IBIT vs MSTR: Two Bitcoin-Per-Share Numbers That Do Not Mean the Same Thing

Why One Is Verifiable

A spot bitcoin trust has exactly one mechanism that removes bitcoin without removing shares: the sponsor's fee. Creations and redemptions move bitcoin and shares together in fixed 40,000-share baskets, leaving the ratio unchanged. The fee accrues daily at an annualised 0.25% of net asset value.

Therefore IBIT's bitcoin-per-share can only fall between filings, and only at a known maximum rate. A filed figure is an upper bound on today's figure. The error is bounded, one-directional, and computable at roughly 0.02% a month.

That is checkable against the filings. Over the six months from December 31, 2025 to June 30, 2026, IBIT's NAV per share fell from $49.61 to $33.48 — a decline of 32.5136%. Bitcoin over the same period fell from $87,463.03 to $59,101.49 (CF Benchmarks), a decline of 32.4269%. The shares slipped 0.1283% against the asset. Expected drag from a 0.25% annual fee over 181 days: 0.1239%.

Unexplained residual: 0.0044 percentage points.

A ratio you can audit to four decimal places is a different object from one you cannot. What the fee does to that ratio over a holding period is the subject of our fee-drag calculator.

What “Verifiable” Actually Bought, This Week

It would be dishonest to present the trust side as effortless, so here is a live example from the week this article was written.

On August 20, 2026, BlackRock's own IBIT product page published two figures that cannot both be right. Its downloadable holdings data gave 762,287.03650 bitcoin against 1,333,840,000 shares — a ratio of 0.000571498. The same page's Basket Bitcoin Amount of 22.65 gave 0.000566250.

The first figure is 0.875% above the June 30 filed maximum. For a fee-only trust that is impossible: the ratio cannot rise. The second sits correctly below it.

The cause is visible in the same page's own numbers — holdings market value exceeded net assets by about 0.75% on the same as-of date, and BlackRock states that holdings data reflects its investment book of record and may differ from the accounting book used to determine net assets. The holdings feed was running roughly 309 baskets ahead of the shares feed.

The point is not that the issuer's feed was wrong. It is that being wrong was detectable. The invariant caught it in one line of arithmetic, without inside information, without waiting for a correction, and without any judgment call. That is the entire practical value of a structural guarantee: not that errors never occur, but that they cannot hide.

No equivalent test exists for Strategy. If its published bitcoin-per-share were off by 0.875%, nothing in the arithmetic would say so.

Why It Fails for Strategy

Strategy's ratio moves in both directions, for four independent reasons, on three different disclosure schedules:

• Bitcoin can rise. 174,895 bitcoin acquired in the six months to June 30, 2026.
• Bitcoin can fall. Approximately 1,395 sold in the same period, and more since — 1,638 in the week to August 2 and 1,690 in the week to August 9, with proceeds funding preferred dividends and preferred repurchases.
• Shares can rise. 58,456,000 Class A shares issued in the half for net proceeds of $8,235,553,000, plus 1,085,000 under incentive plans.
• Shares can fall. A common repurchase programme authorized June 29, 2026, with $1.0 billion of capacity remaining at August 16.

The cadences do not align. Bitcoin holdings are disclosed weekly in 8-Ks. Total shares outstanding are disclosed quarterly. The weekly filings report shares sold under the at-the-market programme, not shares outstanding — so the denominator cannot be reconstructed exactly at all.

The share count rose 3.40% in the 24 days between the balance sheet date and the cover date of the same document.

Three defensible computations:

The third row adds the three disclosed weekly ATM sales after the cover date — 3,011,361 (week to August 2), 6,585,682 (week to August 9) and 3,458,866 (week to August 16) — to reach 397,281,660. That is a lower bound, not a share count — it omits option exercises, employee purchases, vesting, the partially-captured cover week, and settlement timing.

An analyst using only primary sources can produce figures differing by more than seven percent depending on which filing dates they mix, and none of them is today's number. The most defensible is the same-date figure from the 10-Q, already seven weeks stale when this was written.

Strategy is not doing anything wrong. Weekly treasury 8-Ks are far more than the rules require. The problem is structural: a company that raises and deploys capital continuously cannot have a stable per-share asset ratio, and no disclosure regime can build one current metric from two moving parts measured on different days.

