Sylvain, 58, lives in Geneva. Since 2018, he puts a few hundred francs into Bitcoin each month via Relai. Seven years later, he holds a bit over 4 BTC and is preparing his retirement at 65.
For the first time since his first purchase, he thinks about selling a part. Not to panic, not because the thesis has changed. But because at some point, he will have to convert part of this wealth into liquidity to live. And Sylvain realises he has no plan. He spent years learning to buy and hold, zero time on the question: how do I exit.
This question is the blind spot of Bitcoiners. The dominant mantra is « only buy, never sell ». Consistent with an infinite store-of-value thesis. Incompatible with using one's wealth in a human life.
This article describes what the literature and practice say on the subject in May 2026: the 3 exit horizons, the « sell some vs hold all » debate, a central table comparing 4 operational strategies, tax implications by jurisdiction, and the Bitcoin-as-collateral path. It is not a recommendation, it is a toolbox.
Why talk about exit when Bitcoin is designed to be held
The « never sell » mantra has a real foundation. Bitcoin is a fixed-supply monetary asset (21 million21 millionMaximum number of bitcoins that will ever exist, hard-coded in the protocol. This programmed scarcity is a founding feature. The last sat will be mined around the year 2140.See in the lexicon → units), whose main thesis is the preservation of purchasing power over decades. Selling Bitcoin against fiatFiat (fiat currency)State currency with legal tender status (euro, Swiss franc, dollar), issued by a central bank and not backed by a physical asset. By contrast, Bitcoin has an issuance capped at 21 million units, with no central issuer.See in the lexicon → currency that structurally depreciates 2 to 4 % per year is, over the very long term, a net-negative operation. Saifedean Ammous puts it bluntly in The Bitcoin Standard (2018): Bitcoin is designed to save, not to spend. Michael Saylor repeats that he buys Bitcoin to pass it on, not to sell it.
This stance is coherent for an infinite horizon. It is partially defective for the real life of an individual. A human has a limited life expectancy. He also has occasional financial needs that do not wait for retirement or inheritance: home purchase, financing studies, starting a business, health expenses, divorce, etc. At all these moments, the investor needs liquidity. If he does not have it, he sells what he has at hand. And what he has at hand is often Bitcoin, because it is the most liquid asset of his estate.
The operational problem is therefore not to sell or not to sell. The problem is to sell under what plan. An exit planned in advance, calibrated, spread over time, under written rules, produces radically different results from an improvised exit under pressure. The literature on the subject is limited (behavioural finance has mostly looked at entry biases, little at exit biases), but 3 lessons recur.
Lesson 1: the improvised exit happens at the worst moment. When a BitcoinerBitcoinerPerson interested in Bitcoin, who holds some and adheres more or less to its values (individual sovereignty, sound money, decentralisation).See in the lexicon → without a plan ends up selling, it is typically because an unforeseen expense arrives in bear (paradoxically: it is when the market is down that humans are also most likely to have financial worries, by economic-cycle effect). Statistically, unplanned sales fall in bear more often than in bull. It is the opposite of what should happen.
Lesson 2: the total exit at the peak is a fantasy. Nobody knows how to identify the peak in real time. Technical analyses (Pi Cycle Top, MVRV ratio, monthly RSI) give a posteriori signals. Those who try to « sell everything at the top » miss the top in half the cases and stay frozen in bear. Exiting in multiple steps is mechanically superior to trying to time the peak.
Lesson 3: the exit is not a single decision, it is a program. Like the entry (cf. DCA), the exit is spread and mechanised. The rules are written before the emotion arrives. It is the exact opposite of « I will see when the time comes ». When the time comes, the emotion will be there, and the decision will be bad.
These three lessons reverse the debate. « Never sell » is not an exit strategy, it is the absence of strategy. And absence of strategy is the most costly default of experienced Bitcoiners.
The 3 exit horizons: tactical, scheduled, total
Before comparing operational methods, we must clarify what « exit » means. Three horizons exist, with different objectives and different tools.
