The best-performing asset of the last decade

Opening

Every generation inherits a set of assumptions about money it never chose and cannot easily examine.

That the supply is managed by someone. That an account balance is a promise rather than a possession. That savings erode quietly unless actively defended. These are not laws of nature. They are design decisions, made by people, revisable by people, and invisible only because they have never been absent.

In 2008 something appeared that was absent of all of them.

Not a company. Not a product. A set of rules that nobody administers, that no single party can rewrite, and that have now run for seventeen years without a governing body, a bailout, or anyone able to change the terms alone. The rules say there will be twenty-one million units and no more.

That is the idea. What follows is the evidence.

What makes it different

Most assets are defined by what backs them - a company's earnings, a government's promise, an industrial use. Bitcoin is defined by a set of properties no other asset holds at once. Six are worth knowing, because everything else on this page follows from them.

A fixed supply of 21 million. There will never be more. No other asset - not gold, not any currency, not any equity - has a supply that cannot be expanded when demand rises. This is the property the rest depend on. It is also the only supply that does not respond to price: when demand for gold rises, miners produce more of it; when demand for Bitcoin rises, the protocol simply raises the difficulty and issuance stays exactly on schedule. Effort rises; supply does not.

Rules that only consensus can change. Bitcoin is not frozen - it improves, and has (the network has adopted upgrades before, and will adopt others, including defences against future threats). But no single party can impose a change. A rule only takes effect if the network agrees to it, which means improvements pass and self-serving changes - like lifting the supply cap - do not. The protection is not that the rules can never change. It is that no one can change them alone.

Ownership with no intermediary. Held properly, Bitcoin is the only asset you can own outright, with no bank, broker, or custodian standing between you and it. The keys can live in your memory. No one can freeze it, seize it by decree, or fail and take it down with them - because there is no one in the middle.

Portability without permission. Any amount, across any border, in minutes, for a cost that does not scale with the sum. Moving a million dollars of Bitcoin is no harder than moving ten. No other store of value moves like this.

Censorship resistance. No state, bank, or platform can block a valid transaction or exclude anyone from the network. Bitcoin does not ask permission and cannot be told no.

Verifiable, and secured by real cost. Anyone can check the entire ledger independently - you verify, you do not trust. And the integrity of that ledger is protected by the real energy cost of producing it: rewriting history would mean redoing all the work that went into it, faster than the rest of the network produces new work. The cost is the security.

Adoption is no longer the question

For most of Bitcoin's existence, the debate was whether serious money would ever touch it. That debate is over - not because anyone won the argument, but because the money already moved.

The funds arrived. US spot Bitcoin ETFs launched in January 2024 and now hold around $100 billion in assets, having taken in roughly $55 billion of net inflows since launch. A single fund - BlackRock's IBIT - accounts for about 62% of the category. These are not crypto-native vehicles. They are the plumbing of conventional finance, and they were built because the demand was already there.

The corporations followed. Somewhere between 180 and 200 public companies now hold Bitcoin on their balance sheets - roughly 1.2 million BTC in total, about 6% of all the Bitcoin that will ever exist. One company, Strategy, holds around 846,000 of those coins at a cost basis near $63.8 billion (as of September 2026).

And here the story turns in a way worth noticing. Through the summer of 2026, the most committed "never sell" holders sold. Strategy, after years of only accumulating, sold a small portion to fund obligations. MARA sold more than 20,000 coins in a single quarter to cut debt and fund operations. The point is not that conviction wavered - both remain among the largest holders alive. The point is that even the most ideological holders started managing the position rather than simply holding it. Buying is not the whole strategy, and the biggest holders on earth demonstrated it in public.

The states arrived - cautiously. The United States established a Strategic Bitcoin Reserve by executive order in March 2025, though it is capitalised from forfeited coins rather than open-market purchases, and the exact count is contested. El Salvador holds around 7,700 BTC. A handful of US states passed reserve legislation, though only Texas has made even a token purchase.

And here is the honest limit. No G7 or G20 central bank holds Bitcoin as a reserve asset. The institutions that could move the most have not moved at all. Adoption is real, it is large, and it is still concentrated in ETFs, corporations, and a few early states - not in the core of the monetary system.

The money has come from funds, companies and a few states. The institutions that could move the most have not moved at all.

That is what "no longer a question" actually means. Not that Bitcoin has won. That the serious money stopped asking whether, and started deciding how much.

And yet it is astonishingly early

Both things are true at once, and the gap between them is the entire opportunity.

