Strategy's 'Permanent Capital' Meets the mNAV Floor: What JPMorgan's Index Math Actually Prices

Hook

Here is the number that never makes the slide deck: 1.04.

Not 640,000 coins. Not the $42 billion multi-year raise. Not the 0% coupon that turned Michael Saylor into a folk hero on crypto Twitter. The number that decides whether Strategy (NASDAQ: MSTR) is a bitcoin-accumulation machine or a slow-motion deleveraging event is the gap between what the market pays for a share and what that share's bitcoin is actually worth. Traders call it mNAV — modified net asset value. In early 2024 it traded north of 3x. By the fourth quarter of 2025, on the prints I was pulling, it was pressing against 1.0.

I have seen that shape before, and I have watched what happens when it breaks. In the summer of 2020 I was live-blogging transaction hashes out of Uniswap and Aave Discords while a niche lending protocol paid four-figure APYs and the entire timeline called it product-market fit. When the emissions stopped, total value locked did not taper. It fell off a shelf in about nine days. DeFi was not a bug; it was a feature of chaos — but chaos has rules, and one of them is that a subsidy always looks like a business model right up until the terms reset.

JPMorgan just published the arithmetic on what a reset would cost.

Context: how a software company became a bitcoin bank that is not a bank

Back up to February 2025. MicroStrategy — a legacy enterprise-analytics vendor with a mid-single-digit-millions software business — renamed itself Strategy. The rebrand was not cosmetic. It was an admission that the balance sheet had become the product.

The mechanics are worth restating precisely, because most of the discourse around this company is either worship or dismissal, and neither survives contact with a prospectus.

Strategy funds bitcoin purchases through three channels. First, at-the-market equity issuance — selling new shares into the open market when the stock trades above the value of the bitcoin backing it. Second, convertible senior notes, largely unsecured, with staggered maturities running from 2027 out past 2032, several tranches carrying coupons at or near zero. Third, perpetual preferred stock, sold to institutional income buyers at fixed or variable dividend rates that sit in the high single digits to low double digits.

The company's public framing for all of this is "permanent capital." The argument runs like this: because the converts are unsecured, because they have long durations, because they can be settled in shares rather than cash, and because there are no depositors who can queue at a teller window demanding their money back, this capital does not have the run-risk profile of a bank. A bank funds long-dated, illiquid assets with demand deposits. Strategy funds a volatile asset with term liabilities. That is a genuinely different structure.

JPMorgan is the natural foil in this comparison, and not only because it is the largest US bank. It is also the house whose digital-asset research desk — the team run by Nikolaos Panigirtzoglou — has spent 2025 publishing the most detailed public analysis of what happens to Strategy if the index providers change their rules. That analysis is the thing worth engaging with, because it is not an opinion about bitcoin. It is an opinion about plumbing.

And plumbing is where the argument gets interesting.

Core: the flywheel, the floor, and the rule book

Start with what makes the machine work, because the machine is elegant and I want to be fair to it before I take it apart.

Accretive issuance is the engine. If a share trades at $300 while the bitcoin attributable to that share is worth $200, then every new share sold buys $300 of bitcoin and adds $200 of backing to the existing share count. The company can therefore raise its bitcoin-per-share without bitcoin going up at all. Strategy formalized this into a metric it calls "BTC Yield" — essentially the percentage growth in bitcoin-per-share from issuance.

That is the flywheel. Premium generates issuance. Issuance generates bitcoin-per-share growth. Bitcoin-per-share growth justifies the premium. Around and around.

Note the circularity, because it is the whole story. The premium is not a reward for the bitcoin. The premium is a reward for the ability to keep buying bitcoin. When the premium is 3x, the machine is self-reinforcing and every quarter produces a spectacular BTC Yield number. When the premium compresses toward 1x, the arithmetic inverts. Issuing at 1.0x is not accretive to anyone except the underwriters. Issuing below 1.0x actively destroys bitcoin-per-share for existing holders.

That is the floor. mNAV 1.0 is not a support level in the charting sense. It is the point at which the company's primary funding mechanism stops working. Above it, Strategy is a compounding machine. Below it, Strategy is a closed-end fund with a debt stack.

So the entire thesis reduces to a single question: what keeps mNAV above 1.0?

Part of the answer is narrative — retail and momentum buyers who want bitcoin exposure in a brokerage account, an IRA, or a 401(k). Part of the answer is structural: passive index flows. And that second part is exactly where JPMorgan has been aiming.

Here is the plumbing. MSCI, FTSE Russell, and S&P Dow Jones all maintain index eligibility rules. Several of those providers have consulted on, or moved toward, criteria that would exclude or down-weight companies whose balance sheets are dominated by digital assets. MSCI ran a consultation through late 2025 with a decision point in mid-January 2026. FTSE Russell has already tightened criteria in some product families.

