Nine days until the list. Nineteen more until it takes effect at the close.
Those two numbers are the entire schedule of the strangest corporate-finance question of 2026: on 11 November, MSCI publishes the results of its November index review, including which companies fail its new screen for “non-operating” businesses. On 30 November, at the close, any deletion takes effect. And somewhere in between, a software company with a $64 billion market capitalisation and 848,517 bitcoin on its balance sheet has to argue that it is not an investment fund.
I am not being cute. MSCI’s rule, adopted on 15 October after a consultation that ran from 3 August to 30 September, does not name crypto anywhere. The May 2026 simulation that first applied the framework flagged three names: Strategy, Japan’s Metaplanet, and Yellow Cake — a UK-listed company that holds physical uranium. The uranium example is the tell. The screen is not about what you hold. It is about whether holding is the business.
The Rule, Written as Code
Most coverage of this story stops at “MSCI may remove Strategy”, which is like describing a trial by the verdict. The interesting object is the test. Here is the shape of it, compressed from MSCI’s methodology into the only form I can check against a set of accounts:
operating = has_operating_assets(issuer) # step 1: is there a business?
if not operating:
fails = count_failed_ratios(issuer, FIVE) # step 2: five financial ratios
if fails >= 4:
if not in_index(issuer):
return INELIGIBLE # new entrant: one bad period is enough
if failed_in_two_consecutive_periods(issuer):
return INELIGIBLE # existing member: two periods, or an interim at MSCI's discretion
return ELIGIBLE
The five ratios are not exotic. They ask about the size of operating assets, about operating expenses, about operating cash flow, about how much of the earnings line comes from changes in the fair value of what you hold, and about how dependent the company is on raising outside money. Four hits out of five and the issuer is out.
Two details in that code matter more than the headline. First, the two-period buffer for existing index members: MSCI is trying to keep turnover down, so a company that fails once is not immediately removed. Second, the new-entrant asymmetry: a company outside the index that fails a single period cannot get in at all. Read those together and you get the real policy: this rule is designed to stop the next Strategy from being added, and to give the current one a long, deliberate exit.

Why a Fund Is Not Allowed In
There is a reason MSCI’s instinct here is not arbitrary. Equity benchmarks have excluded collective investment vehicles for as long as they have existed, for a simple mechanical reason: if you put a fund inside an index, every index tracker owns the fund, and then the index is partly a claim on itself. It is a circularity that makes the benchmark useless as a measure of anything.
The January 2026 episode is the useful precedent. MSCI looked at digital-asset treasury companies, decided not to exclude them after pushback from Strategy and others, and instead froze increases to their index weights while it opened a broader consultation on non-operating companies generally. That is a regulator’s move, executed by a private company: when you cannot win the specific argument, widen the definition until the specific case is a subset.
So the October methodology is the widened definition, and Strategy is now a subset of it. MSCI’s framing is that the target is companies whose balance sheet looks more like a portfolio than an operating business. Strategy disputes the label, arguing that its bitcoin activity is an operating segment and that the framework repackages the earlier crypto-specific proposal under a neutral name. Both of those can be true at once. Rulebooks usually work that way.
Applying the Test to the Actual Numbers
Here is where the argument stops being philosophical. Strategy reported third-quarter software revenue of $125 million against 848,517 bitcoin held as of 25 October. In 2026, through 25 October, it raised $21.23 billion of capital and spent $14.33 billion on bitcoin. Its quarterly preferred dividend rose to $420 million from $140 million a year earlier, and cumulative preferred dividends paid reached $1.48 billion, against a pool of $5.07 billion of dollar assets that the company says covers dividend and interest obligations for 3.4 years.
Those numbers describe a financing machine, not a software business with a hobby. And the machine has three moving parts that have to keep turning in order: the share price has to stay above the value of the coins per share, so that issuing is accretive; the preferred market has to keep accepting new paper so the dividend can be serviced; and the bitcoin has to stay boring, so that neither of the first two gets interrupted. The entire structure depends on the equity market’s willingness to keep buying a wrapper. That is what MSCI is deciding whether to underwrite.

