How a Bitcoin Treasury Company Creates Yield

There is a claim in the Bitcoin treasury world that sounds like it breaks a rule of finance: a company can increase the amount of Bitcoin sitting behind each of its shares while the price of Bitcoin is falling. Not through trading or derivatives, just share issuance and arithmetic.

It sounds like alchemy. It isn't. It falls straight out of a few basic definitions, and once you see the calculation you can tell which companies are actually doing it and which are just hoping. Let me walk through it, and then show you a real company that wrote the whole mechanism into its financing.

A flywheel doesn't make energy. It stores what the engine produces and gives it back as steady momentum. A Bitcoin treasury company runs on the same principle: it converts the premium in its share price into Bitcoin per share and keeps that momentum, for as long as the engine keeps turning. Let the premium fade and the wheel still spins, but it gathers nothing more.

Flywheel of the Boulton and Watt steam engine. Image credit Newtown Grafitti, shared under license CC BY 2.0.

The building blocks

A Bitcoin treasury company is a listed company whose main asset is Bitcoin. Five numbers describe it:

• M: the company's market cap
• B: the Bitcoin it holds
• P: the Bitcoin spot price
• S: the number of shares
• BPS: Bitcoin per share, which is simply B divided by S

The number everyone watches is the multiple of net asset value, or mNAV:

The denominator, B × P, is the dollar value of the company's Bitcoin. So mNAV is just the market cap divided by the Bitcoin it owns. When mNAV = 1, the company is worth exactly its Bitcoin. When mNAV = 1.5, the market is paying $1.50 for every $1.00 of Bitcoin on the balance sheet. That extra half is the premium.

The engine, in algebra

Now the interesting question. What happens to each shareholder's slice of Bitcoin when the company issues new stock and spends the proceeds on more Bitcoin?

Suppose it issues ΔS new shares at the prevailing price per share, M/S. That raises cash:

which buys ΔB = ΔC/P of new Bitcoin. Bitcoin per share rises if the new figure beats the old one:

Cross-multiply and the S's and B's cancel down to a very clean condition:

In words: each new share must buy more Bitcoin than the average share already represents. Now substitute what ΔB actually is. Since ΔB/ΔS = (M/S)/P = M / (P × S), the condition becomes M / (P × S) > B / S, and the S cancels to leave:

That is the whole secret in one line. Issuing shares to buy Bitcoin makes existing holders richer in Bitcoin if and only if the company trades above the value of its Bitcoin. Nothing about the price direction appears in that condition.

We can go one step further and measure how much richer. Let f = ΔS/S be the fraction of new shares issued. Then ΔC = f × M and ΔB = f × M/P, and the growth in Bitcoin per share works out to:

This little formula is the engine of the whole model. Notice what is not in it: the Bitcoin price, P. It has cancelled out entirely. The accretion depends only on how much you dilute (f) and how big your premium is (mNAV).

A worked example

Consider a treasury company that this morning looks like this:

• Bitcoin held: 10,000 BTC
• Bitcoin price: $100,000, so the Bitcoin is worth $1bn
• Market cap: $1.5bn, which makes mNAV = 1.5
• Shares: 100 million, so each share costs $15
• Bitcoin per share: 10,000 / 100,000,000 = 0.0001 BTC, or 10,000 sats

It issues 10 million new shares, a 10% increase (so f = 0.10), at the prevailing $15. That raises $150m, which buys 1,500 BTC at $100,000.

After the raise:

• Bitcoin held: 11,500 BTC
• Shares: 110 million
• Bitcoin per share: 11,500 / 110,000,000 = 0.00010455 BTC, or about 10,455 sats

Every share now carries 4.5% more Bitcoin than it did that morning. The company diluted its shareholders by issuing stock, and they came out ahead in Bitcoin terms. Nobody's holding was watered down. It was concentrated.

Check it against the formula: (1 + 0.10 × 1.5) / (1 + 0.10) = 1.15 / 1.10 = 1.045. The same 4.5%. Why does it work? Because the company sold shares at a premium and bought Bitcoin at par. The market valued each share as if it were worth $15; the company turned that $15 into a full $15 of actual Bitcoin. The gap between the two is the premium, and the premium is what flows to existing holders.

Why the price drops out

The cancellation of P is the part worth sitting with, so let's stress it. Run the same company through a brutal bear market. Bitcoin halves to $50,000, but the market still awards it the same 1.5 multiple.