Horizon 1: partial tactical exit. Sell 10 to 30 % of the position at cycle peak to secure gains and rebalance the estate. The Bitcoin position remains majority (70-90 %), the investor keeps his asymmetric exposure to the next cycle. This is what Pierre does in the lump-sum vs DCA article. Typical tools: written rule triggered by price threshold (« if BTC exceeds 2x previous ATHATH (All-Time High)Highest historical price of an asset. Breaking the previous ATH is a strong psychological signal in Bitcoin cycles.See in the lexicon →, sell 20 % »), or by technical signal (MVRV > 3, Pi Cycle Top activated, monthly RSI > 80).
Horizon 1 is compatible with a long-term thesis. It accepts that nobody can time the peak and that a partial exit in the peak zone beats a failed total exit. Statistically, selling 20-30 % between +50 % and +20 % before the peak captures most of the achievable gain. It is suboptimal in pure expectation (Bitcoin's right tail remains massive), but it is psychologically tolerable and tax-spreadable.
Horizon 2: exit scheduled for defined use. Progressively convert a significant portion (30 to 70 %) of the position into fiatFiat (fiat currency)State currency with legal tender status (euro, Swiss franc, dollar), issued by a central bank and not backed by a physical asset. By contrast, Bitcoin has an issuance capped at 21 million units, with no central issuer.See in the lexicon → over 5 to 10 years, to finance an identified use: retirement, home purchase, early inheritance. Typical tools: scheduled DCADCA (Dollar Cost Averaging)Buying a small fixed amount at regular intervals (for example 100 EUR a week), regardless of price. Smooths the average purchase cost and neutralises timing bias.See in the lexicon →-out (sell X CHF per month over 5 years), or sales calibrated on price tiers (5 tiers at 20 % each, spaced across successive cycle peaks).
Horizon 2 recognises that Bitcoin must serve something in a human life. It accepts the opportunity cost (selling early misses the future right tail) in exchangeExchangeService that lets you buy, sell and swap cryptocurrencies against fiat money. Examples : Kraken, Coinbase, Bitstamp, Bitvavo. Most are custodial.See in the lexicon → for the security of liquidation. Sylvain is typically in this case for his retirement: he needs to convert ~50 % of his Bitcoin position over 2027-2034 to top up his pension fund.
Horizon 3: total exit. Liquidate 90-100 % of the position because the fundamental thesis has changed: major regulatory change, successful network attack, monetary alternative that makes Bitcoin obsolete (improbable but possible over 30 years), radical personal evolution of the risk profile. Typical tools: total sale immediately after confirmation of thesis change, ideally spread over 2-3 months to smooth tax impact and market slippageSlippageDifference between the expected price and the actual execution price. More pronounced on large orders or thinly traded markets.See in the lexicon →.
Horizon 3 is rare and must be justified by an objective event, not by an opinion or a price move. Selling everything by capitulationCapitulationFinal phase of a bear market where the last sellers give in to panic, often on record volume. Frequently marks the cycle bottom.See in the lexicon → in bear (Thomas case in the volatility psychology article) is not a horizon 3, it is a plan failure.
Sylvain identifies his need as majority horizon 2 (scheduled exit for retirement, 50 % of position over 7 years) plus secondary horizon 1 (tactical exit 10-15 % at the next cycle peak 2028-2029, to secure a portion regardless of retirement calendar). Horizon 3 remains an option if the thesis changes, but Sylvain anticipates nothing on that side.
The « sell some, hold most » rule and its psychological foundation
Once the idea that an exit strategy is necessary is accepted, the question becomes: how much to sell. The rule of thumb that recurs most in the practice of experienced Bitcoiners is: sell a minority fraction (15 to 30 %) at the cycle peak, keep the majority (70 to 85 %) very long term. This is informally called « sell some, hold most ».
Mathematically, in pure expectation, this rule is dominated by « hold everything ». Bitcoin has a fat right-tail returns distribution (cf. halving and cycles): over the 16 years of history, holding 100 % from any date produces on average more than selling a fraction at the peak. The reason: successive peaks have historically exceeded previous peaks by 2 to 5x. Selling at the 2017 peak at 17,000 EUR to buy back in 2022 at 15,000 EUR earns fiatFiat (fiat currency)State currency with legal tender status (euro, Swiss franc, dollar), issued by a central bank and not backed by a physical asset. By contrast, Bitcoin has an issuance capped at 21 million units, with no central issuer.See in the lexicon → short term, but on 2025 (110,000 EUR), the 100 % hold does better.
