Crypto Treasury Premium Collapse: Dying or Pivoting?

The trade these companies ran in 2025 depended on one number: mNAV, or the ratio between a company's stock price and the market value of the crypto it holds. When mNAV sits above 1, a company can issue new shares above the value of its underlying bitcoin, use the proceeds to buy more bitcoin, and end up with more BTC per share than before — a genuinely accretive loop that made 2025's newsletter darlings, Strategy, Metaplanet, SharpLink, look like perpetual-motion machines. When mNAV falls to or below 1, the same move does the opposite: issuing stock to buy bitcoin dilutes existing shareholders instead of enriching them. That's what's happened. Strategy's mNAV peaked near 3.4x in November 2024 and has since fallen to at-or-below parity with its own bitcoin holdings — swinging between roughly a 30% discount and near-even depending on the day and whether you count basic or enterprise value. Metaplanet's mNAV topped 3x, a premium of well over 200%, in July 2025; it recovered to a 17% premium by December 2025 after briefly dipping below 1x that October, then kept falling through 2026 to today's discount of roughly 9-28% below its bitcoin value. Sector-wide, K33 measured the average premium falling from about 3.76x to roughly 2.8x by September 2025 — and it has kept compressing since, with the sector's biggest name now trading at or below the value of its own coins. Cryptopolitan puts the share of treasury companies now trading below the value of their own crypto holdings at around 40% and climbing; The Block had it closer to a quarter back in September 2025. Whichever tracker you use, the direction is the same, and it's the only number in this story that matters.

Two things did this. Bitcoin's own drawdown from October's all-time high removed the tailwind that made paying a premium for indirect BTC exposure look reasonable — if a spot bitcoin ETF gives you the same asset with none of the corporate-structure risk, why pay 2x or 3x book for it once the price stops climbing? And the sector simply got crowded. Dozens of copycat treasury vehicles launched through 2025 chasing the same flywheel, and supply of "buy our stock for leveraged bitcoin exposure" eventually exceeded demand for it. Because these companies are equity-financed and leveraged to bitcoin's moves in both directions, the same mechanism that made them outperform bitcoin on the way up is now making them underperform it on the way down — Metaplanet, SharpLink and Semler are all down 74-90% from their highs, multiples of bitcoin's own decline. In July 2026, treasury companies flipped from net buyers of bitcoin to net sellers for the first time since the model existed, using proceeds to pay down debt and rebuild cash rather than add exposure at a valuation that no longer rewards it.

Restructuring is the more accurate word, but it will look like dying for the weakest names. The companies most exposed — small, altcoin-linked, or newly launched treasuries with no operating business behind the balance sheet — are the ones most likely to merge, delist, or wind down over the next year, because they have no way to fund operations once the premium that justified their existence disappears. The survivors are splitting into two camps. One is disciplined, cash-generating majors-only treasuries: Strategy, sitting on roughly 21 months of reserves funded by its dividend business, has simply stopped issuing stock and is waiting for premiums to return rather than diluting shareholders to keep buying. The other is companies with real hard assets behind the ticker — miners with power contracts and data-center infrastructure — pivoting that infrastructure toward a trade that actually still commands a premium in 2026.

Because AI infrastructure, not bitcoin accumulation, is this year's trade that investors will pay up for. Firms like MARA and Bitdeer already have the two assets an AI data center needs most: contracted power and physical land. Bitdeer's Norway campus lease and similar deals let bitcoin miners redirect BTC-sale proceeds and fresh capital into GPU hosting and compute revenue instead of buying more bitcoin at a mNAV that punishes them for it. It's the same capital that funded the 2025 treasury boom, rotating toward the sector currently rewarded with a premium instead of a discount. Vetle Lunde at K33 and others tracking the space describe this less as crypto companies abandoning bitcoin and more as treasuries repricing what kind of balance sheet Wall Street will pay extra for right now.

The bull case is straightforward: a sustained bitcoin rally back toward or past October's highs would re-inflate premiums for the strongest, most liquid names and restart accretive issuance, and AI-pivot miners would benefit twice, from bitcoin's recovery and a genuine data-center re-rating already underway. The bear case is that bitcoin chops sideways or falls further, keeping most mNAVs pinned at or below 1 indefinitely, that dilutive issuance keeps eroding per-share bitcoin exposure at the weaker names, and that AI pivots arrive too late to compete against pure-play data-center operators the market already trusts. The real stress test isn't imminent — it's the first wave of DAT convertible-debt maturities due in 2027 and 2028, when refinancing costs meet balance sheets that never got the chance to rebuild through issuance. Between now and then, watch Q3 earnings from Strategy, Metaplanet, MARA and Bitdeer in October and November: they'll show whether any large treasury resumes net bitcoin buying, or whether AI capex has quietly become the actual plan.