Eight words, no filing, no ratio. On September 27 — the year is unstated, which I will return to — the CEO of Strive, a listed small-cap that has been converting its balance sheet into Bitcoin, told a reporter something to the effect of: let's increase the leverage a bit more. No leverage multiple. No cost of capital. No maturity schedule. No counterparty. A media appearance, an adjective, and a plan to borrow against an asset with a 70% drawdown history.
The market read it as bullish. I read it as a disclosure event that failed to disclose. A CEO who describes leverage in diminutives is either withholding the number or does not know it, and both are material. Code does not lie; people do. Capital structures move slower than code — but they print their own audit eventually, and it usually arrives as a collateral call.
Start with what Strive is not. It is not a protocol, not a chain, not an application. There is no code to audit here, no sequencer, no token. What exists is corporate treasury engineering — a financing machine that converts equity premium into Bitcoin and reports the result as a per-share metric.
That template has one inventor and a crowded field of imitators. Strategy proved the pattern at scale. Metaplanet carried the Asian distribution. Semler Scientific wrapped it in medical device cash flow. Twenty One Capital entered with fresh capital and a mining adjacency. Strive sits at the small end of this ladder, and the relevant question is never whether the company likes Bitcoin. It is whether the financing condition that makes the strategy work still holds.
The mechanism, precisely. Treasury company shares often trade at a multiple of the Bitcoin they hold per share. Call that ratio mNAV. When mNAV exceeds 1, the company issues shares at a premium and buys coin at par. Each new share acquires more Bitcoin than it dilutes. That is accretion, and it is real. When mNAV falls below 1, the identical operation runs in reverse: each new share acquires less Bitcoin than it dilutes, and the model becomes a value incinerator with a ticker.
Everything else — the preferreds, the convertibles, the ATM programs, the "Bitcoin yield" dashboards — is instrumentation on top of that single inequality.
And we are in a bear market. In a bear market, the premium is where the fault line runs.
Let me state the arithmetic explicitly, because an adjective has been substituted for it.
Take a company holding 1.0 BTC per share, trading at a 2.0x premium. Issue 10% more shares into that premium. Proceeds: 0.1 shares × 2.0 = 0.2 BTC-equivalent. New holdings: 1.2 BTC. New share count: 1.1. Bitcoin per share: 1.091. Accretion of roughly 9%. The machine works.
Now run the same issuance at 0.9x — a discount. Proceeds: 0.09. Holdings: 1.09. Share count: 1.1. Bitcoin per share: 0.991. You destroyed 0.9% of per-share coin while telling shareholders you acquired more Bitcoin. Both quarters get described in the press release as "increased Bitcoin holdings." Only the second is a slow liquidation.
Leverage does not change the sign of that equation. It changes how fast you arrive at the flip, and how much it costs to stand there.
Work the carry. Borrow at cost c, deploy into Bitcoin with expected return r. The spread r − c is the carry, and in a low-coupon convertible regime, c looks almost free. But low-coupon convertible debt is not free capital. It is a call option sold to the bondholder, struck at your equity, settled in your dilution. In a drawdown, conversion never triggers — bondholders hold par, refinancing resets the coupon higher, and your equity becomes the residual claim on a levered Bitcoin position being marked down daily.
That structure has an event horizon. Going from 1.0x to 1.2x exposure is not a 20% increase in risk. It is a change in the shape of the loss distribution — the left tail thickens, and the route to the right tail starts running through covenant thresholds, margin mechanics, and a counterparty's willingness to keep accepting volatile collateral.
I have seen this shape before. In 2020 I modeled the stETH and Compound interaction during the DeFi summer and wrote a fifteen-page assessment titled "The Illusion of Arbitrage." The spread looked durable right up until liquidity thinned, oracle feeds lagged, and the arbitrage became a liquidation cascade. In 2022 I reconstructed the Terra/LUNA fail-safe and found the only marginal buyer of LUNA was the burn mechanism itself. Forty billion dollars in panic-selling volume later, the design was described as unforeseeable.
It was foreseeable. It was arithmetic. When the marginal buyer is the mechanism, the mechanism is the exit liquidity.
