjustment preserves block timing, not deterrence. Until someone rigorously defends a viable lower threshold, the current budget is the only security regime that has ever been shown to work.
3. The Infrastructure Ratchet: Bull Cycles Make Collapse Worse
The most counterintuitive dynamic in Bitcoin’s design: every meaningful price increase compounds the eventual fragility.
The mechanism runs in four steps. Price rises and miner revenue spikes. High margins trigger massive capex – public miners issue equity and convertible debt, operators sign long-term power deals, new ASIC generations get ordered in bulk, 100-500 MW facilities get built in remote locations chosen for cheap energy. That capital is hyper-specialized: ASICs do SHA-256 and nothing else, stranded-location data centers don’t convert cleanly to cloud or AI workloads, and power contracts revert to the utility when mining stops. Then, when price fails to keep pace with the required path, the network’s breakeven is higher than before, the pool of capital facing negative NPV is larger, and the incentive to defect is stronger.
The pattern has already run three times. The 2017 bull took hashrate from ~5 to ~60 EH/s; the crash to $3,200 cut it roughly in half and made S9-era hardware nearly worthless. The 2021 bull took hashrate to ~250-300 EH/s on billions in new debt and equity; the 2022 crash cut it 40-50%, but recovery was possible because the subsidy was still 6.25 BTC. The 2024-2025 bull pushed hashrate to ~900-1,000 EH/s. At $75,000, the network is already underwater, and the sunk capital at risk is an order of magnitude larger than prior cycles.
| Cycle peak | Peak hashrate (EH/s) | Est. global mining capex | Avg all-in breakeven at peak |
| 2017 | ~60 | $2–4B | $5,000–$8,000 |
| 2021 | ~250–300 | $15–25B | $20,000–$30,000 |
| 2025 | ~900–1,000 | $40–80B | $60,000–$90,000 |
The common rebuttal is that miners will just shut down and wait for the next cycle. Three realities block that off-ramp. Most of the capital is debt-financed or equity-diluted, and public companies hit covenants and shareholder pressure long before they can quietly mothball. Hardware loses most of its resale value within a year or two of unprofitability. And power contracts revert or turn punitive when mining stops. The infrastructure can’t be turned off and stored at scale without enormous losses, and losses create desperation.
This is also why the security budget has to be modeled dynamically, not as a smooth average. Prior analyses, including Budish (2018, updated 2024), evaluate incentives at a steady state. But Bitcoin’s price path is volatile – multi-fold booms followed by 70-90% drawdowns – and security depends on staying above a price floor that ratchets upward with each cycle’s capex. The real tests come at moments of maximum volatility, which are also the moments when fee revenue is most depressed.
The strongest objection is empirical: Bitcoin survived an 84% drawdown in 2018 and a 77% drawdown in 2022, and re-equilibrated both times. True, and the resilience was real. But each recovery was achieved against a lower floor. The 2018 recovery needed miners viable at a $3,000-$8,000 breakeven. The 2022 recovery, $15,000-$25,000. The current cycle requires defending a breakeven around $38,000-$65,000 – a test 5-8x harder in nominal terms. A system that re-equilibrated at a $5,000 floor has not demonstrated it can do so at $40,000. Every system that eventually fails showed resilience in prior cycles, by definition. The three-cycle capex progression above is the evidence that this test is qualitatively different.
4. The Low State Is Not Stable
Even skeptics tend to believe in a minimal equilibrium: hashrate at a few percent of peak, slow blocks, near-zero fees, a small group of ultra-low-cost or ideological miners keeping the chain alive. The protocol’s incentives rule this out. Once honest revenue falls below a critical threshold for a sustained period, defection becomes the dominant strategy for every rational remaining participant, including the cheapest ones.
The low-cost tail. The usual counterargument is that miners with $0.01-0.02/kWh power can mine profitably at almost any revenue level, creating a floor. Optimists put this tail at 30-50% of hashrate, implying a cascade floor as low as $1,000-$10,000. Neutral analysts see 10-30%, implying $10,000-$30,000. My view: even the cheapest reported rates rarely reflect all-in costs (depreciation, maintenance, labor, opportunity cost), and the tail shrinks fast in low-revenue regimes because miners stop when expected future revenue stops justifying the operational grind of harvesting stranded energy. That implies a cascade floor around $30,000-$75,000 (the higher end only plausible in a truly steep cascade – like 30-50% in a single 2016 block period).
A bigger tail does lower the price at which trouble starts, but no version of the tail is stable. Run the numbers on a post-2036 low state: ~56 BTC daily subsidy at $50,000-$100,000 means $3-6 million per day in total network revenue. Spread across even a collapsed 10-50 EH/s, that’s pennies to a few dollars per EH/s per day – not enough to cover depreciation, maintenance, security, and compliance even with free power. Meanwhile the energy itself faces rising competition: AI training and HPC tenants will pay multiples more for the same flared gas and excess hydro, with predictable revenue and less regulatory baggage. The “immortal” tail erodes from both ends.
