Fail Conventionally

Something is moving among the largest investors in the world. Pension funds, university endowments, family offices and sovereign wealth funds are building positions in bitcoin. Since spot bitcoin ETFs launched in the United States, their names have started to appear in public filings: Harvard, Brown and Dartmouth among the universities, Mubadala and the Abu Dhabi Investment Council among the sovereigns. And yet the amounts remain small, and much of what these institutions do stays out of view.

A new study by the asset manager Bitwise explains why. Between late March and April 2026, Bitwise interviewed the people responsible for crypto decisions at 15 large institutions. The report and its numbers are summarized in our news coverage. Its most revealing line comes from a public pension fund. Some of the people the fund answers to, an official said, "would be a lot happier if performance were a lot more mediocre, but we were not in the press." Bitwise summed it up in one sentence: the only thing worse than losing is winning with crypto.

This essay argues that this sentence, not any chart of inflows, explains how institutional money reaches bitcoin. The institutions are not waiting for proof that bitcoin works. Most of them decided that long ago. They are waiting for permission, the point at which holding bitcoin no longer puts a career at risk. A great deal of capital is waiting for exactly that point right now.

The buyer everyone is waiting for

Bitwise makes no secret of where it thinks the next wave of demand will come from. In July 2026, its chief investment officer Matt Hougan put it bluntly: "The end boss of investing is institutional capital." Financial advisers, pension funds, endowments and sovereign wealth funds, in his reading, are the buyers who will take over from companies like Strategy. That thesis is examined in Who Buys Bitcoin After Strategy. Bitwise sells products to exactly these institutions, and that belongs next to every one of its conclusions.

What these investors allocate is small by design. In the Bitwise interviews, holdings ranged from 0.5% to 13% of investable assets, with most between 1% and 2%. That is close to what the large banks now tell their own clients. BlackRock described 1% to 2% as a reasonable allocation in December 2024 and advised against going beyond 2%. Morgan Stanley's investment committee set up to 4% for its most growth-oriented portfolios in October 2025, and Bank of America let its advisers recommend 1% to 4% from January 2026.

Among bitcoiners, a popular calculation circulates from here: if only a few percent of the world's institutional capital moved into bitcoin, the price would multiply. The scale behind it is real, as a later section shows. The multiplication is a forecast, and this essay does not make it. The more useful question is a different one. If the case is so well known and the allocations are so small, what is holding the money back?

The case is made. Permission is not.

The Bitwise interviews give a clear answer. For the institutions, the question of whether bitcoin has merit was usually the easiest hurdle. What slowed them down was custody, the question of where the asset belongs in a portfolio built from stocks, bonds and alternatives, and above all the question of how a position would look from outside.

At one sovereign wealth fund, the allocation was examined by the leadership of the country's central bank. The questions were mostly about perception: how the position would look, and whether other sovereigns and central banks had moved first. "Frankly speaking, it was really hard to pitch our first crypto investment," the fund told Bitwise. Among the nonprofit clients of one consultant, committee members still ask "what is this rat poison that Warren Buffett told me is going to zero?" The same consultant does not expect those clients to allocate because of "performance on a chart." They will allocate "because the next generation joins the committee."

None of these are objections to bitcoin. They are objections to being seen holding it.

John Maynard Keynes described this problem in 1936, in the twelfth chapter of his General Theory. A long-term investor who goes against average opinion, he wrote, looks eccentric and rash to everyone else. If he succeeds, that only confirms his rashness in their eyes, and if he fails in the short run, he can expect little mercy. His conclusion: "Worldly wisdom teaches that it is better for reputation to fail conventionally than to succeed unconventionally."

The rule has a second half, and it is the one that matters now. Once enough peers hold something, the convention flips. A senior investment professional at a large US endowment put it to Bitwise this way: "If your peers own crypto and you don't, you are structurally short relative to what you're being graded on." From that point on, the career risk no longer sits with the institution that buys. It sits with the one that did not.

This is why adoption does not arrive at a steady pace but in clusters. Bitwise found endowments asking each other about their exposure more directly than before, and one named peer benchmarking as the main reason it allocated. Each disclosed position lowers the cost of the next. Bitwise describes the likely path as "more likely exponential than linear." It is also why so much seems to be happening at once. It usually is, because every move makes the next one cheaper.