Yet in practice, the retail investor who applies « sell some, hold most » holds the long-term position better than the one who promises himself « never sell ». Three psychological mechanisms explain this paradox.
Mechanism 1: realising a partial gain reduces loss aversion on the rest. Once Sylvain has sold 15 % of his position at the 2028 peak and cashed in, say, 65,000 CHF in fiat (beyond his initial investment), psychologically he considers that « the entry cost is recovered ». The rest of the position (85 %, i.e. 3.57 BTC) can now fall by -80 % in the 2030 bear without triggering panic, because it is « the house » falling, not the initial capital.
Mechanism 2: rebalancingRebalancingRebalancing your portfolio by selling part of what has risen and buying what has fallen, to return to a target allocation.See in the lexicon → prevents over-concentration. A BitcoinerBitcoinerPerson interested in Bitcoin, who holds some and adheres more or less to its values (individual sovereignty, sound money, decentralisation).See in the lexicon → who never sells sees his Bitcoin/estate ratio mechanically rise with each bull. Sylvain started in 2018 at 5 % of his estate in Bitcoin, it is now 14 %, and at the next peak it will be 25-30 % without him doing anything. At 30 %, concentration becomes uncomfortable for his profile. Selling 15-20 % at the peak brings the ratio back into the target range (10-15 %) and lets him sleep.
Mechanism 3: creating a positive sale precedent makes future sales easier. The investor who has never sold a single satoshiSatoshi (sat)The smallest unit of bitcoin. 1 BTC = 100 million satoshis. Named after the creator. In 2026, talking in sats becomes common as the price of one BTC rises.See in the lexicon → in his life is psychologically blocked the day he must sell. The first sale act is traumatic. The investor who has already sold 15 % in 2028 will find it easier to sell 20 % in 2032 then 25 % in 2036 to finance his retirement. The gesture is rehearsed.
The 3 mechanisms together produce a counter-intuitive effect: selling a minority fraction at the cycle peak allows holding the majority better very long term. The statistical opportunity cost is compensated by behavioural robustness. The Sylvain case materialises it: his plan includes a 15 % tactical sale at the 2028-2029 peak for this precise psychological reason, not to optimise expected return.
The argument works less for psychologically disciplined investors (rare) or for those whose Bitcoin exposure remains very minority (< 5 % of estate), where over-concentration risk does not exist. For Sylvain however, it fully applies.
Central table: 4 exit strategies compared
Operational synthesis of the 4 most used exit strategies. None is universally superior. The choice depends on profile, horizon, psychological tolerance and tax jurisdiction.