The attack side of the ledger. While honest NPV trends toward zero, the payoff from a single well-timed defection stays large: short positions ahead of a reorg, double-spends against remaining exchange hot wallets, deep out-of-the-money puts on miners and Bitcoin ETFs, censorship ransom, or a pre-mined hidden chain. Notably, these derivative markets did not exist at any meaningful scale in prior periods with meaningful hashrate capitulation. Crisis execution is messy – liquidity evaporates, exchanges halt, counterparties break – and the derivative payoff is uncertain and nearly impossible to preduct. But it only needs to be large enough in expectation to tip an actor who is already near-indifferent between honest mining and defection. A $500M notional put position taking a 50% execution haircut still dwarfs the cumulative honest-mining trickle available to the same actor over any realistic horizon. For anyone controlling 10-20 EH/s of a collapsed tail, defection dominates.
Who actually holds the tail. Hashrate today is concentrated: Foundry USA at ~30-35%, AntPool at ~18-22%, then SpiderPool, ViaBTC, F2Pool, and smaller pools. By country, the US holds ~37-38%, Russia ~16% (~175 EH/s), underground China ~12%, with smaller shares in Paraguay, UAE, and Ethiopia. US-regulated public miners have effectively zero attack probability – legal exposure, fiduciary duty, and large disclosed BTC treasuries align them with preservation. The two realistic attackers in a deep capitulation are Russian operators (cheap surplus energy, state-tolerant framework, sanctioned entities with strategic motives beyond economics) and Iranian state-linked mining (subsidized, sanctions-evasion driven, opaque reserves).
The attack capability requires no mobilization. It emerges passively. At 900 EH/s, Russia’s ~175 EH/s is about 19% – nowhere near majority. If a price decline triggers 60% capitulation among economic miners, the network falls to ~360 EH/s and Russia’s unchanged hashrate becomes ~49% of what remains. At a 95% capitulation, it’s 80%+. Russia and Iran combined clear the 51% threshold somewhere below ~$40,000. Russia alone clears it near $30,000. No new machines, no coordination – just other people turning theirs off.
One scenario is worse still. An operator who shuts down but keeps the hardware, the facility, and the power connection hasn’t destroyed their attack capability. They’ve put it in cold storage. If the network keeps shrinking around them until their idle fleet exceeds 51% of remaining hashrate, reactivation is a unilateral decision: turn the machines back on. Stranded, idle hashrate, not remaining miners, is, in my view, the greatest risk to Bitcoin’s security in a hashrate capitulation scenario.
Two attackers is plenty for the game theory. The cascade requires only one rational defector with temporary majority control. Russia and Iran lack enforceable coordination, which cuts both ways: mutual deterrence is fragile, and aligned interests could just as easily produce joint extraction.
To be precise about what this argument claims: game theory does not guarantee an attack. Real-world frictions – coordination uncertainty, residual strategic value of a sanctions-evasion channel, diplomatic blowback, bureaucratic inertia, unknown private BTC holdings – can delay or prevent execution even after the incentives flip. What the analysis shows is that below a certain price, the economic mechanism that has protected Bitcoin since 2009 stops operating, and security comes to depend entirely on factors the protocol does not control. That margin of safety is vanishingly thin, and it thins further the longer capitulation persists, as price falls further, and if prices fall faster.
Investors don’t wait for the attack. Family offices, hedge funds, and institutions exit when the tail risk becomes credible, not when it executes. Once hashrate capitulation is visible on-chain and Russia-plus-Iran’s share of the remnant crosses majority, the perception shift alone triggers preemptive de-risking: anticipation depresses price, lower price accelerates capitulation, visible capitulation validates the thesis, more capital exits. Bitcoin can suffer a terminal repricing driven purely by perceived fragility.
Volatility moves the thresholds up. The figures above assume gradual declines that the two-week difficulty adjustment can track. A violent 15-25% drop changes the math: 30-60% of hashrate can vanish in days while difficulty barely moves, spiking the surviving sovereign share immediately. In a flash crash, the “tail risk is live” configuration becomes visible at $50,000-$65,000 – well above the steady-state floors. The difficulty algorithm keeps blocks coming; it does nothing to restore deterrence or reverse the composition shift toward sovereign operators.
5. Tipping Points and Cascade Dynamics
The collapse unfolds as a cascade with identifiable thresholds:
| Indicator | Trigger level | Why it matters |
| Daily total miner revenue | Sustained below $1–3M/day | Even ultra-low-cost miners can’t cover minimal opex; honest NPV clearly negative |
| Global hashrate | Below 5–10% of peak for 3–6+ months | Remaining tail too small to resist coordinated defection |
| Average block interval | Sustained above 1–2 hours | User experience collapses; activity and fees evaporate |
| Fee revenue share | Stuck below 10–15% of revenue | No fee takeover; subsidy decay dominates |
| Exchange behavior | Major platforms freeze withdrawals or delist for weeks | Liquidity and price discovery die |
When three or more of these are crossed simultaneously for several months, collapse moves from likely to near-certain.
