Strive bought 4,045 bitcoin last week and finished it owning less bitcoin per share. That is not a view on price. It is division.
On Oct. 9 the Dallas-based treasury company (Nasdaq: ASST) disclosed 32,773 BTC against roughly 101,169,907 Class A shares. By Oct. 16, per the Form 8-K filed Monday, it held 36,818 BTC against 114,847,610 shares. Two ratios: 0.000324, then 0.000321. A 12.3% expansion of the coin stack produced a 1% contraction in what each share owns. Roughly $334.9 million left the building to make that happen β 4,045 coins at a weighted average of $82,803, fees and expenses included.
Treasury companies sell this as accretion. Few publish the denominator beside the numerator. Strive did, which is why the claim is auditable.
The mechanics behind the week matter more than the headline. Strive absorbed Asset Entities in September 2025, inheriting a warrant stack that sat on its balance sheet as contingent dilution. Between Oct. 12 and Oct. 13, holders exercised 273,493,909 of those warrants for 13,674,703 Class A shares β a 20:1 conversion β delivering $369.2 million in gross proceeds. The residue lapsed unexercised on Oct. 13. The company now reports no traditional warrants outstanding.
That cash did not sit. Cash and equivalents moved from $310.3 million to $317.5 million across the week β a $7.2 million drift against a $369.2 million inflow. The rest went into coin. Strive also carries 505,000 shares of Strategy's STRC preferred, marked at $49.3 million as of Oct. 16.
Then the forward leg. On Oct. 15 the company priced 7 million shares of SATS β a perpetual Series A preferred paying 15% β at $90 per share against a $700 million stated amount. Expected net proceeds are roughly $630 million. Settlement is Tuesday, Oct. 20. As of Oct. 16, no SATS shares were outstanding.
Start with the tranche split, because it is the only place the 8-K lets you see execution quality. The 1,223 coins bought Oct. 14β16 cost $81,734 each. Back out of the weekly total and the earlier 2,822 coins β announced Oct. 14 β priced at roughly $83,266. The second tranche cleared $1,532 per coin cheaper, a 1.8% improvement on the same asset inside the same week. That is not skill; it is the tape. But it confirms the desk is buying in slices, not blocks, and it confirms the warrant cash was deployed within roughly 72 hours of settlement.
The warrant arithmetic is exact enough to reconstruct. $369.2 million divided by 13,674,703 shares is $27.00 β the exercise price, to the cent. That was the cheapest equity Strive will ever issue, and it is now spent. Every future coin must be financed at the common's market price or at the cost of the new preferred.
The cash reconciliation deserves the same treatment. $369.2 million came in from exercises; roughly $334.9 million went out to bitcoin; cash rose $7.2 million. That leaves about $27 million that coin purchases do not explain, and the 8-K does not decompose it. Treasury vehicles live or die on residuals like that.
Now the preferred. Seven million shares against a $700 million stated amount implies $100 stated value per share, sold at $90 β a 10% discount to par. The dividend accrues on par, not on the check. Eighteen months of 15% on $100 is $22.50, exactly the per-share reserve figure in the pricing release: $157.5 million across the issue. The effective cash cost of SATS capital is 16.67%, not 15%, because the coupon is indexed to a stated amount no buyer actually paid.
Set that against the liquidity stack. Cash of $317.5 million plus $49.3 million of STRC is $366.8 million of liquid or near-liquid value. The 18-month reserve consumes $157.5 million of it β 43%. And the reserve is not a segregated trust in the 8-K; it is a stated intention to set aside "the equivalent of" that sum from cash and securities held. Equivalent is carrying a lot of weight in that sentence.
The structural problem follows. Bitcoin generates nothing. A perpetual preferred paying 15% on par, with no maturity and no disclosed conversion, generates a fixed claim every quarter regardless of what the coin does. Strive is running negative carry of roughly 16.67% annualized against a zero-yield reserve asset. For that to be rational, BTC must compound faster than the cost of the instrument funding it. Strategy solved this with low-coupon convertibles that convert into equity at higher prices. Strive sold straight perpetual preferred at a headline rate roughly four times the sector norm β and sold it to buy more of the same volatile asset. Bear markets strip the leverage, leave the logic.
One more line deserves attention. The 505,000 STRC shares are not a hedge. They are second-order beta on the same underlying: in a drawdown, STRC marks down while the 15% obligation keeps accruing. Correlated collateral is a familiar failure mode. In a 2024 custody review I found a multi-sig threshold that diverged from the filed policy by exactly one key β and the correlation problem rhymed: every component moved together in precisely the stress case that mattered.
None of this is an argument that Strive is insolvent. It is an argument that the balance sheet has acquired a duration mismatch: a perpetual liability priced off par, against an asset with no cash flow, funded partly by a reserve that is an intention rather than an account.
Here is the part that should concern anyone reading the 8-K as a risk document. The dividend reserve β the single largest discretionary commitment attached to $630 million of new capital β does not appear in it. The pricing release states the intent to fund 18 months of distributions. The CEO restated it on X the same morning: "Strive plans to maintain current 18-month dividend reserve." The legally binding filing is silent. The code whispers what the auditors ignore, and here the marketing speaks while the filing stays quiet. Yellow ink stains the white paper.
This is not a scandal. It is a disclosure hierarchy, and the hierarchy is the signal. The number that matters to SATS holders lives in a press release and a post, not in the document the SEC receives. Logic holds when markets collapse; prose does not. When the reserve is drawn to pay distributions during a drawdown, the only question that will matter is whether "equivalent" meant segregated or meant available. Entropy increases, but the hash remains β and so does the obligation, whether or not it was filed.
Tuesday's settlement is the real test, not the coin count. Watch three things: whether the next 10-Q discloses the reserve as a restricted line or a footnote; whether any of the $630 million is redirected from "bitcoin and bitcoin-related products" back into the reserve; and what the company does when the common trades below the cost of the preferred that funded it. Entropy is cheap. The 16.67% is not.