Mentioned in This Article
1. What did MARA Holdings report for the third quarter?
MARA Holdings, the Bitcoin (BTC) miner formerly known as Marathon Digital Holdings, reported net income of $632.1 million, or $1.28 per diluted share, for the third quarter of 2026 in results released on November 9. The figure compares with $123.1 million in the same quarter a year earlier, even though revenue fell 35% to $164.2 million. After the quarter closed, the company sold 13,597 BTC for roughly $1.1 billion.
The headline profit is striking next to the revenue decline, and the gap between the two is the core of the story. Revenue is what the company earns from running its mining operation, while the profit figure also captures changes in the value of the Bitcoin it keeps on its balance sheet. In this quarter the second effect was much larger than the first.
The company reported the results in its quarterly shareholder letter, which accompanied its regulatory filings. Investors tracking the stock tend to read the letter for operating metrics such as hashrate and coins produced, and for balance-sheet items such as cash, holdings and debt, all of which carried notable changes this time.
2. Where did the $632.1 million profit come from?
Most of the profit came from an unrealized fair-value gain of $885.3 million on the company's digital assets, not from selling the coins it mined. Under current accounting rules, a company that holds crypto must mark it to market each quarter, so a rise in the Bitcoin price during the period is recorded as income even if nothing has been sold.
That treatment helps explain why a quarter with falling revenue can still produce a large profit. The gain reflects the value of MARA's roughly 34,000 BTC at the end of September compared with the start of the quarter. If the price moves the other way in a later quarter, the same rule produces a loss, which is why results at miners that hold Bitcoin can swing sharply from one period to the next.
Put simply, the reported profit tracked Bitcoin's price at the quarter's end more closely than it tracked the performance of the mining business. The operating side of the company generated $164.2 million in revenue, a fraction of the paper gain that lifted the bottom line.
3. Why did revenue fall 35% when MARA mined more Bitcoin?
Revenue dropped to $164.2 million from $252.4 million a year earlier because the average Bitcoin price the company realized fell 37%, which outweighed gains in output and efficiency. Mining revenue is paid in coins, so when each coin is worth less in dollar terms, the same or greater production can still translate into lower sales.
MARA mined 2,247 BTC during the quarter, and its hashrate, a measure of computing power directed at the network, reached 71.2 exahashes per second (EH/s), up 18% from a year earlier. A higher hashrate generally gives a miner a larger share of the block rewards available on the network, and the company described itself as operating more efficiently than before.
The 37% decline in realized price is the figure the analysis of the quarter singles out as the one to watch. It shows the underlying economics of mining: output can rise while income falls, because the price of the asset being produced is outside the miner's control. For a company with large fixed costs for power and equipment, that exposure is the main source of earnings volatility.
Hashrate and price are separate forces. Network difficulty adjusts to the total computing power competing for rewards, so a miner that grows its hashrate is partly running to keep its share as others expand. That is why efficiency, meaning the cost of power per coin produced, matters as much as raw capacity when the price of Bitcoin declines.
4. How big is MARA's Bitcoin and cash position?
As of September 30, MARA held 34,012 BTC, a total that includes receivables, along with $946.5 million in cash and cash equivalents. Together, those assets give the company a large balance sheet relative to its operating revenue, which is typical of public miners that keep much of what they produce rather than selling it immediately.
Holding coins is a strategic choice. It gives shareholders direct exposure to Bitcoin's price, but it also means the company's reported results rise and fall with the market. Cash, on the other hand, gives the company flexibility to pay down obligations, fund new projects or repurchase shares, and the post-quarter sale was a move to convert a portion of the Bitcoin position into that kind of flexibility.
The position described in the filing is a snapshot at the end of the quarter. Because the 13,597 BTC sale happened afterward, the company's holdings and cash at the time of the announcement differed from the September 30 figures, with a meaningful share of the coins converted into cash and then applied to debt.
5. What did MARA do with the 13,597 BTC it sold?
The company directed most of the roughly $1.1 billion in proceeds toward debt, using about $750 million to repay Bitcoin-backed term loans. The sale took place after the quarter ended, at a price of about $81,039 per coin according to the company's shareholder letter.
Repaying those loans released collateral. The coins that had been pledged against the borrowing, between 17,477 and 18,750 BTC, were freed up, leaving pledged Bitcoin below 5% of the company's total holdings. Lower collateral requirements reduce the risk that a sharp price decline would force a lender to demand more coins or liquidate a position.
MARA has reduced its debt by $1.4 billion, or 37%, since the start of the year. For a miner with a large treasury, that deleveraging changes the profile of the balance sheet: the company still owns a large Bitcoin position, but a far smaller portion is encumbered, and interest costs on the repaid loans fall away.
Selling coins to retire debt also reflects a broader pattern among listed miners, which have used their holdings as a source of funding for capital projects. In this case the move reduced leverage rather than financing expansion directly, although the company is also pursuing growth in other areas.
6. What is the $500 million share repurchase program?
MARA announced a share repurchase program of up to $500 million. A buyback lets a company purchase its own shares on the market, which reduces the number of shares outstanding and can raise earnings per share, and it is often read as a sign that management considers the stock undervalued or that it has capital to spare.
The announcement came alongside the debt repayment, which suggests the company is trying to strengthen its balance sheet and return capital at the same time. The program is an authorization up to a ceiling, and the pace and size of any purchases depend on later decisions by the company.
For shareholders, the buyback offers a second lever beyond Bitcoin's price. If the company chooses to use cash for repurchases rather than hold it, it signals confidence in its own outlook, though it also uses resources that could otherwise go toward acquisitions or equipment.
7. How is MARA expanding into AI and high-performance computing?
MARA is broadening its business beyond Bitcoin mining into artificial intelligence and high-performance computing. The company approved a strategic acquisition of the Long Ridge natural gas plant and secured a $79 million GPU-as-a-service contract through Exaion.
GPU-as-a-service means renting out graphics-processing hardware, the chips used to train and run AI models, to customers who pay for access over a contract term. For a miner, the appeal is that it already operates large data-center sites with power supply and cooling, and contracted revenue is more predictable than mining income that varies with the Bitcoin price.
Owning a natural gas power plant such as Long Ridge would give the company a direct source of electricity, which is the central input for both mining and computing workloads. Power access has become a defining constraint for data-center development, and miners with energy assets have drawn interest from companies seeking capacity for AI.
The analysis of the quarter notes that public miners combine an operating business with a Bitcoin treasury. In the latest quarter the treasury's gain far exceeded revenue, so reported profit followed Bitcoin's price. Contracts like the Exaion agreement could, over time, smooth earnings by adding income that does not depend on the coin's market value.
8. What should investors watch next?
The most important measure to follow is the realized Bitcoin price, since its 37% drop was the main reason revenue fell. If prices recover, revenue should improve even without additional production, while a further decline would put pressure on the mining business regardless of hashrate growth.
A second item is how much of the $500 million buyback the company actually carries out, and how it balances repurchases against the AI and power projects. The size of the remaining Bitcoin holdings after the post-quarter sale, and whether the company continues selling coins to reduce debt or fund expansion, will also shape how closely its earnings track the Bitcoin market.
Finally, the progress of the Long Ridge acquisition and the Exaion contract will show whether diversification can reduce the company's dependence on price swings. The shareholder letter did not set out a date for those milestones in the material reviewed, so the next quarterly report is the likely point at which they will be updated.
Readers comparing quarters should also separate the three layers of the results: the mining operation, the Bitcoin treasury and the financing decisions. Each moves for different reasons, and the third quarter showed all three at once, with a falling realized price, a large unrealized gain and a major debt reduction.