In the first quarter of 2026, Bitcoin mining company MARA Holdings executed a sweeping strategic pivot. Between March 4 and March 25, the company sold approximately 15,133 Bitcoins in several batches, generating around $1.1 billion in proceeds. Its holdings dropped sharply by about 28%, from 53,822 to 38,689 Bitcoins. Looking at the entire quarter, MARA sold 20,880 Bitcoins at an average price of $70,137 each, totaling roughly $1.5 billion. Alongside this large-scale sell-off, MARA announced it would no longer make major purchases of ASIC miners, instead shifting its focus toward artificial intelligence and high-performance computing infrastructure.
This pivot was directly driven by financial pressures. In Q1 2026, the price of Bitcoin fell from about $87,000 to $68,000—a 22% decline—which significantly reduced the fair value of MARA’s Bitcoin holdings on its balance sheet. The company reported a net loss of approximately $1.3 billion, with unrealized losses from digital asset fair value changes accounting for $1 billion. Meanwhile, mining output for the quarter was just 2,247 Bitcoins, with an average production cost of $76,288 per coin. By the time of sale, market prices had dropped to $70,137, meaning the mining business was operating at an accounting loss.
Beyond the financial statements, MARA is tackling deeper structural challenges: a business model reliant solely on Bitcoin mining is increasingly vulnerable to price swings, especially as halving cycles intensify and energy costs rise. Pivoting to AI infrastructure marks a strategic move for this former second-largest Bitcoin reserve holder as the industry undergoes significant transformation.
How Can Financial Losses and Hashrate Expansion Occur Simultaneously?
In Q1 2026, MARA achieved a record energized hashrate of 72.2 EH/s, up 33% year-over-year. However, operational expansion did not translate into improved financial performance. Revenue for the quarter was $174.6 million, down 18% year-over-year and below analyst expectations of $192.7 million. The core reason for the revenue decline was not reduced output—hashrate growth partially offset lower production—but rather a roughly 18% drop in average Bitcoin prices compared to the same period last year, sharply reducing fiat-denominated income per block.
On the cost side, MARA’s power purchase expenses rose from $35,728 to $40,047 per Bitcoin year-over-year. Including depreciation, management, and debt costs, the all-in cost per Bitcoin ranged from $60,000 to $70,000. This means Bitcoin’s price lows during the quarter were close to—or at—the company’s breakeven point.
Another major source of losses was accounting treatment. MARA measures its Bitcoin holdings at fair value, so price declines are directly reflected in current-period profit and loss, resulting in paper losses. These unrealized losses do not equate to cash outflows, but they highlight the balance sheet risks of holding Bitcoin during market downturns. After selling Bitcoin to repurchase convertible notes at a discount, MARA’s total debt fell from $3.3 billion to about $2.3 billion—a reduction of $1 billion—while saving approximately $88.1 million in interest expenses.
How Has MARA’s Position Shifted in the Global Mining Industry?
In terms of hashrate, MARA remains in the industry’s top tier. However, its market cap ranking has slipped from the largest Bitcoin mining company to seventh place. This shift is not solely due to MARA’s performance decline—some competitors have gained greater market recognition through faster AI transformation and optimized capital structures, resulting in more resilient valuations.
MARA’s ranking among publicly traded companies holding Bitcoin has changed even more dramatically. After its Q1 sell-off, MARA dropped from second to fourth place on the public company Bitcoin holdings leaderboard. Strategy (formerly MicroStrategy) still leads with about 843,738 BTC, while Japan’s Metaplanet surged to third place thanks to frequent purchases in Q1 2026. MARA’s decline, amid others’ increases, reflects a deepening divergence in public company Bitcoin reserve strategies: Strategy leverages preferred stock financing to sustain ongoing acquisitions, while MARA is channeling funds into new business directions beyond hashrate.
How Does the Long Ridge Acquisition Support MARA’s AI Infrastructure Transition?
On April 30, 2026, MARA announced a $1.5 billion acquisition of Ohio’s Long Ridge Energy & Power facility, which includes a 505 MW combined-cycle natural gas power plant and over 1,600 contiguous acres of land. In the second half of 2025, the plant posted annualized adjusted EBITDA of about $144 million, with roughly 76% of its capacity locked in swap agreements at a weighted average price of $38.80/MWh. All-in operating costs are below $15/MWh. The facility also boasts about 100 million cubic feet of associated natural gas supply daily, supporting over 20 years of self-sustained fuel.
