MARA Sells 15,133 BTC for $1.1B to Retire Convertible Debt at a Discount
MARA Holdings disclosed in an SEC filing on March 26 that it sold 15,133 Bitcoin for approximately $1.1 billion between March 4 and March 25, using the proceeds to repurchase roughly $1 billion of its zero-coupon convertible notes at an average discount of 9% to par. The transaction reduces the company’s outstanding convertible debt by approximately 30%, from around $3.3 billion to $2.3 billion, and is expected to generate $88.1 million in cash savings before transaction costs. At the time of writing, Bitcoin is trading at $68,566, down 2.19% over the prior 24 hours, which places the residual treasury of 38,689 BTC at a market value of roughly $2.65 billion.
The Mechanics of a Reverse Convertible Play
The structure of the repurchase deserves careful attention, because it is not simply a distress liquidation dressed up in strategic language. MARA’s convertible notes are zero-coupon instruments, meaning they carry no periodic interest burden; their cost to the issuer is dilution risk, the possibility that noteholders convert debt into equity at prices that disadvantage existing shareholders. By repurchasing those notes at a 9% discount to face value during a period when depressed share prices have eroded the conversion option’s appeal to holders, MARA’s management has executed a liability-management trade with a measurable positive spread. That spread, $88.1 million in gross savings, is real capital recovered from a market dislocation rather than an accounting fiction.
The note-level detail matters here. MARA will repurchase $367.5 million of its 2030 notes for $322.9 million and $633.4 million of its 2031 notes for $589.9 million, with both transactions scheduled to close on March 30 and 31, respectively, pending customary conditions. After closing, $632.5 million of 2030 notes and $291.6 million of 2031 notes will remain outstanding, alongside $48.1 million of 1.0% notes due 2026, $300 million of 2.125% notes due 2031, and $1.025 billion of zero-coupon notes due 2032. The residual debt load remains substantial, but its composition is more manageable; the longest-dated maturities carry no coupon, preserving near-term cash flow for operational and strategic deployment.
CEO Fred Thiel framed the decision in terms of optionality rather than necessity. “Our decision to sell a portion of our Bitcoin holdings reflects a strategic move designed to strengthen our balance sheet and position the company for long-term growth,” he said in a press release, adding that the repurchases preserve shareholder value and provide greater financial flexibility as the company expands into digital energy and AI and high-performance computing infrastructure. That framing is consistent with the arithmetic: buying back dilutive instruments below par when your share price is depressed is a textbook value-accretive capital allocation decision, not a distress signal.
Treasury Reduction in Context
The sale represents 28% of MARA’s Bitcoin treasury as it stood at the end of February, when the company held 53,822 BTC. The reduction is large enough to warrant scrutiny of whether the company’s identity as a Bitcoin accumulator remains intact. The honest answer is that it does, but with meaningful qualification. At 38,689 BTC, MARA retains the second-largest corporate Bitcoin position among publicly listed companies, behind only Twenty One Capital, according to Bitcointreasuries.net. The treasury is worth approximately $2.65 billion at current prices, which compares favourably with the $2.3 billion in remaining convertible debt; the company is not, by any standard measure, underwater on a net asset basis.
The wider context for the sale is a Bitcoin price that has retreated substantially from its October 2025 all-time high above $126,000. The stock itself reflects that compression; MARA shares remain roughly 44% lower over the past six months even after rising more than 5% on Thursday’s session, touching an intraday high of $9.29 before settling around $8.74. That six-month drawdown is the backdrop against which the discount repurchase becomes available: convertible noteholders who bought in at higher implied Bitcoin and equity valuations are willing to accept 91 cents on the dollar to exit, and MARA’s management has correctly identified that window as finite. With 107,560 blocks remaining until the next Bitcoin halving at time of writing, the structural case for a recovering hash price over the next several quarters is not trivial, which makes locking in this discount now rather than later a defensible call on timing.
Mining Economics Are the Forcing Function
To understand why MARA is reconfiguring its balance sheet rather than simply holding through the cycle, one has to look at the underlying mining economics. Bitcoin’s hash price, a measure of daily revenue per unit of mining power, has fallen to approximately $33 per petahash per second per day, according to CoinShares research head James Butterfill, roughly half the $64 per petahash recorded in July 2024. That compression is partly a function of Bitcoin’s price decline from its peak and partly a consequence of the post-halving reduction in block subsidy, which cut the per-block reward from 6.25 BTC to 3.125 BTC. Miners operating on levered balance sheets during a period of compressed hash price face a structural mismatch between asset values and debt obligations that cannot be resolved through operational efficiency alone.