The turning point that made the first moves possible was the approval of spot bitcoin ETFs in January 2024. The SEC's chair at the time stressed that the agency "did not approve or endorse bitcoin." For institutions, the approval worked as a form of permission all the same. Bitwise found that ETFs did more than simplify access. They made the asset class look legitimate, and they made a bitcoin position look ordinary from the back office. Holding bitcoin through a product from BlackRock is a different conversation with a board than holding a coin the public still associates with the darknet.

Bitcoin first, everything else on probation

The study shows another thing clearly. Every institution in the sample that holds digital assets holds bitcoin. For nearly all of them it was the first, largest and longest-held position, and Bitwise calls it the only digital asset with consistent institutional conviction. Several institutions weight a basket of leading assets by market value, which still puts about 80% of their holdings in bitcoin. Most simply hold bitcoin on its own.

Ether and Solana are treated differently. Institutions that hold them keep smaller positions over shorter horizons and describe them as "venture-stage technology bets", with explicit conditions for selling. Several said they would exit if real use does not translate into value for the token within a few years. Some hold neither.

The difference is not a legal one. In March 2026, the SEC and CFTC listed bitcoin, ether, Solana and thirteen other assets as examples of digital commodities, all in the same category. The institutions draw the line themselves.

Bitcoiners have long argued that networks like Ethereum and Solana are closer to young fintech companies than to money: more widely distributed than a bank, but still shaped by foundations, core teams and roadmaps. The institutions in the study did not use that language. But they hold these assets the way one holds a stake in a young company, with a thesis, milestones and a date by which the team must deliver. Our reading, and it is our reading rather than something the interviewees said, is that this is the real dividing line. A store of value cannot have a management, because a management is precisely the risk its holder wants to avoid.

The histories of the networks show where the distinction comes from. In July 2016, after an attacker drained a large investment fund built on Ethereum, the network's developers shipped a hard fork that moved about 12 million ether from the affected contracts into a recovery contract, and the large majority of the network followed. In February 2024, Solana stopped processing transactions for about five hours because of a software bug. The restart required validator operators to agree on a common starting point and switch to patched software, and the Solana Foundation warned that operators who failed to upgrade would lose their delegation status. In both cases, a small and identifiable group decided what happened next.

Bitcoin has had such moments too, and it would be dishonest to leave them out. In August 2010, a bug allowed a transaction in block 74,638 to create about 184 billion bitcoin. A patched client appeared within five hours, and the corrected chain overtook the faulty one the next day. Today, the two largest mining pools found about 45% of all blocks in the month to late September 2026. What separates the cases is what the coordination was for. The 2010 fix restored a rule every participant already relied on, that no more than 21 million bitcoin can exist. The 2016 fork reversed an outcome the rules had allowed. And a pool, however large, cannot change bitcoin's rules, because every node checks them independently. Why that makes bitcoin so hard to change is the subject of Bitcoin Changes Only by Consensus.

That is also what censorship resistance means in practice. A stablecoin can be frozen by its issuer, as Circle and Tether did with part of the funds stolen from Bitget in September. The dollar that the GENIUS Act made programmable can be stopped by design, as our analysis of stablecoins and the dollar describes. Bitcoin has no issuer to ask. An institution that buys bitcoin is paying for exactly that absence.

Digital gold, filed next to gold

Most interviewees placed bitcoin next to gold, the comparison explored in Bitcoin vs. Gold. Several endowments built bitcoin and gold positions in parallel. One institution keeps bitcoin in its gold allocation and told Bitwise: "We could be having this conversation in 10 years and we're telling you we gave up on gold and it's all bitcoin now." A large endowment whose rules bar it from holding spot commodities, even through an ETF, still describes bitcoin as the digital equivalent of gold and the anchor of the asset class.

In the most striking case, a sovereign wealth fund is paying for part of its crypto allocation by selling foreign currency and gold reserves. That is a state institution moving money out of the oldest reserve asset and into the newest one.

The comparison also has limits worth naming. Not every institution accepts it. One foundation in the study rejects the digital gold framing and treats the whole sector as disruptive technology. And gold still outweighs bitcoin in many of the portfolios that hold both. At the end of March 2026, Harvard's endowment held about 199.8 million dollars in a gold ETF and about 117.0 million dollars in a bitcoin ETF.

Do sovereigns hold more than anyone can see?