What Deletion Does — and the Thing It Doesn’t
The consensus reading is that index removal forces selling. It doesn’t, not directly, and the distinction is where most of the analysis goes wrong.
Index funds that track the affected benchmarks own MSTR shares. They do not own bitcoin, and they have no contractual right to exchange those shares for a slice of the treasury. If they sell, they sell stock. Strategy’s financing agreements do not force a bitcoin sale if the price falls below its average purchase cost, and the company has said as much. So the direct effect on the bitcoin market from a 30 November deletion is approximately zero.
The indirect effect is not zero, and it is worth putting a number on. Funds tracking the relevant indexes hold roughly 13 million MSTR shares, about 3.1% of basic shares outstanding. Reports have put that block near $1.8 billion, or about 60% of the stock’s 30-day average daily trading volume. That is the size of a potential forced sale in one name — not a forecast, and not a bitcoin trade, but a very large order arriving in a stock whose entire purpose is to be a liquid wrapper.
From there the chain runs the other way:
- Index trackers sell, and the marginal buyer of MSTR disappears for a week.
- The premium to net asset value compresses.
- Each new share sold buys less bitcoin than it did before, so issuance becomes dilutive instead of accretive.
- The company raises less, and buys less.
Step four is the one that touches the bitcoin market, and it is a slowdown, not an event. The honest framing of 11 November is not “will Strategy be forced to sell” — it is “can Strategy keep buying at the pace the market has priced in.”
| Company | What it holds | Position in the October list |
|---|---|---|
| Strategy | 848,517 BTC | Flagged for potential deletion |
| Metaplanet | Bitcoin treasury | Flagged for potential deletion |
| Upexi | Solana treasury | Flagged for potential deletion |
| Yellow Cake | Physical uranium | Flagged for potential deletion |
| SKEL | Investment holding, Iceland | Flagged for potential deletion |
Of the 22 companies on MSCI’s potential list, 14 are crypto-related, and they account for about 94.5% of the list’s free-float market capitalisation — almost all of it Strategy. A further group, including SharpLink and Solana Company, sits on the observation list. But five names were flagged for actual removal and two of them have nothing to do with crypto. That is the part worth respecting about the rule: it is not a crypto rule that happens to catch a uranium fund. It is a fund rule that happens to catch crypto.
The Calendar, Which Is the Actual Story
MSCI declines to exclude digital-asset treasury companies, freezes weight increases, opens a broader consultation
A simulation under the new framework flags Strategy, Metaplanet and Yellow Cake
Consultation opens
Consultation closes
MSCI adopts the non-operating company screen; deletions to take effect at the 30 November close
Final list of affected constituents published
Index changes take effect at the close
Look at the spacing. MSCI took ten months to move from “not now” to “here is a rule”, and then gave the affected companies nineteen days between the list and the effective date. That is not how a regulator behaves and it is not how a market behaves. It is how a company behaves when it wants the outcome without wanting the fight — and it is the most consequential thing an index provider has done to a bitcoin-adjacent business since the ETFs launched.

What I Think Happens
I think Strategy stays on the list and gets deleted, because the alternative requires MSCI to argue that a company earning $125 million a year from software is operationally indistinguishable from a fund with a small side business, and the ratios are not that close. I also think the market has already traded most of it — the stock fell 7.86% to $130.49 on Friday, on top of a bad week for everything — and that the actual damage will show up not in bitcoin’s price but in the pace of Strategy’s purchases through the first half of 2027.
The part I keep coming back to is smaller. Saylor has said the company can raise as much as $30 billion more in preferred stock. That is a capacity estimate, not committed funding, and the difference between those two things is the entire question MSCI is asking. A company that funds itself by selling paper is a company whose cost of capital is set by people who are not its customers, not its shareholders, and not its board — and the index provider is simply the first of those people to write the test down.
So: if MSCI can decide whether a bitcoin treasury is a company, what exactly is the thing you own when you buy the stock — a share of a business, or a claim on a financing loop that needs the index to keep working?
(The End)