| Strategy | Advantages | Disadvantages | Target profile |
|---|---|---|---|
| Total exit at peak Sell 90-100 % at once when the market looks overheated |
Theoretical capture of the maximum if timing is right. Fast liquidation, compact tax case. | Nobody can time the peak in real time. Very high risk of selling 3-12 months too early or too late. Massive tax impact on a single year. Psychologically hard to execute (paralysis at trigger moment). | Rare. Investors with very strong conviction they can time (often wrongly), or investors forced by a single massive expense (cash home purchase for example). |
| Progressive exit by price tiers 5 tiers at 20 % each, triggered at predefined price levels |
Independence from calendar. Captures part of the rises without trying to guess the peak. Pre-committed written rules that resist emotion. | If the peak exceeds planned tiers, the sale happens too early. If the peak does not reach the tiers, no sale happens and exit is pushed back one cycle. Possible tax concentration if several tiers trigger the same year. | Investors convinced by halvingHalvingScheduled event every 210,000 blocks (roughly every 4 years) that cuts the miner reward in half. This mechanism makes Bitcoin issuance decline towards the total cap of 21 million.See in the lexicon → cycles, who accept missing part of the right tail in exchangeExchangeService that lets you buy, sell and swap cryptocurrencies against fiat money. Examples : Kraken, Coinbase, Bitstamp, Bitvavo. Most are custodial.See in the lexicon → for objective discipline. Suited to tactical horizon 1. |
| Reverse DCADCA (Dollar Cost Averaging)Buying a small fixed amount at regular intervals (for example 100 EUR a week), regardless of price. Smooths the average purchase cost and neutralises timing bias.See in the lexicon → (DCA-out) Sell X CHF (or X % of residual position) each month over 24-48 months |
Complete independence from timing. Optimal tax spreading over 2 to 4 years. Reproducible and mechanical like entry DCA. Suits identified objectives perfectly (retirement, long-term financing). | High opportunity cost if the market is rising during the exit (BTC sold early are worth less than those sold late). Requires execution discipline over 24-48 months without deviating. | Investors with identified fiatFiat (fiat currency)State currency with legal tender status (euro, Swiss franc, dollar), issued by a central bank and not backed by a physical asset. By contrast, Bitcoin has an issuance capped at 21 million units, with no central issuer.See in the lexicon → need and established calendar (retirement at 5-7 years, programmed home purchase). Suited to scheduled horizon 2. This is Sylvain's main strategy. |
| Exit triggered by technical rules Sell 5-10 % per month when MVRV > 3, Pi Cycle Top activated, monthly RSI > 80 |
Adapts to actual market conditions rather than a fixed calendar. Captures the real peak better than price-tier exit. Historical signals have had a decent success rate (but not perfect). | Technical signals work until the day they no longer work. No guarantee they hold over the next cycles, especially with post-ETF institutionalisation which may dampen cycles. Requires regular indicator monitoring. | Investors comfortable with on-chain analysis and technically active (5-10 % of the time). Can complement DCA-out as an accelerator at identified peak. |
Practical combination. In reality, few investors use a single pure strategy. Sylvain's plan combines DCA-out (strategy 3, scheduled over 7 years 2027-2034 for retirement) and a 15 % partial exit at the next peak identified by technical rule (strategy 4, triggered if MVRV > 3 between 2028 and 2029). This combination targets a scheduled portion (retirement security) and an opportunistic portion (capture of a cycle peak).
Tax aspects by jurisdiction (CH, FR, DE, IT) in May 2026
Exit taxation is, in some jurisdictions, more determining than the timing strategy. Quick overview of CapBitcoin's 4 target regimes on 27 May 2026. This information is indicative and calls for verification by a local tax adviser before any action. Full detail will be covered in the upcoming taxation topic (sprint 6).
Switzerland (Sylvain's case). For an individual who holds Bitcoin in his private estate (not in his business), realised capital gains on movable assets are exempt from income tax, in line with federal and cantonal tax practice applied to individuals not qualified as « professional securities traders ». However, the Bitcoin position remains subject to the annual wealth tax (cantonal rate, ~0.3 to 0.6 % in Geneva on the value at 31 December). Sylvain therefore has, for his scheduled exit, practically no tax constraint on the gain: he can sell 1 BTC at 100,000 CHF in 2028 without paying tax on the gain, provided he stays in the private individual status (no frequent trading, no leverage, no debt financing). The « professional securities trader » condition is defined by 5 criteria in AFCAFC (Administration Fédérale des Contributions)Swiss federal tax administration, which sets the criteria for crypto wealth and professional income.See in the lexicon → circular n° 36 (short holding duration, high transaction volume, use of leverage, etc.). Sylvain is far outside these criteria.