Abu Dhabi's Mubadala reported 14,721,917 shares of BlackRock's bitcoin ETF at the end of March 2026. That raises a question often asked about sovereign buyers. Why would a state fund hold its bitcoin through a US product, when the United States could freeze it in a conflict? The question is not paranoid. Since Russia's invasion of Ukraine, around 210 billion Euro of the Russian central bank's assets have been immobilized in the European Union alone.

Most public knowledge about institutional bitcoin comes from Form 13F. Investment managers with discretion over 100 million dollars or more in listed securities file it every quarter. It shows spot ETF positions at quarter end. It does not show bitcoin held directly or stakes in private venture and hedge funds.

Put the permission problem next to that rule, and a pattern follows. The institutions most exposed to headlines have the strongest reason to choose vehicles that make none. Bitwise found exactly that. The public pension funds in its sample hold digital assets only through venture and hedge funds, which fold quietly into existing buckets for alternative investments. One institution said a 13F-visible ETF position creates public visibility it would rather avoid. And one sovereign wealth fund is building domestic custody to satisfy a government mandate for direct control of its assets. Bitwise's conclusion is that estimates of institutional ownership based on 13F filings "should be seen as a floor, not a ceiling."

That gives weight to a reading that circulates widely among bitcoiners: that some sovereigns hold a visible ETF position partly as a gesture toward Washington and Wall Street, while building larger holdings of actual bitcoin in their own custody. The study supports the mechanism behind that reading. It does not support naming anyone. The report identifies no institution and puts no number on what sits outside the filings. Any claim about a specific fund's hidden holdings is a guess, however plausible it sounds.

Visibility also misleads in the other direction. In May 2026, Texas announced that it will move its reserve out of the ETF and into bitcoin held directly with a custodian, and that the custodian must run a public website showing the holdings. Leaving the ETF, in that case, means more transparency, not less. The government of Bhutan is watched on-chain by analytics firms, which have tracked about 1 billion dollars in transfers from its wallets to exchanges and trading firms since July 2025. Those transfers were widely read as sales. The head of its investment arm said he did not recall the last time it sold any bitcoin. A missing filing does not prove an absence, and a moving wallet does not prove a sale. What can be said with confidence is narrower: the visible holdings are a lower bound, and the institutions with the most to lose from publicity are the most likely to sit below it.

Why this money stays

Sovereign wealth funds are the slowest movers in the study, at 1.0% to 1.5%. One told Bitwise that even with the support of the country's president, building the legal and regulatory framework to invest state capital took more than a year. Others have announced timelines of several years.

But once they are in, several hold for reasons that go beyond returns. Some sovereigns told Bitwise they invest in digital assets partly to attract foreign capital, to support a national crypto initiative or to position their country as a hub for the industry. One described its allocation as a "multi-year bet on achieving global recognition rather than near-term return." Perception cuts both ways here. A bitcoin position can signal that a country is open to new technology, or it can look like a reckless bet on something the public still links to crime. Which reading wins depends on who has already moved. That is the permission problem again, played out between states.

A position with a purpose beyond returns can survive returns that disappoint. None of the 15 institutions Bitwise interviewed reduced its allocation during the roughly 50% decline between the fourth quarter of 2025 and the second quarter of 2026, and several bought more. Asked what would make them sell, none named price. They named a failed thesis, a regulatory reversal or a crisis of credibility across the industry.

Who goes first, and who goes last

If permission is the bottleneck, the institutions that need the least of it should move first. That is what the interviews show. Bitwise found that allocation size falls as the number of people who must approve it rises. Where one person can decide, the money moves. Where a committee must agree, it often stalls.

Family offices often answer to a single principal and can decide within a day. They reported the highest allocations, with a typical target of 5% and positions of up to 13%. One consultant said the families "don't have investment committees, and they don't have public-facing risks that trip them up." Endowments come next. Their decisions often rest with a chief investment officer and a small team, and most hold 0.5% to 2%. Public pension funds answer to boards, elected officials, retirees and the local press. They hold 1.5% to 4.5%, and one team stopped talking to the press about its allocation altogether, "because crypto is so polarizing to the public." Sovereign wealth funds move last.