The industry response has been a broad pivot toward AI and high-performance computing infrastructure, as covered in detail in our analysis of how the AI pivot is reshaping miner capital structures. MARA’s agreement to acquire a majority stake in Exaion’s AI-focused data centres places it within a peer group that includes Bitdeer, which has sold its entire Bitcoin treasury to fund a cloud-and-AI pivot, and Cipher Mining, now rebranded as Cipher Digital, and the entity formerly known as Bitfarms, now operating as Keel. Butterfill projects that Bitcoin miners collectively could derive as much as 70% of their revenue from AI by the end of 2026. Whether that projection proves accurate depends heavily on AI infrastructure demand curves that remain contested, but the directional shift in capital allocation is already observable across the sector.
The Convertible Note Model and Its Limits
The convertible note financing mechanism that has defined listed Bitcoin miners’ capital strategies since 2021 was, in its original formulation, a one-directional vehicle for accumulation. Strategy, formerly MicroStrategy, pioneered the approach and has raised tens of billions through zero-coupon convertible issuance to fund a treasury that now exceeds 738,000 BTC. The logic was straightforward: issue cheap debt, buy Bitcoin, allow the asset’s appreciation to make the conversion feature attractive to noteholders while shareholders benefit from leveraged upside. That logic held convincingly when Bitcoin was trending toward $126,000; it becomes considerably more complex when the asset has retraced nearly 46% from that peak, as is the case today.
MARA’s reverse application of the same instrument, buying back notes at a discount rather than issuing new ones, demonstrates that the convertible structure contains embedded optionality for the issuer as well as the holder. Cointelegraph’s reporting on the SEC filing confirms that the 9% discount to par is a direct function of depressed share prices reducing the notes’ conversion value. The lesson for institutional observers is that convertible note structures are not passive instruments; they create dynamic two-way markets in which the issuer’s optimal action shifts as underlying asset and equity prices move. MARA has recognised that shift and acted on it with measurable financial benefit.
Who Benefits, and What Comes Next
The clear beneficiaries of this transaction are MARA’s existing equity holders. Eliminating the conversion overhang associated with $1 billion in notes removes a source of potential dilution that, at sufficiently high Bitcoin and equity prices, would have transferred value from shareholders to noteholders. The $88.1 million in gross savings is a secondary benefit; it is real, but it is the dilution reduction that carries the longer-term structural significance. Bondholders who tendered at a 9% discount have accepted a worse-than-par exit in exchange for liquidity, which suggests their assessment of MARA’s equity recovery prospects over the next four to five years was sufficiently pessimistic to make that trade rational for them.
The company’s stated intention to sell Bitcoin “from time to time” as part of its 2026 capital and liquidity strategy signals that this transaction is not a one-off. Management is treating the treasury as a working asset rather than a permanent store, which is a meaningful philosophical departure from the pure accumulation model exemplified by Strategy. That departure carries risk: if Bitcoin recovers sharply toward its prior highs over the next 12 to 24 months, MARA will have sold 15,133 BTC at prices between roughly $68,000 and $73,000 that it could have retained. However, the counterfactual in which the company does not reduce its debt load and then faces rolling maturities during a sustained bear cycle is considerably more damaging to equity holders than the opportunity cost of a partial treasury reduction at current prices. The balance sheet improvement is concrete; the forgone upside is speculative.
The more consequential question for the sector is whether MARA’s active treasury management model, combined with a genuine operational pivot toward AI infrastructure, produces a revenue profile that justifies a re-rating of the stock relative to pure-play Bitcoin miners. The market’s initial reaction, a gain of more than 5% on a day when Bitcoin itself fell roughly 3%, suggests investors are beginning to assign value to the deleveraging independently of the underlying asset price. That divergence, between equity performance and Bitcoin price, is precisely what management needs to sustain as it builds out the AI revenue base. Whether the Exaion acquisition and adjacent initiatives deliver enough recurring, non-Bitcoin revenue to anchor that valuation gap is the central empirical question for MARA over the next four quarters, and the answer will determine whether this balance sheet restructuring is remembered as prudent capital stewardship or as the moment the company began trading its core identity for a more uncertain hybrid one.