France. Cryptoasset capital gains are taxed at the PFUPFU (Prélèvement Forfaitaire Unique)French 30 percent tax on capital gains, including crypto gains. Also called the « flat tax ». Made up of 12.8 percent income tax and 17.2 percent social levies.See in the lexicon → (single flat rate withholding) of 30 % since 2019 (12.8 % income tax + 17.2 % social levies), for individuals in occasional private management. The option for the progressive scale is possible if more favourable. Note: only the exchangeExchangeService that lets you buy, sell and swap cryptocurrencies against fiat money. Examples : Kraken, Coinbase, Bitstamp, Bitvavo. Most are custodial.See in the lexicon → against fiatFiat (fiat currency)State currency with legal tender status (euro, Swiss franc, dollar), issued by a central bank and not backed by a physical asset. By contrast, Bitcoin has an issuance capped at 21 million units, with no central issuer.See in the lexicon → currency or against good or service triggers taxation, crypto-to-crypto exchange remains tax-neutral. Pricing: a French citizen who sells 50,000 EUR of gain pays 15,000 EUR of PFU. Spreading over several tax years is possible and advised for large positions. The « habitual trading » status (BIC) remains rare but triggers progressive scale taxation up to 45 %, to be avoided by low-frequency strategy.
Germany. The German regime is very favourable to long-term holding: capital gains on Bitcoin held more than 12 months are fully exempt, in line with §23 EStG§23 EStG (Spekulationsfrist)German tax provision that fully exempts Bitcoin capital gains after a holding period of more than 12 months.See in the lexicon → (Spekulationsgewinne). Bitcoin is treated as a private asset under the Spekulationssteuer regime, and holding > 1 year purges taxation. For Bitcoin held less than 12 months, the gain is taxed on the progressive scale (up to 45 % plus 5.5 % Soli). Operational consequence: a German who exits his position after 1 year of holding pays nothing. It is the most favourable regime of CapBitcoin's 4 jurisdictions for a long-term exit strategy.
Italy. Since the 2023 finance law then the 2025 revision, capital gains on cryptoassets are taxed at a flat rate of 26 % for individuals (raise to 33 % proposed in the 2026 finance law, applicable from fiscal year 2027 according to the latest indications). The annual 2,000 EUR allowance was removed in the 2025 version. Capital losses on cryptoassets are deductible from other cryptoasset gains over 4 years. Spreading over several tax years is possible. An Italian who exits 100,000 EUR of gain in 2026 pays 26,000 EUR, in 2027 will potentially pay 33,000 EUR (subject to final adoption of the 2026 measure).
Synthesis for Sylvain. The Swiss regime is exceptionally favourable for his scheduled exit: no tax on the gain. His constraint is therefore not tax-based (spreading, optimisation), but purely linked to timing and psychology. If he moved to another country before his exit, the calculation would be radically different. For a French or Italian investor in the same situation, DCADCA (Dollar Cost Averaging)Buying a small fixed amount at regular intervals (for example 100 EUR a week), regardless of price. Smooths the average purchase cost and neutralises timing bias.See in the lexicon →-out spreading becomes even more relevant fiscally: it avoids the concentration of a large gain on a single year.
Bitcoin-as-collateral: borrow against your BTC rather than selling
An alternative path to selling exists and deserves consideration for occasional liquidity needs. Instead of selling Bitcoin to obtain fiatFiat (fiat currency)State currency with legal tender status (euro, Swiss franc, dollar), issued by a central bank and not backed by a physical asset. By contrast, Bitcoin has an issuance capped at 21 million units, with no central issuer.See in the lexicon →, the investor deposits his BTC as collateral with a specialised lender and receives a loan in EUR or USD. The loan is repaid later, in fiat or by recovering the deposited BTC. Bitcoin is never sold, the long-term position is preserved, and the borrower has the liquidity needed for his use (home purchase, financing, health expense, etc.).
Operational mechanism. Sylvain wants to buy an apartment at 350,000 CHF to rent out as a retirement supplement. Rather than selling 3.7 BTC to generate the sum, he deposits 5 BTC (~470k CHF) as collateral with a lender (Ledn, Unchained Capital, two surviving post-MiCAMiCA (Markets in Crypto-Assets)European regulation 2023/1114 that frames crypto services across the EU since 2024. Creates the CASP status.See in the lexicon → actors in 2026). The accepted LTV (Loan-to-Value) ratio is typically 30-50 %. At 40 % LTV, he borrows 188,000 CHF, at a variable rate of around 9-12 % annual (rates have risen sharply since 2022). He combines this loan with a classic bank mortgage to close the property financing.
