The Tower of Babel

There is a moment in the building of every speculative structure when the thing being built stops making sense on its own terms and continues only because everyone has agreed not to notice. The digital-asset treasury company — the DAT — is the purest such structure the current cycle has produced, and to see why it is so dangerous you have to first appreciate why, for a time, it genuinely worked. A DAT is, at its core, a public company whose business is to own cryptocurrency. It raises money by selling its own stock and bonds, uses the proceeds to buy bitcoin (or, increasingly, ether or solana), and holds it. That is the whole business. It builds nothing, sells nothing, invents nothing. It is a publicly traded pile of coins.

The strange and crucial fact is that these piles of coins have traded, for most of the boom, at a premium to the coins inside them — sometimes an enormous one. The metric that captures this is called mNAV, the "multiple on net asset value," which compares a company's market value to the market value of the crypto it holds. An mNAV of 1.0 means the company is worth exactly its coins. An mNAV of 2.0 means investors are paying two dollars for every one dollar of bitcoin the company owns — which, stated plainly, is irrational, because you could simply buy the bitcoin yourself for one dollar. And yet at the peak of the mania, the mNAVs of the hottest DATs did not sit at 2; they ran, in some cases, above twenty-five. Investors were paying twenty-five dollars for a dollar of coins. To understand why that was not pure madness — and why its unwinding is so dangerous — you have to understand the machine.

The machine that prints value from a premium

The magic of the DAT, pioneered and perfected by Michael Saylor's Strategy (the company formerly known as MicroStrategy), is a genuine financial flywheel, and within its own logic it is real. It works like this. Suppose a company holds $1 billion of bitcoin and, because investors are enthusiastic, its stock trades at a market value of $2 billion — an mNAV of 2.0. The company now issues new shares, selling $1 billion of fresh stock into that enthusiasm. It uses the proceeds to buy another $1 billion of bitcoin. It now holds $2 billion of bitcoin, but here is the trick: because it sold the new shares at a premium to the coins behind them, the existing shareholders end up owning more bitcoin per share than they did before. The dilution is accretive. Selling overpriced stock to buy bitcoin makes each old share richer in bitcoin terms. The premium, in other words, is not just a quirk of sentiment; while it lasts, it is a tool the company can wield to manufacture bitcoin-per-share out of thin air, which in turn appears to justify the premium, which enables more issuance, which creates more bitcoin-per-share. Around and around. It is a machine that converts investor enthusiasm directly into shareholder value, and for a glorious stretch it minted fortunes.

This is the engine that made Strategy one of the great-performing stocks of the era and inspired, by various counts, dozens upon dozens of imitators — one estimate put it at 89 copycats spawned directly by Strategy's example, and the total population of crypto-treasury companies now runs to something like 190 to 195 firms holding bitcoin, ether, solana, and more. Everyone saw the same magic trick and tried to repeat it: raise stock at a premium, buy coins, watch the flywheel spin. For a while, in a rising crypto market with abundant enthusiasm, it spun for many of them at once. The Tower rose, floor upon floor, each builder copying the last, all speaking the same fluent language of mNAV and accretive dilution and bitcoin-per-share, all reaching toward the same limitless sky.

The flywheel has a reverse gear

Here is the catastrophe hiding inside the machine, and it is not a bug but a defining feature: the flywheel runs in both directions. Everything that makes the DAT a value-creation engine when the premium is positive makes it a value-destruction engine when the premium disappears. The entire mechanism depends on one condition: the stock must trade above the value of its coins, at an mNAV greater than 1.0. Above 1.0, issuing shares to buy crypto is accretive and the company can keep accumulating. But the moment the mNAV falls to 1.0 — the moment the market will only pay a dollar for a dollar of the company's bitcoin — the magic dies, because now issuing new shares to buy crypto is no longer accretive; it is merely dilutive, transferring value from existing holders to new ones for no gain. The accumulation has to stop. The flywheel seizes.

And below 1.0 — when the company trades at a discount to its coins, which a growing number of DATs now do — it gets actively destructive. A company worth less than the crypto it holds is the market's explicit verdict that something is wrong: that the firm will dilute its holders, or mismanage the assets, or, most dangerously, be forced to sell its crypto to meet obligations. Because here is the part the euphoric flywheel narrative glossed over: many DATs did not just sell stock to buy crypto. They issued debt — convertible bonds and loans — to lever up their coin purchases and amplify the flywheel. That debt is a fixed obligation that must be serviced and repaid on a schedule, regardless of what bitcoin does. In a rising market, debt supercharges the gains. In a falling market, it becomes a noose. If crypto drops and the company's mNAV is below 1.0, it cannot issue equity to raise cash without destroying shareholder value, and it may be left with only one option to meet its debts: sell the bitcoin. Selling pushes the price down further, which pushes more DATs underwater, which forces more selling. The same reflexivity that drove the Tower upward drives it down, and faster, because fear moves quicker than greed.