The public record follows the same order. The state investors that have moved did so in small, deliberate steps. Luxembourg's sovereign wealth fund put 1% into bitcoin ETFs in October 2025. The Czech National Bank created a 1 million dollar test portfolio in November 2025, kept apart from its reserves. Texas bought 5 million dollars of a bitcoin ETF the same month. These positions look less like bets sized for returns than like steps sized to learn, and to be seen learning. The early movers carry the career risk that the later ones are spared.

How large is large

The weight of the pension and sovereign question comes from the size of the pools behind it. Global pension assets reached 68.3 trillion dollars at the end of 2025, according to the Thinking Ahead Institute. State-owned investors, a group that includes central banks, sovereign wealth funds and public pension funds and therefore overlaps with the pension figure, managed about 63.8 trillion dollars in August 2026, according to Global SWF. On September 29, 2026, one bitcoin cost about 83,600 dollars, and the roughly 20.09 million bitcoin in existence were worth about 1.68 trillion dollars.

A 1% share of global pension assets would be 683 billion dollars, about 41% of bitcoin's market value.

This is where the popular calculation takes one more step and turns that ratio into a price. The step is a forecast, and a weak one. Money flowing into an asset does not add to its market value one for one, every buyer needs a seller, and allocations drift with prices and with committee decisions. What the arithmetic shows is scale, and nothing more. The pools of capital now debating an allocation are large relative to the asset, and a position that looks small inside a portfolio is not small for bitcoin.

It also shows how early the process still is. Bitwise cites estimates that retail investors control more than two-thirds of the crypto market. CoinShares found that 13F filers held 20.8% of the assets of US spot bitcoin ETFs at the end of March 2026. If the shift continues, it changes who holds bitcoin and why. Bitwise expects a majority of institutional investors to hold digital assets within five years. That is an expectation from a firm with an interest in it, not a finding. The direction the interviews describe is clear. The pace is not.

Is the study too bullish?

Three objections deserve to be taken at full strength.

The first is that Bitwise asks the people who already agree with it. Bitwise manages only digital assets. Fifteen interviews with institutions selected in an undisclosed way, by a firm that sells to them, are not a representative study, and a broadly diversified manager would likely hear a different mix of answers. The public filings from the same months show institutions selling during the drawdown in which none of Bitwise's interviewees did. Harvard's endowment reduced its position in BlackRock's bitcoin ETF by 55% between the end of September 2025 and the end of March 2026 and sold its ether ETF entirely. Across all 13F filers, CoinShares counted endowment holdings down 40% in the first quarter of 2026 alone.

The second is that conventions run both ways. The mechanism that makes adoption reflexive can make retreat reflexive too. Bitwise itself warns that a long downturn could make holding crypto the reputationally exposed choice again, and that a major crisis in the sector could reset the career-risk calculation for years. An argument built on social permission has to accept that permission can be withdrawn.

The third comes closest to bitcoin's own purpose. Institutions that arrive through ETFs and custodians rebuild the intermediaries that bitcoin was designed to make optional. An ETF share is a claim on a custodian, not a key. If adoption means that the largest holders never touch the asset, something is gained in legitimacy and something is lost in what the asset is for. That tension runs through Bitcoin Isn't the Money You Were Promised.

Each objection is partly right. The first limits how much weight Bitwise's numbers can carry, but the argument here does not rest on them. It rests on a mechanism that predicts institutions moving in clusters and in both directions, and the filings show exactly that. The second is conceded in full, which is why this essay claims that permission sets the pace, not that the pace can only rise. The third is the most serious, and the answer is not complete. But at the edge of institutional adoption, the movement is toward the asset itself: a sovereign wealth fund building custody so that it controls the coins directly, a US state leaving its ETF for bitcoin held directly through a custodian. And nothing an institution does takes away the ability of an individual to hold their own keys, as described in Self-Custody Best Practices.

Permission, not proof

The institutions that allocate in the coming years will not be braver than the ones that allocated in the last few. They will simply no longer be alone. Keynes's rule does not say that conventional investors are wrong. It says they are late, and that being late is the price they pay for being safe.

That is why so much capital is waiting now, and why so much of what moves does so quietly. The case for bitcoin was made years ago, by people who held it when doing so looked eccentric and rash in the eyes of average opinion, in the phase described in Nobody Was Watching. What remains is a social process in which each institution waits for the one before it. That process is slow, it can reverse, and much of it will happen where no filing can see it. But what decides it is something bitcoin already has and its competitors cannot easily acquire: no one to blame if it fails, and no one to ask if it should change.