Almost nobody owns it. The honest estimate - and it is an estimate, because there is no registry of Bitcoin owners, only models built on wallet data and surveys - is that somewhere around 300 to 375 million people hold any Bitcoin at all. That is roughly 4% of humanity - and that is the generous end. Tighter on-chain reconstructions put it closer to 100–150 million. Either way, more than 95% of the world owns none.

In the United States, the market with the best data, 19% of adults say they have ever used or invested in crypto, and 9% say they currently own any crypto. Ever tried, once, is a fifth of the country. Currently holding is under a tenth.

And in money terms, it barely registers. Here three statements are all true at the same time, and holding them together is the whole point. At Bitcoin's price in September 2026 - around $81,000, a market value near $1.6 trillion:

• Bitcoin is worth about 5% of all the gold ever mined - and about 12% of the gold held specifically as investment.
• It is roughly 1% of the world's broad money supply.
• It is around 0.1% of the global balance sheet - every asset McKinsey counts, real and financial, added up.

Which number is "right" depends only on what you compare it to. Against the entire wealth of the world, Bitcoin is a rounding error. Against monetary gold - the asset it most directly competes with - it is already a meaningful fraction.

It helps to see the gap as a function of price rather than a snapshot. Gold's above-ground stock is worth somewhere around $29 trillion (end-Q2 2026); the portion held as investment, closer to $13 trillion. Against those fixed targets, and Bitcoin's roughly 20 million coins, the arithmetic looks like this:

These are illustrative prices, not targets. They are the same fact viewed from different prices - a way of seeing how much room sits between Bitcoin and the one asset it most resembles. Even a return to its own previous high would leave it below a fifth of investment gold. For Bitcoin to match the entire above-ground gold stock, a single coin would have to be worth roughly $1.45 million. That figure is not a forecast. It is simply what "catching gold" costs, in arithmetic.

One more figure to sit with. Fewer than a million addresses hold a whole Bitcoin or more. Whatever the exact human count behind them, the number of people on earth who own a single entire coin is smaller than the population of many single cities. A complete Bitcoin is already a rare thing to hold, at a point when, by every wealth measure above, adoption has barely begun.

That is the tension this asset lives inside. The smart money has arrived - and almost everyone else, and almost all of the world's capital, is still on the outside.

Performance

Over the decade to September 2026:

Housing measured by the Case-Shiller national index, which tracks prices only - rental income and holding costs are excluded, so it is not directly comparable to the total-return figures above.

Bitcoin has been the best-performing major asset of the last decade.

Fidelity Digital Assets puts the wider record at 11 of the past 15 years as the top-performing asset of any class.

One note on method, because it matters more here than for other assets: Bitcoin moves in multi-year cycles, so any single figure is a window rather than a property. Fidelity's research team addresses this directly, deliberately selecting periods that cover complete market cycles - and its own March 2026 measurement of the preceding decade put Bitcoin above 70% a year. The lower figure in the table reflects six further months of decline. Same asset, same decade, two honest numbers.

Risk-adjusted

The second objection is that the returns simply compensate for extreme volatility. Measured across a full cycle, that has not been the case.

Jurrien Timmer, director of global macro at Fidelity, put Bitcoin's five-year Sharpe ratio at +0.97 as of early 2024, against +0.74 for the S&P 500 and +0.73 for a US 60/40 portfolio. Correlation with the S&P over that period was +19% - weak enough that Bitcoin was not simply a leveraged expression of equity risk.

Volatility is falling

• Early years: annualised volatility above 200%.
• Recent cycles: roughly 45–55%, against ~13% for the S&P 500.
• By late 2023, on a 90-day realised basis, Bitcoin was less volatile than 92 individual constituents of the S&P 500 - individual companies, not the index itself.

The asset most commonly described as too volatile to own has been getting steadily less volatile for its entire existence, while the assets it is compared against have not.

Zero is also a decision

For most of Bitcoin's existence, holding none of it was simply where portfolios started. Nobody was asked to justify zero. The only question considered worth debating was whether there was a reason to move away from it.

That assumption is what has quietly broken - and not because Bitcoin advocates argued against it. Because the institutions ran the numbers.

Fidelity Digital Assets published its position in March 2026. Its research team's view is that the central question is no longer whether Bitcoin merits consideration, but rather what an investor's current allocation is and why - and that a zero weighting now requires a well-informed rationale rather than being the default that needs none.

Three findings from that work are worth separating out, because they point somewhere unexpected.

Where the improvement actually appeared. Adding Bitcoin to a conventional 60/40 portfolio raised both returns and volatility, but the risk-adjusted measures improved - and the most significant improvement in the Sharpe and Sortino ratios came when moving from a 1% to a 3% allocation. Not from a large position. From a small one, sized deliberately.