The premium era is over

This is not a hypothetical. It is happening now, in 2026, and the numbers are stark. The mNAVs that once exceeded 25 have been collapsing toward 1.0 across the sector, and a significant and growing number of DATs are now trading below 1.0 — at a discount to the very coins they hold. The premium, as one industry headline flatly declared, era "is over." Standard Chartered, hardly a crypto-skeptic institution, has warned explicitly of an mNAV collapse across the digital-asset-treasury space. The poster child of the reversal is Metaplanet, the Japanese DAT that became a sensation: it traded at a premium of roughly 237% to its bitcoin holdings at its peak and then plunged to a 10% discount — a breathtaking round trip from "investors will pay $3.37 for a dollar of my bitcoin" to "investors will only pay 90 cents." Of the roughly 190-odd treasury companies now in existence, only a handful still trade above their net asset value — most sit in an mNAV band of roughly 0.8 to 1.7. The land grab is over; 2026 is, in the words of analysts covering the space, a "survival of the fittest," in which the weakest DATs — those with debt, poor management, and no business beyond holding coins — are expected to fail outright or be absorbed, their carcasses becoming acquisition targets for the few survivors.

Consider what an mNAV below 1.0 actually represents, because it is one of the most damning verdicts a market can render. It means investors would rather own the underlying bitcoin directly than own it through this company — that the corporate wrapper, far from adding value, is now subtracting it. The premium was the entire reason for the DAT to exist; a bitcoin treasury company trading at a discount to its bitcoin is a contradiction in terms, a structure whose only purpose was to be worth more than its contents now worth less. And the discount is not merely an indignity; it is a trap, because it shuts off the equity-issuance lifeline at exactly the moment a leveraged, crypto-exposed company most needs cash. The discount is the market pre-pricing the forced selling it expects.

The reason they were supposed to exist

It is fair to ask why a DAT should command any premium at all — why a rational investor would ever pay more than a dollar for a dollar of bitcoin held inside a corporate shell. The bull case had real answers, and they are worth stating, because their erosion is part of the story. First, access: many institutions, retirement accounts, and funds are restricted from holding crypto directly but can freely buy stocks, so a DAT was a way to get bitcoin exposure inside a conventional brokerage account — a wrapper that solved a real plumbing problem. Second, leverage: a DAT could borrow against its position and amplify returns in a way an individual holder could not easily replicate, offering a geared bet on crypto. Third, the flywheel itself: the demonstrated ability to grow bitcoin-per-share through accretive issuance was a genuine, if circular, source of value that direct ownership did not provide. For these reasons, a modest premium was arguably defensible, and the early movers earned it.

But each of those pillars has weakened, and that is why the premium is collapsing rather than merely wobbling. The access argument has been gutted by the proliferation of spot bitcoin and ether exchange-traded funds, which now let almost anyone hold crypto directly, cheaply, in any ordinary account — removing the very plumbing problem the DAT existed to solve. Why pay a premium for bitcoin in a leveraged corporate wrapper when you can buy a low-fee ETF that holds it one-to-one? The leverage argument cuts both ways and looks far less appealing after investors have seen what leverage does on the way down. And the flywheel argument is self-refuting the moment the premium falls, because the flywheel requires the premium to function — it cannot generate the premium it depends on. Strip out access (now solved more cheaply elsewhere) and the flywheel (now stalled), and what remains is leverage, which is not a feature investors pay a premium for; it is a risk they demand a discount for. The rationale for the premium has been dismantled piece by piece, which is exactly why the mNAVs are converging on, and breaking below, 1.0.

Why the copies are weaker than the original

It is worth distinguishing the original from the imitations, because the distinction matters for how the unwinding plays out. Strategy itself — the prototype — is the strongest of the species: it accumulated an enormous bitcoin position (hundreds of thousands of coins) early and cheaply, built a deep and relatively sophisticated capital structure, and has a first-mover scale and credibility the others lack. Even Strategy is not immune to the reflexivity, and its own forced-seller risk has been examined in its own right. But it at least has scale, history, and a cost basis far below current prices. The nearly two hundred imitators, by and large, do not. They came later, bought their coins at higher prices, raised their debt on worse terms, and built their towers on the assumption that the premium that existed when they launched would persist. Many are small, thinly capitalized, run by promoters who saw a hot trade rather than operators who understand leverage, and they are precisely the firms most likely to be pushed into forced liquidation when the premium vanishes and the debt comes due.