The sequence: revenue falls below breakeven and the highest-cost miners capitulate first, dropping hashrate 30-60% while difficulty lags and blocks slow. Slower blocks discourage usage, fees fall further, and the next layer of miners exits. The remaining sub-10% tail earns dollars or pennies per EH/s while attack NPV dominates. The first defector mines a hidden chain or executes a targeted double-spend; exchanges detect conflicting chains and freeze; price crashes; more miners exit; panic spreads. Eventually blocks arrive days apart and the ledger becomes a museum piece – intact but worthless.
Recovery becomes impossible early in this sequence. Liquidity dies, trust dies, the social layer fractures under stress, and once hardware is scrapped, restoring meaningful hashrate takes years and billions with no profit case. A tail-emission rescue fork during maximum stress – frozen exchanges, crashing price, divided community – is historically a fantasy; every contentious fork under pressure (ETC, BCH, BSV) split the chain rather than saving it.
The best counterargument is hardening: after a first credible majority event, exchanges raise confirmation requirements, institutions hedge, developers respond, and the window closes. The mechanism is real. The problem is timing and resources. Hardening requires capital deployment into a collapsing asset by entities being bankrupted by the same collapse – a miner with $500M of debt, tripped covenants, and hardware worth a fraction of book cannot raise money to restore hashrate. And it requires social coordination at the exact moment the community is most divided and its largest operators are filing for bankruptcy. Hardening could work if the first incident is contained while economic resources remain. Once the thresholds above are crossed together, those resources are gone precisely when they’re needed.
6. Implications: The $40,000 Ignition Zone
The binary endgame is a long-term structural claim, but specific price zones can flip the system from “delayed reckoning” to “visible tail-risk regime” in this cycle (after the next halving in 2028, or if Bitcoin hits a new all time high and more CapEx enters, or if hashrate is permanently removed from circulation or upgraded with more efficient hardware, the ignition zone will move to a new range). Around $40,000, three things converge.
First, the tail risk becomes observable. Hashrate falls 40-70% within weeks, difficulty lag concentrates the survivors, and the Russian-plus-Iranian share of active hashrate crosses majority somewhere in the $35,000-$45,000 range. At that point sophisticated capital doesn’t need an attack to happen. It concludes the deterrence mechanism no longer operates and exits.
Second, the dominant marginal buyer disappears. Since late 2023, MicroStrategy has absorbed an estimated 20-30% of aggregate fiat inflows into BTC through equity and convertible issuance. Below $40,000-$45,000, equity raises become impossibly dilutive, convertible arbitrage loses appeal, covenants approach breach, and the “Bitcoin yield” story breaks. No forced selling required – the mere absence of the most committed buyer is mechanically equivalent to a large increase in selling pressure.
Third, institutional capital has no ideological floor. Allocators are marked-to-market and bound by drawdown limits. When the headlines read “hashrate collapses 60%” and “Russian operations dominate remaining proof-of-work,” Bitcoin gets reclassified from macro hedge to tail-risk asset, and the de-risking is self-reinforcing.
Two cycle-specific amplifiers make this worse than prior bears. AI and HPC demand is bidding away the stranded energy that the “immortal tail” depends on, shrinking it faster than the optimist models assume. And equity markets trading at 22-26x forward earnings, with AI infrastructure spending flattering reported margins, leave little risk appetite to backstop miners or corporate treasuries if multiples normalize.
Put together: the realistic ignition zone for this cycle is $38,000-$48,000, with flash-crash dynamics capable of making the tail risk visible as high as $50,000-$55,000. A floor of $25,000-$35,000 is possible only if the low-cost tail is much larger than current estimates and stays committed despite years of negative honest-mining NPV. If the core argument here is right, even sophisticated optimists should concede the mechanism and confine the debate to where the floor sits.
Epilogue: The Cost That Never Shows Up on the Ledger
If Bitcoin collapses the way its incentives imply, the erased market cap will dominate headlines for a few weeks and then fade. The deeper loss is the opportunity cost. For over a decade, Bitcoin has absorbed an extraordinary share of engineering talent, intellectual energy, political capital, and risk-taking – much of it sincere and idealistic, none of it redeemed by sincerity. Failed experiments are cheap and often productive. The tragedy is that Bitcoin succeeded just enough, for just long enough, to trap enormous human and financial capital inside a system whose rules made durable success less likely as it grew. Every cycle pulled more people in, delayed the reckoning, and raised its cost. Belief was asked to do the work of economics.Full disclosure: After years of studying Bitcoin’s incentive structure and concluding it faces structural long-term risks, I maintain zero long exposure to Bitcoin and hold a small, conditional bearish position in my personal portfolio, sized modestly and expressed only through regulated instruments. This analysis reflects my independent views. It is not financial advice, an investment recommendation, or a solicitation. Do your own due diligence.
Leave a Reply