For a trust, bitcoin-per-share is a measurement. For Strategy, it is a snapshot with a two-directional, unbounded error term. Never put them in the same column without dates attached to each.

One further precision note, in Strategy's favour on cadence and against it on decimals: IBIT reports bitcoin to eight places. Strategy's 10-Q reports the “approximate number of bitcoins held” as 846,000, rounded; the weekly 8-Ks are precise to the whole coin. The quarterly financial statements are the least precise source for the numerator.

Strategy's Own Metric Is a Third Calculation

Strategy reports “Bitcoin Per Share” as a KPI, stating in its Q2 results that it grew holdings 11% to 846,000 bitcoin and grew Bitcoin Per Share by 5%.

On basic common shares from the balance sheet, December 31, 2025 to June 30, 2026:

• Bitcoin: 672,500 → 846,000 (+25.8%)
• Common shares: 312,062,000 → 371,603,000 (+19.1%)
• Ratio: 0.00215502 → 0.00227662 (+5.64%)

Consistent with the company's “+5%.” But Strategy's published KPI uses Assumed Diluted Shares Outstanding — basic shares plus assumed conversion of convertible notes, convertible preferred, options, RSUs and PSUs. On ADSO, the 10-Q's KPI table gives 0.00210824 (210,824 sats, on 401,283,000 assumed diluted shares) at June 30, 2026. By mid-August the figure on Strategy's own dashboard — an issuer disclosure channel, not an SEC filing — stood at 0.00196684, consistent with the ATM share sales disclosed in the weekly 8-Ks since.

Three different denominators, three different numbers, all correct for their own definition. Always name the denominator.

Strategy is explicit about what its metric is not: BPS excludes debt and preferred senior claims, is not a financial performance, valuation or liquidity measure, and ownership of the company's securities does not represent an ownership interest in, or a redemption right to, the bitcoin the company holds. It also states that debt- or preferred-financed purchases can raise BPS while increasing senior claims the metric does not reflect.

mNAV, and Why There Is No Figure Here

mNAV compares the market value of a company's equity to the net asset value of its bitcoin. Above 1.0 the market pays more than the underlying bitcoin; below 1.0, less. It matters because a company above 1.0 can issue shares and buy bitcoin in a way that raises bitcoin-per-share for existing holders, while one below 1.0 dilutes by doing the same thing.

This site cannot publish a current mNAV under its own rules, because the numerator is a live market price and market prices are not in filings. Rather than source a share price to an aggregator, here is what the balance sheet supports directly.

At June 30, 2026:

On the liquidation-preference basis, $80.37 per common share against 371,603,000 shares. That is book value, not market value — though Strategy carries bitcoin at fair value, so book sits closer to economic value here than for most companies.

To get mNAV, divide market capitalisation by the bitcoin fair value of $49,672,080k, or by the residual after debt and preferred. Do that arithmetic with a price you sourced yourself, on a date you recorded.

Two cautions. The denominator convention is contested: mNAV against gross bitcoin ignores $6.71 billion of debt and $15.46 billion of preferred liquidation preference standing ahead of the common, and published figures rarely say which convention they use — Strategy's own dashboard now defines mNAV net of debt and preferred claims, while much circulating commentary uses gross bitcoin. And Strategy changed the calculation method for its BTC Yield and related KPIs effective January 1, 2026, stating that prior-period figures are not directly comparable — so historical charts of these metrics spanning definition changes are comparing different quantities.

The Capital Structure

IBIT has no debt, no preferred, no obligations and no counterparties beyond its custodian. Strategy has all four.

Convertible Notes, June 30, 2026

Net carrying value $6,670,114k; aggregate fair value $6,259,499k. The notes were trading below carrying value in aggregate — the 2029s at $1,296,855k against $1,492,981k carrying, the 2030Bs at $1,776,440k against $1,991,618k.

None was convertible during the half and no conversion requests were received. On May 19, 2026, Strategy repurchased and cancelled $1.50 billion principal of the 2029 notes for $1.38 billion, recognising a $113.9 million gain on extinguishment.

Each series carries a holder put — a date on which holders can demand cash repurchase at par, stated per series in the table above from the 10-Q's own disclosure. Those dates, not the stated maturities, are the operative deadlines. The nearest is September 15, 2027.