Why does that matter for a company with no operational dependence on index membership? Because of who owns MSTR shares. A meaningful slice of the register is held by passive vehicles — broad-market index funds, thematic ETFs benchmarking to standard indices, and institutional mandates whose investment policy statements prohibit single-name concentration outside an approved benchmark. If MSTR leaves the benchmark, those holders do not get to make a judgment call about bitcoin. Their mandates force them to sell. The selling is mechanical, price-insensitive, and calendar-driven.

JPMorgan's framing is that this is a forced-seller event measured in billions of dollars of notional, landing on a stock whose marginal buyer has historically been the retail momentum crowd. That is a bad combination. Forced sellers meet price-sensitive buyers, and the clearing price is wherever the buyers decide it is.

Now the part the bulls get right. None of this is a solvency event. Strategy's converts are unsecured and long-dated, and the company can settle them in shares. There is no margin call, no counterparty who can accelerate, no depositor queue. A bank facing a deposit run must sell assets into a falling market to meet withdrawals — that is the classic fire-sale death spiral. Strategy does not have that specific vulnerability. The structure genuinely is more durable than a bank's.

But durability is not the same as permanence, and this is where I want to plant a flag.

"Permanent" is doing an enormous amount of work in that phrase. The equity capital is permanent — nobody can redeem a common share. The convertible debt is not permanent; it matures, and refinancing a maturing convert in a market where mNAV has collapsed means either paying cash the company would rather spend on bitcoin, or issuing shares at a bad price. The preferred stock is perpetual in legal form but it is not free. Those dividend obligations are cash. On the order of $800 million a year across the preferred series, by my read of the filings, and that number grows every time another series is sold.

So the honest description of the capital structure is this: the equity is permanent, the debt is term-funded, and the preferreds are a permanent cash obligation layered on top of a volatile asset. That is not the same thing as permanent capital. It is a capital structure with a fixed charge attached.

In the void, we found our value in the noise — and the noise here is a dividend schedule nobody is modeling.

Which brings me to the parallel I cannot unsee.

I have spent thirteen years watching crypto protocols invent financial structures that looked inventive at the top of a cycle and looked obvious at the bottom. The most reliable pattern is the subsidy that gets mistaken for a moat.

Take liquidity mining. A DEX or a lending market pays outsized emissions to attract deposits. TVL goes vertical. The dashboard shows growth, the governance forum shows confidence, and the token price holds because emissions are being recycled into the pool. Then the emissions taper, or the token price falls, or a fork offers a better rate. The TVL leaves. It almost always leaves within weeks, because it was never loyalty — it was yield. The APY was the product, and the APY was the project paying itself.

The same pattern shows up in the rollup stack. Post-Dencun blob space was priced so cheaply that posting data to Ethereum L1 became nearly free, and rollup fees collapsed to fractions of a cent. Everyone declared the scaling problem solved. What actually happened is that rollups have been running on a subsidized input cost. Blob demand is compounding faster than blob supply, and the fee market for blob space is a fee market — it will clear at whatever price clears it. When that happens, the subsidy ends and every rollup that built its unit economics on near-zero data availability costs will have to re-price. Some of them will double their fees and act surprised when users notice.

Strategy's premium is the same shape. It is not a fundamental property of bitcoin. It is a temporary price the market pays for access, and access can be repriced by a rule change, a sentiment shift, or simply time.

The subsidy framing is not just rhetorical. Look at how the company's own disclosures evolved. BTC Yield is a metric that only exists because issuance is accretive. When issuance stops being accretive, BTC Yield goes negative, and the metric stops being reported in the same font. I have watched this exact move in DeFi dashboards — when emissions dry up, the APR widget quietly disappears and is replaced by a "real yield" chart.

Metric substitution is the tell. Whenever a treasury company starts introducing new units of account, ask what the old units were saying.

There is a second, more important piece of the puzzle, and it has almost nothing to do with Michael Saylor.

I want to talk about the other permanent capital, the one that does not appear in any 8-K.

Go find the most reliable demand for dollar-denominated crypto assets on earth, and it is not in Greenwich or Miami. It is in Lagos. In Buenos Aires. In Istanbul. In Karachi. It is a market maker in Kano moving naira into USDT because the naira's purchasing power has been shredded and the parallel-market spread makes the official rate a fiction. It is a freelancer in Argentina holding stablecoins because a 200% annual inflation print turns a peso savings account into a countdown timer. It is a small importer in Turkey settling invoices in USDC because the correspondent-banking rail costs more in fees and delays than the entire stablecoin transaction.

Chainalysis has ranked Nigeria and India at or near the top of its grassroots adoption index for consecutive years. That ranking is not driven by ideology. Nobody in those markets is holding stablecoins because they read a governance forum. They are holding them because the alternative is watching their savings evaporate at a rate they can measure monthly.