This is the danger of a Tower built by imitation: the copies are always weaker than the original, because the original captured the genuine opportunity and the copies captured only the appearance of it, arriving after the easy gains were gone and the conditions had begun to turn. A mania does not just inflate one company; it spawns a whole population of look-alikes, each one a leveraged bet on the same underlying asset, each one structurally identical, each one exposed to the same trigger — and because they are all the same trade, they are all exposed to failing at the same time, in the same crypto downturn, for the same reason. The nearly two hundred DATs are not two hundred diversified businesses. They are two hundred leveraged wrappers around the same two or three coins, correlated to the hilt, primed to unwind in unison. That is not an industry. It is a single position wearing two hundred different tickers.

The lesson the closed-end funds already taught

None of this is unprecedented; markets have run this exact experiment before, in a quieter asset class, and the result is instructive. Closed-end funds — pooled investment vehicles that hold a fixed portfolio and trade as a stock — have existed for a century, and they exhibit precisely the same behavior as DATs: they can trade at a premium or a discount to the net asset value of what they hold. And the iron lesson of closed-end-fund history is that premiums are temporary and fragile, while discounts are common and persistent. Funds that trade above their net asset value almost always see that premium erode, because there is no durable reason to pay more than a portfolio is worth, and arbitrage and disappointment grind the premium away. Most closed-end funds, most of the time, trade at a discount — often a stubborn one of ten or twenty percent — because the market demands compensation for fees, illiquidity, and the friction of the wrapper. The natural resting state of a vehicle that holds assets and trades as a stock is a discount, not a premium.

DATs are closed-end funds in all but name, holding a single volatile asset instead of a diversified portfolio, often with added leverage — which makes them more prone to discounts, not less, because leverage and single-asset concentration are exactly the features that frighten investors into demanding a margin of safety. The premium era was the anomaly, sustained by mania and the flywheel narrative; the discount era now arriving is the reversion to the historical norm that the closed-end-fund record would have predicted all along. The builders of the Tower believed they had invented a new kind of structure that could trade above its contents forever. They had in fact rediscovered the closed-end fund, an old structure whose century-long history says the premium never lasts and the discount usually wins. The only novelty was the leverage and the volatility of the asset, both of which make the eventual discount deeper and the unwinding more violent. There was nothing new under the sun except the speed.

The confusion of tongues

The Tower of Babel, in the old story, fell not because it was poorly built but because its builders lost the ability to understand one another — the single shared language that had let them cooperate scattered into mutual incomprehension, and the project collapsed into confusion. The digital-asset-treasury boom has its own shared language: the fluent, confident dialect of mNAV and accretive dilution and bitcoin-per-share and the eternal flywheel, spoken in unison by nearly two hundred builders all certain the premium would hold and the sky had no limit. That language worked only as long as everyone believed it. The premium was never a fact about the coins; it was a fact about the belief, a collective agreement to pay more than a dollar for a dollar of bitcoin because everyone else was doing the same and the flywheel made it self-fulfilling. And belief, unlike bitcoin, cannot be held in a treasury. It evaporates.

What we are watching in 2026 is the confusion of tongues — the moment the shared language stops being believed, the premium scatters, and the structure that depended entirely on collective faith begins to come apart floor by floor. The mNAVs collapse toward and through 1.0; the issuance machine seizes; the leveraged copies face their debt with no equity lifeline; the forced selling begins; and a structure that looked, at its height, like a perpetual-motion machine for minting wealth reveals itself as what it always was: a leveraged, reflexive bet on a volatile asset, dressed in the respectable clothes of a "treasury strategy," and worth a fortune only so long as everyone agreed it was. None of this means bitcoin is going to zero, or that every DAT will fail, or that the strongest few will not survive and even prosper. It means that the premium — the thing that made a DAT more than a coin in a corporate costume — was the most fragile part of the whole edifice, and it is going, and as it goes it takes the flywheel, the accumulation, and in many cases the solvency with it. The Tower reached impressively high. But it was built on a shared sentence that everyone agreed to keep repeating, and the market has begun, at last, to fall silent. When the language goes, the Tower follows. It always has.