Preferred Stock

Five series at June 30, 2026: STRF (10.00%), STRC (variable), STRE (10.00%), STRK (8.00%) and STRD (10.00%). Classified in mezzanine equity because redemption events exist outside the company's sole control.

Preferred issuance raised $7,534,966k gross in the half. Preferred dividends declared were $758,224k, against $58,142k in the first half of 2025. On July 31, 2026, Strategy announced it would hold the STRC rate at 12.00% annually for semi-monthly periods from August 16, 2026. (8-K, August 3, 2026) What these instruments are — and how Bitcoin treasury preferred stock differs from a spot ETF — is covered in our digital credit explainer.

An Operating Company That Holds Bitcoin

The software business, six months to June 30, 2026: revenues $246,668k, gross profit $164,904k, remaining performance obligations $546.9 million, of which $320.1 million is expected within twelve months.

Net cash provided by operating activities for the same six months: $9,850k.

Preferred dividends paid in the same period: $629,175k.

The operating business covers roughly 1.6% of the preferred dividend bill. The remainder comes from capital raising — $8.24 billion of common and $7.53 billion of preferred issuance in the half — and, since late July 2026, from selling bitcoin. In the week to August 16, Strategy raised $333.7 million from stock sales and allocated $52.4 million to STRC dividends, $132.2 million to STRC repurchases and $149.1 million to its USD reserve. None went to bitcoin. (8-K, August 17, 2026)

Owning MSTR means owning a software business, a bitcoin position, a leveraged capital structure with fixed dollar obligations, and an ongoing dependence on capital-markets access to service them. Owning IBIT means owning bitcoin less 0.25% a year. These are different risks, not different amounts of the same risk.

There is also surface area unrelated to bitcoin: a board, executives, auditors, litigation exposure, tax positions, key-person risk. Strategy's principal accounting officer changed effective June 30, 2026, when Jeanine Montgomery retired and CFO Andrew Kang was designated to the role. (8-K, July 6, 2026) A trust has none of this.

What a Drawdown Does to Each

The first half of 2026 is a filed example. Bitcoin fell roughly 32% on every index used by the entities here — $87,463.03 to $59,101.49 on CF Benchmarks (IBIT's index), and to $58,714 at June 30 on Coinbase, Strategy's stated principal market. The index always matters: Grayscale's funds put the same day at $58,745.18 on CoinDesk. Same asset, same date, three filed prices.

IBIT. Net assets fell from $67,401,155,244 to $43,386,012,125; NAV per share from $49.61 to $33.48. The per-share decline tracked bitcoin to within 0.13%, essentially all fee. Nothing else happened, because there is nothing else in the structure. A holder's bitcoin-per-share was unaffected by the price move.

Strategy. The senior claims do not shrink when bitcoin falls. $6.71 billion of debt principal and $15.46 billion of preferred liquidation preference are fixed dollar amounts; the equity absorbs the decline first. Preferred dividends remain payable from an operating business generating $9.85 million of operating cash flow per half. That is the mechanism by which bitcoin sales began — not a change of view about bitcoin, but an obligation falling due.

This is stated mechanically on purpose. Nothing here predicts how MSTR's share price will behave, and Strategy itself states that its trading price is influenced by factors beyond holdings and share count and can deviate significantly from the fair market value of its bitcoin.

How to Read Either Number

For IBIT: take the most recent filed figure, treat it as an upper bound, and apply at most 0.25% a year of decay. If a published figure exceeds the filed maximum, the published figure is wrong — not the filing. The trust's fact sheet carries the site's live ratio with its provenance.

For Strategy: take the numerator and denominator from the same filing, or accept a known mismatch. State both dates. Name the denominator — basic, ADSO, or something else. Never compare a Strategy ratio to a trust ratio without both.

This site publishes verified ratios for spot bitcoin trusts because the structure permits verification — how that category came to exist is our history of the road to the spot Bitcoin ETF. It does not publish a live MSTR ratio, because the structure does not permit it — and a number presented with the site's usual provenance treatment would be claiming a guarantee that cannot be given. That is not a gap in coverage. It is the same standard applied honestly to a case where it produces a different answer.

FAQ