Here is the part that matters for Strategy specifically: this demand is structurally different from the demand the company is cultivating. Lagos demand is usage-driven, small-ticket, and reflexive to monetary conditions. It does not care about mNAV. It does not care about index inclusion. It will not sell because MSCI changed a rule, and it will not buy more because a convertible note was oversubscribed. It just keeps showing up, every month, because the local currency keeps failing.

Strategy's demand is capital-markets-driven. It is large, it is concentrated, and it is highly sensitive to a handful of institutional decisions. Those are two entirely different asset bases, and the crypto market has a habit of confusing the loud one for the real one.

Let me put my audit hat on, because this is where thirteen years of reading filings is actually useful.

If I were handed Strategy's disclosure package tomorrow and asked to price the real risk, I would not start with the bitcoin count. I would start with four lines.

Line one: the maturity ladder. Which converts come due in the next twenty-four months, and what is the cash cost of settling them if the share price is below the conversion price. A convert that settles in shares at a depressed price is dilution; a convert that settles in cash is a bitcoin sale.

Line two: preferred dividend coverage. Where does the cash come from? The software business generates revenue in the low hundreds of millions annually — nowhere near the preferred obligations. So the cash comes from either operating cash plus reserves, or from new issuance. If it comes from new issuance, the preferreds are being paid with the proceeds of more preferreds, which is a structure that works exactly until it does not.

Line three: the mNAV print, weekly. Not the chart. The actual ratio, computed with the fully diluted share count, not the basic count. This is the single number that predicts the direction of everything else.

Line four: the index exposure. Which mandates are legally compelled to hold this, and what is the dollar value of the overhang if the rule changes. JPMorgan has done the work here and the work is worth reading even if you disagree with the conclusion.

Based on my audit experience with treasury-heavy balance sheets — and I have torn apart a few distressed ones — the failure mode is never the headline asset. The asset can fall 70% and the company survives. What kills these structures is a fixed cash obligation meeting a closed funding window. The dividend is the fixed obligation. The closed funding window is mNAV below 1.0. Those two things do not need to happen simultaneously for long, but if they do, the timeline compresses fast.

Contrarian: everyone is arguing about the wrong balance sheet

The consensus bear case on Strategy is a balance-sheet argument: too much leverage, too much concentration, a bitcoin drawdown triggers a cascade. The consensus bull case is the mirror image: unsecured converts, long durations, no forced selling, diamond hands at scale.

Both sides are arguing about the same set of numbers, and both sides are missing the actual variable.

The real variable is not leverage. It is the buyer base.

Strategy's entire model depends on a specific, fragile demographic: institutional allocators who want bitcoin exposure but are not permitted to buy bitcoin directly. That is a regulatory-arbitrage trade, and regulatory arbitrage has a shelf life. Spot bitcoin ETFs already solved most of it. Every quarter that ETFs get cheaper, more liquid, and more widely approved inside retirement account platforms, the reason to own a premium-wrapped bitcoin proxy gets weaker.

That is the blind spot. Not "can Strategy survive a bear market" — it obviously can, structurally. The question is what happens when the wrapper stops being necessary. Because the moment an allocator can buy bitcoin itself at 20 basis points with same-day settlement, paying 1.5x net asset value for a leveraged, dividend-bearing, index-inclusion-dependent proxy stops being a clever trade and starts being a governance question.

The bulls will say the premium reflects the leverage and the optionality. Fair. But leverage cuts both ways, and optionality has an expiration date printed on it.

The second blind spot is the one that stings. Everyone treats JPMorgan as the adversary here. Read the report again. JPMorgan is not arguing that bitcoin is bad. It is arguing that index rules are rules, and rules do not negotiate. That is not a bear thesis on crypto. That is a bear thesis on a specific capital structure that depends on rules not changing.

The story isn't in the numbers — it's in the pulse. And the pulse of the institutional buyer right now is telling you that bitcoin exposure is becoming a checkbox, not a conviction trade. Checkbox allocations do not pay premiums.

Takeaway

Watch four things over the next two quarters, in this order.

The mNAV print, weekly, computed on fully diluted shares. If it holds above roughly 1.5x, the flywheel keeps turning and the bears are early. If it prints below 1.1x for consecutive weeks, the machine has stopped buying and started defending.

The January index decision window, and the dollar value of the mandates that would be forced to sell if the rules change.

Preferred dividend coverage in the next quarterly filing — specifically whether cash obligations are being met from operations or from fresh issuance.

And the number nobody puts on the slide: how much of the buying is coming from allocators who cannot buy bitcoin directly, versus how much is coming from people who simply like the ticker. One of those groups is permanent. The other one is a subsidy, and subsidies have terms.