The Strive flywheel has the same topology. Accretion depends on the premium. The premium depends on the accretion. Neither depends on an operating business. That is a reflexive loop, not a cash flow, and reflexive loops have a direction, not a floor.
A single small-cap's Bitcoin bid is a rounding error against spot volume. That is not the point. The point is what its financing condition reveals about the class. When several treasury companies depend on the same premium, a drawdown compresses every premium simultaneously, and every issuer faces the same decision at the same moment. The first mover pauses issuance. The weaker later movers keep issuing into a discount — because halting the flywheel means reporting a flat Bitcoin-per-share quarter, and flat is narrative death for a company whose entire pitch is per-share coin growth. The incentive is to keep issuing past the point of accretion. That is the collective tail risk: not one company deleveraging, but a cohort crossing the inequality together.
The disclosure layer is where this gets genuinely uncomfortable. "Let's increase the leverage a bit more" was delivered to media, not to the SEC. Convertible issuance, preferred placement, and ATM activity all surface in filings — with a lag, and often only in aggregate. If a leverage signal reaches a reporter before it reaches a Form 8-K, that is a selective disclosure problem, not a marketing one.
I spent part of 2024 picking through the custody arrangements of the major spot Bitcoin ETF issuers and finding conflicts of interest that no prospectus summarized. Early in 2026 I audited an AI-agent crypto platform and found its contracts kept no decision trail for autonomous execution. The pattern repeats: the information is never in the sentence. It is in the appendix nobody reads. Leverage with no decision trail is leverage with no accountability, and the difference between the two is a footnote.
Note also the dateline. September 27, year unstated. If this signal was released while the company was mid-transformation — a shell merging, a strategy being re-labeled, management roles shifting — the probability that leverage is being amplified without a complete public financing disclosure goes up, not down. Strive was founded around 2022 as an anti-ESG asset manager, and its founder's attention has since migrated to politics. When a founder becomes a letterhead, the leverage decision belongs to whoever remains in the room. That is a governance question with a financing consequence, and it is unanswerable from the outside.
Three questions determine whether this is a growth story or a warning.
How much leverage, against what covenant, at what collateral haircut? "A bit more" is not a number.
What is the all-in cost of capital, and where is the refinancing wall? A near-zero coupon with a three-year maturity is a near-zero coupon with a three-year fuse.
Can the premium hold, and what is the plan for the day it doesn't? Every treasury company has a story for mNAV above 1. Almost none have one for mNAV below 1, because the story only exists on one side of the inequality.
Audit the promise, not the poster.
Now the steelman, because the bulls are not wrong about everything.
At a sustained premium, this is the cheapest legal channel for a public company to accumulate Bitcoin with other people's money. The coupon can be near zero. The dilution is accretive while mNAV sits above 1. A spot ETF holder cannot manufacture that. Retail Bitcoin exposure is inert — it tracks the price and does nothing else. A treasury vehicle is a machine, and machines can compound per-share coin faster than the underlying appreciates. That is a legitimate, non-trivial innovation, and dismissing it as a gimmick misses the engineering that makes it work.
The bulls are also right about timing. Buying aggressively into a drawdown, mechanically and at scale, is precisely what the strategy was designed for, and the entities that execute it through a full cycle tend to be the ones still holding when the cycle turns. Strive's weakness is not the direction of the trade. It is the financing stack underneath it.
What the bulls get wrong is the classification. High yield is a warning, not a welcome — and a premium-funded accretion engine is high yield wearing a corporate structure. Calling reflexivity a business model is the error. When your accretion depends on your premium, and your premium depends on your accretion, you have rented sentiment, not built an operating business. That is survivable for a while. It is not survivable on the day the rent comes due and sentiment is at its cheapest.
The next Strive filing is the only document that matters. Read the leverage line, the coupon, the conversion terms, and the maturity stack. Then compute market cap against Bitcoin per share, and track that ratio weekly — it, not the CEO's adjective, is the solvency question.
Forensics doesn't take holidays. The arithmetic of a flywheel reversing is always faster than the arithmetic of it spinning up. The question is not whether Strive can buy more Bitcoin. It is whether anyone has stress-tested the quarter in which it can no longer finance the buy.