MARA Holdings reported third-quarter net income of $632.1 million, or $1.28 a diluted share, in its shareholder letter on 9 November. The letter says net income “includes an $885.3 million gain related to the fair value of digital assets”, 1.4 times the profit, by our arithmetic.
Without those gains, income before taxes of $631.1 million would have been a loss of about $254 million, by our arithmetic. Revenue fell 35% to $164.2 million from $252.4 million.
Business Pill · Vertical Integration
An evergreen Business Engineer explainer on vertical integration. It teaches the general idea only and does not discuss MARA.
The key insight: MARA Holdings reported third-quarter net income of $632.1 million, including an $885.3 million gain on the fair value of its bitcoin: without those gains, income before taxes would have been a loss of about $254 million, by our arithmetic. Revenue fell 35% to $164.2 million.
The letter says the rise in net income was “primarily driven by an $885.3 million unrealized bitcoin mark-to-market gain in Q3 2026”, against a $343.1 million gain a year earlier. A 37% decline in the average price of bitcoin mined lowered revenue by $94.8 million, partly offset by an $11.7 million benefit from higher production.
MARA mined 2,247 bitcoin, up 5%, and held 34,012 at quarter-end, including 19,260 loaned or pledged as collateral. For the first nine months, the statements show a net loss of $1.24 billion.
Figures are from MARA Holdings’ third-quarter 2026 shareholder letter. The pre-tax result without the fair-value gains is our arithmetic.
The AI Bet
The letter says Exaion, MARA’s majority-owned subsidiary, “signed a four-year $79M GPU-as-a-service contract on 100% take-or-pay terms with its first US customer”. It cites a pipeline of over $3 billion in total contract value, adding: “These are indications of demand and are not yet contracted or committed”.
MARA says it is targeting two data-center leases before year-end, but that “market conditions are fluid and that timing could slip into next year”.
Power
At Long Ridge, MARA says it received FERC approval and expects to close the acquisition of a gas plant authorized to sell 485 MW; with its existing mining there, the site would have over 1 GW of total potential capacity. It also secured rights to develop a Matagorda County, Texas site with up to 2 GW.
Debt and Buybacks
After quarter-end MARA sold 13,597 bitcoin at about $81,039 for about $1.1 billion and used $750 million to repay its bitcoin-collateralized term loans. It says it has cut total debt by $1.4 billion, or 37%, since the start of the year, and the board authorized up to $500 million of share repurchases.
As we read the letter, the income statement still moves with the bitcoin price, while the strategy now points elsewhere: MARA describes itself as an energy-backed digital infrastructure enterprise spanning power, land and compute.
As we read the AI side, the letter’s newest data point is Exaion’s four-year, $79 million GPU-as-a-service deal on take-or-pay terms, which it calls an early proof point; the $3 billion pipeline and the two leases MARA targets before year-end are not yet signed.
MARA’s chairman and CEO, in the shareholder letter
“Bitcoin built the foundation. Power created the opportunity. Our strategy now is simple: maximize the value of every megawatt (“MW”) we control.”
Three Implications
FOR AI INFRASTRUCTURE CUSTOMERS MARA says Long Ridge received FERC approval and that, with its mining operations there, the site will have over 1 GW of total potential capacity; it has also secured rights to a Matagorda County, Texas site with up to 2 GW.
FOR SHAREHOLDERS After quarter-end the board authorized a share repurchase program of up to $500 million, and MARA repaid $750 million of bitcoin-backed debt with proceeds from bitcoin sales.
FOR THE BITCOIN BALANCE MARA held 34,012 bitcoin at quarter-end and, after it, sold 13,597 bitcoin at about $81,039 for about $1.1 billion.
The framework argues: “Companies that solve the power equation will dominate the AI era; those that don’t will watch their chips sit idle.”
As we read it, MARA is betting that the power it controls is worth more as AI and HPC capacity than as mining capacity. The framework predates these results and does not discuss MARA.
What the Risk Factors Say
MARA’s 10-Q says that, other than the additional risk factors it sets out, it is not aware of material changes to the risk factors in its annual report. By our comparison, three factors are new against its 2025 10-K: the Long Ridge acquisition’s closing conditions, the development of a site that depends on obtaining power and other approvals, and its level of debt.
The Long Ridge factor says that if the acquisition is not consummated by 30 November 2026, a deadline that may be extended to 30 June 2027 if certain regulatory conditions have not been satisfied, MARA could be required to pay the sellers a termination fee of $75.0 million.
What Is Not Established
The letter does not name Exaion’s US customer or any prospective data-center tenant, and the $3 billion pipeline is not contracted. Long Ridge had not closed when the letter was published, and the size of any AI revenue is not broken out.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
MARA’s third-quarter net income of $632.1 million includes an $885.3 million fair-value gain on digital assets; without those gains, income before taxes would have been a loss of about $254 million, by our arithmetic. The letter describes Exaion’s $79 million contract as an early proof point.
A note on sourcing. Results, holdings, contracts and quotations are from MARA Holdings’ third-quarter 2026 shareholder letter (8-K exhibit, 9 November 2026). The 1.4x ratio and the pre-tax result without the fair-value gains are our arithmetic. Estimates on the embedded card are consensus figures cited by Wall St Engine.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