Where it did not appear. Whether the allocation was funded from equities or from bonds made very marginal difference to the outcome. The same held for rebalancing: applying some mechanism mattered far more than which one was chosen.

What that leaves. By Fidelity's own reading, the biggest difference in results came from the decision to get off zero at all. The execution choices that allocators spend most of their time on turned out to matter comparatively little.

Which reframes the question.

It was a position like any other - and, on the historical record, a position with a cost.

One caveat on all of these numbers

They are institutional figures, produced under institutional constraints.

A fund operates within a mandate, a governance process, position limits, and evaluation periods measured in quarters. In that context a 3% allocation is a significant decision. Fidelity's own analysis acknowledges the point directly, noting that portfolio managers rarely optimise for long-term compound growth alone, and that the mathematically efficient position size can be larger than intuition suggests. Its own forward-looking optimisation - assuming Bitcoin returns fall to 25% a year with volatility of 50% - produced a maximum-Sharpe portfolio holding 9.4% Bitcoin and no bonds at all.

An individual investor has none of those constraints, and none of the protections that come with them. No mandate to answer to, but also no risk committee, no diversification requirement, and nobody positioned to stop a decision made badly. The constraints are different rather than absent, and they cut in both directions.

Which is why the right size is not a number that can be read off someone else's table. It depends on the capital, the horizon, and what happens to everything else if this part goes wrong.

This page does not answer that question.

What the numbers do not say

The drawdowns are real and severe - and they have never broken the thesis. Bitcoin has fallen roughly 84% peak to trough (2018) and roughly 77% (2022). After reaching approximately $125,000 in October 2025, it trades around 38% below that today. In each previous case the supply schedule continued untouched, the network kept producing blocks, and the asset went on to exceed its previous high. The protocol did exactly what it was designed to do while the price did something entirely different. What the drawdowns establish is not that the thesis is fragile, but that holding through one requires having decided in advance that you would.

The current risk-adjusted picture is poor, and that is not hidden here. Fidelity noted that Bitcoin's Sharpe ratio was negative as of February 2026. The same research observed that historically, allocations made at negative Sharpe readings produced positive volatility-adjusted results over the following three-year periods - an observation on a small sample, not a rule, and certainly not a prediction. It appears here because a page that quoted only the favourable readings would not be worth trusting on the others.

Past performance is not a forecast. A 60% compound rate cannot persist indefinitely - at that rate the asset would exceed the value of everything else within decades. The rate declines as the asset matures. How far and over what period is genuinely unknown, and anyone giving you a precise answer is selling something.

What could go wrong

None of what follows is a reason not to hold Bitcoin. They are the reasons position size is a decision rather than an afterthought - and the reasons that decision is worth making deliberately.

• Regulation. Not prohibition - states have largely stopped attempting it - but restrictions on access, custody and on- and off-ramps. Holders in some jurisdictions have already lost access to venues they used for years.
• Concentration. A meaningful share of supply now sits with exchanges, ETPs and a handful of corporate treasuries. A decentralised protocol does not guarantee decentralised ownership - and coins held through an intermediary carry that intermediary's risk, not Bitcoin's.
• The security budget. Mining is currently funded mostly by new issuance, and that issuance ends. Whether transaction fees alone can secure the network at scale is a real open question, though the answer is decades away.
• Cryptography. Sufficient advances in quantum computing would eventually require migration to different signature schemes. The migration is technically understood and the timeline long - but it is not nothing.
• The record is not sterile. A value overflow bug in 2010 and a chain split in 2013 both required coordinated intervention. The network has run continuously since, but "flawless" overstates it.

What the models say - and what they quietly assume

If adoption is early and the asset is scarce, the obvious question is: how far could it go? A number of serious firms have tried to answer with actual models. Their outputs are worth understanding - but only if you understand what produces them, because the number is almost never the interesting part.

Every one of these is a conditional, not a forecast. The honest form is always: if these markets are captured at these rates, the model implies this price. Change one assumption and the answer moves by hundreds of thousands of dollars per coin.

How you even value a thing like this. There are two families. Demand-side models estimate a set of addressable markets - store of value, corporate treasuries, emerging-market savings, central-bank reserves - and ask what share of each Bitcoin might capture. Supply-side models ignore markets entirely and fit a mathematical curve to Bitcoin's price history. The first asks where could the demand come from. The second asks has the past pattern held. Neither is a crystal ball, and both say so in their own documentation.

ARK Invest is the clearest example of the demand-side approach. Their 2030 scenarios run roughly $300,000 in the bear case, $710,000 in the base case, and $1.5 million in the bull case. But the scenarios are just the visible tip. Underneath, ARK assigns each addressable market a penetration rate - what fraction of gold's role Bitcoin takes, what share of corporate cash, and so on - and the price falls out of the sum.