• Bitcoin held: 10,000 BTC, now worth $500m
• Market cap at mNAV 1.5: $750m, so each share is now $7.50
• Issue 10% new shares again: 10 million at $7.50 raises $75m
• Buy $75m of Bitcoin at $50,000, which is… 1,500 BTC

The exact same 1,500 BTC as before. Half the dollars raised, but Bitcoin at half the price, so the coin count is identical. Bitcoin per share climbs by the same 4.5%. The yield a treasury company pays is denominated in Bitcoin, not dollars, and in Bitcoin terms it does not care which way the price is moving.

Case study: Metaplanet

You can watch this exact condition operate in a real company, because one of them wrote it straight into the financing. Metaplanet, the Tokyo-listed treasury company, raises much of its Bitcoin capital through moving-strike warrants, which convert into new common shares. On its 2026 series it attached what its CEO called a first-of-its-kind mNAV clause: the warrants can be exercised only when the shares trade at or above 1.01 times mNAV. In plain terms, new common shares can be created only when doing so is accretive to Bitcoin per share. The company said as much directly, describing the capital as unlocked only when it is accretive to BTC per share. Strategy applies the same discipline less formally, issuing common through its at-the-market programme only when mNAV is above 1, precisely to avoid diluting holders.

Now the striking part. Metaplanet was doing this in March 2026 with Bitcoin trading around $73,000, well below the average cost of $107,000 it was bought for, so the treasury was underwater in dollar terms. It still placed about 107 million new common shares at a 2% premium to the market, with mNAV sitting near 1.1. Because those shares cleared the gate, every one of them left existing holders with more Bitcoin behind their stock, not less, even as the dollar value of the treasury was falling. The yield is measured in Bitcoin, and Bitcoin per share went up while the price went down. That is the worked example above, happening for real, in a drawdown.

And it was not a one-off. Through the sell-off that ran from late 2025, Metaplanet kept posting a positive quarterly Bitcoin yield: roughly 11.9% in the final quarter of 2025 and 2.8% in the first quarter of 2026, on its own figures. Common holders were being paid in Bitcoin in a falling market, for exactly as long as the company could keep issuing above par. This was a real ‘flywheel’ effect: the company stored momentum to smooth out the downturn.

The catch

But the gate cuts both ways, and the same company shows what happens when it closes.

A moving-strike warrant tied to 1.01x mNAV is worthless the moment the premium disappears. As the drawdown deepened, the premium compressed, and by the middle of 2026 Metaplanet's mNAV had slipped to about 0.92, below parity. Now the warrants could not be exercised at all. The equity engine had not broken. It had switched itself off by design, because issuing common below NAV would destroy Bitcoin per share rather than create it. Quarterly Bitcoin yield turned slightly negative, and management started talking about the opposite move, buying shares back, since below NAV a buyback is what concentrates Bitcoin per share.

This is the sign of (mNAV − 1) flipping, made contractual. The accretion formula never cared about the Bitcoin price. But whether mNAV clears 1 cares enormously, and mNAV is one of the most price-sensitive numbers in the whole system. The premium is really a market judgement about future accretion and the future Bitcoin price. It swells in bull markets, shrinks in bear markets, and can turn negative.

So the honest statement is the one the algebra forces on us. A treasury company can pay a Bitcoin yield in a falling market, but only for as long as it can hold a premium. The engine runs on the premium, not on the price. And that, not “number go up,” is the real test of one of these companies: can it keep mNAV above 1 when the price is against it? A company that can is quietly paying its common holders a real Bitcoin yield through the storm. A company whose premium evaporates the moment sentiment turns has no such engine, whatever the investor deck claims.

What's next

Which raises the obvious question. When the premium thins and the common-share gate swings shut, is the company simply stuck, unable to buy another coin without hurting its holders?

Not quite. This is exactly the point where the more interesting operators reach for a different instrument. Metaplanet's own response was to lean on perpetual preferred shares, capital that carries a fixed dividend rather than a claim on Bitcoin per share, and so lets the company keep buying without diluting common holders at all. Strategy built an entire stack of these. In the next post I'll give the same arithmetic treatment to perpetual preferred stock, such as Stretch, and show how it keeps the Bitcoin accruing to common holders even after the common-issuance engine has stalled.

For now, the takeaway is small and slightly surprising, and it is not a thought experiment. A Bitcoin yield does not have to come from a rising price. It comes from a premium. Metaplanet paid one to its common holders through a brutal drawdown, right up until the premium ran out, at which point the machine did exactly what the formula says it should. It stopped.