This acquisition increased MARA’s owned and operated capacity by 65%, reaching a total of about 2.2 GW across PJM, ERCOT, SPP, and international markets. According to plans, Long Ridge can scale to roughly 600 MW of AI load in the short term, with total potential capacity exceeding 1 GW. The first phase of AI and critical IT load construction is expected to begin in the first half of 2027, with commercial operation targeted for mid-2028.
Beyond Long Ridge, MARA entered a strategic partnership with Starwood Digital Ventures in February 2026, aiming to convert some existing mining sites into AI data fabrics. The collaboration targets over 1 GW of instant IT capability, with MARA retaining up to 50% equity in the joint venture for future cash flow participation. The company also acquired a 64% stake in Exaion to strengthen its enterprise AI and sovereign cloud infrastructure capabilities.
Strategically, MARA is shifting from "using electricity to mine" to "owning electricity and allocating it to the highest-value applications"—including AI training and inference, critical IT loads, and Bitcoin mining. About 90% of MARA’s self-operated mining capacity could be reallocated to AI and IT operations in the future. This approach, treating power resources as core assets to enter the AI compute market, is highly differentiated among US mining companies.
Selling Bitcoin and Pivoting to AI: What Are the Core Market Disputes?
MARA’s large-scale sell-off and strategic shift have sparked significant debate within the market.
Bearish perspectives argue that MARA sold a substantial amount of Bitcoin at an average price of $70,137 near the market lows in March 2026. Just months earlier, the company had financed Bitcoin purchases by issuing zero-coupon convertible notes. This rapid shift from "buy and hold" to selling and financing may signal serious doubts about the sustainability of its Bitcoin reserve accumulation model. While the AI pivot offers a new narrative, the sector is capital-intensive and highly competitive; MARA has yet to sign any HPC leases, so its AI business model remains unproven.
Supportive perspectives contend that MARA’s transformation is based on quantifiable asset advantages. Its owned sites have power costs around $0.04/kWh, among the lowest for US miners. Leveraging this, redirecting power resources to AI infrastructure is a process of repricing existing assets—not abandoning its original business. Discounted debt repurchases generated about $71 million in accounting gains and eliminated ongoing interest expenses, materially improving capital structure. From a long-term asset allocation standpoint, owning power plants and data centers provides more stable cash flows and diversified monetization opportunities compared to simply running mining rigs.
How Do Regulatory Policies Impact MARA and the US Bitcoin Mining Industry’s Long-Term Trajectory?
Entering 2026, US federal regulation of Bitcoin mining has produced complex, dual signals.
On the supportive side, Republican senators introduced the "Mined in America Act" in March 2026, proposing a voluntary federal certification program to support domestic crypto mining and gradually phase out reliance on foreign-made hardware. Additionally, the SEC and CFTC issued joint guidance clarifying that mining rewards under Bitcoin’s proof-of-work mechanism are not considered securities, eliminating longstanding regulatory uncertainty.
On the restrictive side, the Biden administration reintroduced the "Digital Asset Mining Energy Tax" proposal in March 2026, aiming to impose a 30% consumption tax on electricity used for Bitcoin mining. If enacted, this would directly erode the cost advantage of US miners, including MARA.
The tension between these policy lines is a key variable for the industry’s long-term development. Large-scale enterprises with stable power supply and compliant operations are better positioned than smaller, decentralized miners to withstand the impact of electricity taxes and meet certification requirements. MARA’s shift from "electricity consumer" to "electricity owner" can be seen as a proactive measure against potential policy risks: controlling generation capacity enables vertical integration, allowing cost adjustment and reducing sensitivity to external electricity pricing and tax changes.
Is the Collective "Hashrate Migration" of Bitcoin Miners a Structural Upgrade or Strategic Retreat?