What makes ARK worth reading is not the headline. It is what they revealed when they updated the model. Between versions, they raised the digital-gold contribution - because gold itself rallied, enlarging the target - and cut the emerging-market contribution by around 80%, because stablecoins, not Bitcoin, won the payments role ARK had once reserved for it. Two large assumptions moved in opposite directions, and the headline barely changed. That is the tell: the final number is a balance of guesses, and the guesses are where the real argument lives.

By ARK's own bucket math, moving institutional penetration from the base rate to the bull rate adds something like $380,000 per coin from that single lever. One assumption. That is how much of the "headline number" is actually a judgment call about how fast institutions allocate.

Other models land in the same wide territory by different roads. VanEck's 2050 framework, built on Bitcoin settling a share of global trade, produces a base case near $2.9 million - but its single most sensitive input is not trade share at all; it is how much of the supply ends up permanently vaulted rather than circulating. Power-law models, which fit a curve to fifteen years of price history, put a 2030 median somewhere in the $390,000–$640,000 range - but they carry no notion of why anyone buys, and they fail the moment adoption stops following the past curve.

The spread across all of these - from a few hundred thousand to several million - is not a sign that someone is wrong. It is an honest measure of how much depends on assumptions nobody can verify in advance.

And this is the line that matters most. Every price in this section is arithmetic applied to a gap - the distance between where Bitcoin sits today and where gold, or global money, or the world's capital already is. The models measure the size of that gap with real rigour.

They describe a possibility and price it. Whether the possibility becomes real is exactly the thing no model contains.

The close

The models describe how large the gap could be. What none of them settles is whether it closes - and that is not a gap in the math. It is the whole question, and it sits outside every model.

Strip out the price and what remains is this.

For the first time, there exists a form of property that cannot be inflated, that - held without an intermediary - cannot be seized by decree or frozen by a third party, and that does not depend on anyone's continued good behaviour. Not because someone promised. Because there is nobody to promise, and therefore nobody to break the promise.

That qualification matters. Bitcoin held on an exchange or through a fund is a claim on an institution, and inherits everything that comes with institutions. The property described above belongs to the protocol. Whether it belongs to you depends on how you hold it.

It is a strange thing to have built. It is stranger still that it was built by someone who then walked away, leaving no name, no company and no claim - which is precisely why it works.

The returns are a consequence, not the point. They are what happens when the world gradually discovers that something it assumed impossible has been running quietly since 2009. If that discovery continues, the market will keep repricing what it finds. If it stops, so does the story. The thesis is not a price target. It is a claim about what people eventually do when a genuinely scarce, genuinely neutral monetary asset becomes available to them - and, on the evidence of this page, the discovery has barely started.

Believing that claim is a decision about the world.

What follows

The thesis tells you to hold. It doesn't tell you how much Bitcoin the same capital can end up holding. A given amount buys a certain quantity at spot - and a larger quantity if the position is built rather than simply bought. That difference compounds alongside the asset itself, over the same years, from the same starting sum.

It also comes with a cost, in risk, that can be calculated precisely rather than guessed at.

Neither of those is a question about Bitcoin. They are questions about how a position is constructed - and they are answered before the capital is committed, or not at all.

Sources and notes

Bitcoin decade return calculated from early 2016 to September 2026 closing prices. S&P 500 figures are total return including dividends. Gold measured from approximately $1,100 (early 2016) to approximately $4,300 (September 2026). Housing measured by the S&P Case-Shiller US National Home Price Index, which rose from 166.23 in January 2015 to 331.62 in June 2025; price appreciation only, excluding rental income and holding costs. Sharpe ratio and correlation: Jurrien Timmer, Fidelity - five-year measurement as reported February 2024. Allocation findings, the 11-of-15 record, the decade CAGR above 70%, the mean-variance optimisation result and its stated assumptions, the note on portfolio-manager constraints, and the note on window selection: Chris Kuiper, CFA, "Getting Off Zero: Evaluating Bitcoin in 2026", Fidelity Digital Assets, 25 March 2026. Fidelity presents the optimisation result as model output under stated assumptions, not as a recommendation. February 2026 negative Sharpe observation: Fidelity Digital Assets research, 2026. Volatility comparison against S&P constituents: 90-day realised volatility, late 2023. Drawdowns measured peak to trough.

All figures as of September 2026 and not updated continuously. Verify current values before relying on them.

This page is educational content. It is not investment advice and contains no recommendation to buy or sell any asset. Bitcoin is volatile and capable of total loss.