Advantages. No sale triggered, therefore no capital-gain tax (determining for taxing jurisdictions like France or Italy). Bitcoin position maintained: if BTC doubles within 2 years, the gain remains captured. Liquidity obtained quickly (generally 24-72h after collateral deposit).
Serious drawbacks. The high interest rate (9-12 %) absorbs part of Bitcoin's potential appreciation. Above all, the margin call risk: if BTC drops -40 % from the loan moment, LTV jumps to 67 % and the lender demands additional collateral or automatically liquidates part of the BTC to bring LTV back under threshold. For Sylvain borrowing at 40 % LTV in May 2026 at 94,000 CHF/BTC, his forced-liquidation threshold lies around 56,000 CHF/BTC (-40 %). If a new -80 % bear occurs (not historically improbable), Sylvain loses part of his BTC against his will, at the worst moment of the cycle.
Counterparty risk. The lender itself can go bankrupt (cf. Celsius, BlockFi, Voyager in 2022). Surviving actors in 2026 (Ledn, Unchained) have more prudent models, strict collateral separation, regular audits, but zero risk does not exist. MiCA regulation in Europe now frames these activities, which reduces risk without eliminating it.
When it makes sense. Bitcoin-as-collateral becomes relevant when: (a) the liquidity need is occasional and identified (home purchase, emergency), (b) the jurisdiction heavily taxes gains (France 30 %, Italy 26 %), (c) the investor has strong conviction on medium-term Bitcoin rise, (d) the LTV ratio stays conservative (30 % max to resist a -70 %). For Sylvain in Switzerland, favourable taxation makes this path less obvious: better to sell 3-4 BTC without paying tax than to pay 9-12 % annual interest. Sylvain will therefore not use Bitcoin-as-collateral. For an equivalent French Pierre, the calculation would change.
Disclaimer
Educational and informational content only: not investment, tax or legal advice. Bitcoin carries significant risks, including high volatility and the possible loss of invested capital. Each reader remains responsible for their decisions; when in doubt, consult a qualified professional in your jurisdiction.
Going further
Bitcoin exit strategy interfaces with all other articles of the Invest topic. To dig deeper:
- Invest Bitcoin guide: overview of strategies and asset allocation.
- Bitcoin DCA method: the entry strategy mirroring the exit DCADCA (Dollar Cost Averaging)Buying a small fixed amount at regular intervals (for example 100 EUR a week), regardless of price. Smooths the average purchase cost and neutralises timing bias.See in the lexicon →-out.
- Lump-sum vs DCA: the 47-year-old Pierre case that informs long-term planning mechanics.
- Bitcoin halving cycles: the cyclical structure underlying tactical exit triggers.
- Gold and Bitcoin allocation: to understand Bitcoin's place in a diversified estate, and therefore how much to exit.
- Bitcoin volatility psychology: why unplanned sales fall at the worst moment.
For operational implementation of the exit:
- Store Bitcoin guide: hardware walletHardware walletSmall dedicated device (Ledger, Trezor, Coldcard, BitBox, etc.) that keeps the private key away from a potentially compromised computer. Signs transactions inside the device itself.See in the lexicon → self-custodySelf-custodyModel in which you hold your own private keys. Your bitcoins depend on no third party. This is Bitcoin's founding promise.See in the lexicon →, step before any transfer to exchangeExchangeService that lets you buy, sell and swap cryptocurrencies against fiat money. Examples : Kraken, Coinbase, Bitstamp, Bitvavo. Most are custodial.See in the lexicon → for sale.
- Buy Bitcoin guide: the platforms used mirror-wise for selling (Relai, Bitstamp, Kraken).
The upcoming taxation topic (sprint 6) will deepen jurisdictional regimes (Switzerland, France, Germany, Italy) with detailed declarative modalities and optimisation cases by relocation, donation, inheritance. Recommended readings beyond CapBitcoin: Andreas Antonopoulos on the monetary dimension of Bitcoin (Mastering Bitcoin, 3rd edition 2023), Lyn Alden on macro-allocation (Broken Money, 2023), and Galaxy Digital's annual report on retail behaviour (2024 edition, publicly available).