MARA’s strategic shift is not an isolated phenomenon. In Q1 2026, several Nasdaq-listed mining companies launched transformation initiatives almost simultaneously. Core Scientific’s AI hosting revenue has surpassed its Bitcoin mining revenue, becoming its largest income source. IREN signed a five-year, $3.4 billion AI cloud services contract with NVIDIA, with AI cloud revenue up 94.2% quarter-on-quarter. By quarter-end, listed miners had signed over $70 billion in AI and HPC-related contracts.
The underlying driver for these pivots is the structural deterioration of mining unit economics. Bitcoin’s halving reduced block rewards from 6.25 BTC to 3.125 BTC over a year ago, and the effects are fully reflected in revenue structures. Meanwhile, network hashrate continues to climb, but Bitcoin prices have not kept pace. Mining costs per coin are approaching—or even exceeding—market prices at times, and the marginal returns from expanding hashrate are diminishing.
MARA’s approach differs from its peers: Core Scientific and IREN primarily provide hosting services for AI compute via third-party contracts, while MARA emphasizes deep vertical integration through direct ownership of power assets. Which model will maintain a long-term competitive advantage depends on a combination of capital efficiency, contract pricing power, and operational execution.
Conclusion
MARA Holdings’ strategic pivot in Q1 2026 represents a sample-level structural adjustment for Bitcoin mining amid deepening halving cycles and surging AI compute demand. The company deleveraged and restructured capital by selling about $1.5 billion in Bitcoin, shifting resources from traditional hashrate expansion to AI and HPC infrastructure. Despite short-term financial pressure from Bitcoin fair value changes, MARA’s core asset base—centered on power resources—and its acquisition of the Long Ridge power plant have laid the foundation for a digital infrastructure platform transition.
The current market assessment of MARA’s transformation hinges less on the validity of the AI narrative and more on whether the timeline for converting that narrative into actual cash flow matches the patience of capital markets. Long Ridge’s initial AI load is scheduled for commercial operation in mid-2028, meaning MARA will rely primarily on Bitcoin mining revenue to sustain operations and fund large-scale capital expenditures until then. Whether MARA becomes the next successful case of power asset monetization after CoreWeave, or faces cash flow and financing challenges amid intense capital competition, remains to be seen in upcoming financial results.
FAQ
Q: How much Bitcoin did MARA sell in Q1 2026, and why?
MARA sold about 20,880 Bitcoins at an average price of $70,137 per coin in Q1 2026, totaling roughly $1.5 billion. The proceeds were mainly used to repurchase convertible bonds at a discount, reducing total debt from $3.3 billion to $2.3 billion, and to fund expansion into AI and data fabric infrastructure.
Q: What are MARA’s current hashrate and Bitcoin holdings?
As of the end of Q1 2026, MARA’s energized hashrate hit a record 72.2 EH/s, up 33% year-over-year. After the sell-off, the company’s Bitcoin holdings dropped from 38,689 to about 35,303 coins.
Q: What are MARA’s core assets for its AI infrastructure transition?
On April 30, 2026, MARA announced the $1.5 billion acquisition of Ohio’s Long Ridge Energy & Power facility, which includes a 505 MW natural gas power plant and over 1,600 acres of land, supporting more than 600 MW of AI load long-term. The company also formed a strategic partnership with Starwood Digital Ventures to convert existing mining sites into AI data fabrics.
Q: How has MARA’s ranking changed among public company Bitcoin holders?
After selling Bitcoin in Q1 2026, MARA dropped from second to fourth place among public companies holding Bitcoin.
Q: What key risks does MARA face in its AI transition?
MARA has not yet signed any HPC leases, so its AI business model lacks critical catalysts for validation. Long Ridge’s initial AI load construction is expected to begin in the first half of 2027, with the first capacity targeted for commercial operation in mid-2028. During this period, the company may face substantial capital expenditure pressures and market doubts about revenue visibility. Additionally, ongoing Bitcoin price volatility continues to directly impact MARA’s mining profitability.
Q: How do US regulatory policies affect the Bitcoin mining industry?
In 2026, US federal regulation is diverging: on one hand, the Mined in America Act proposal supports domestic mining and establishes a voluntary certification system; on the other, the Biden administration’s proposal seeks to impose a 30% consumption tax on electricity used for Bitcoin mining. Meanwhile, the SEC and CFTC have clarified that Bitcoin mining rewards are not securities, eliminating a key regulatory uncertainty in the sector.
