MARA Holdings (NASDAQ: MARA) Q2 2026: Revenue Falls 27%, Net Loss Hits $611M

MARA Holdings (NASDAQ: MARA) reported a sharp deterioration in second-quarter results on August 7, 2026, with revenue falling 27% year-over-year to $174.9 million and GAAP net loss widening to $611.3 million, reversing the $808.2 million net income posted in Q2 2025. GAAP EPS of -$0.70 missed the consensus estimate of $0.19 by $0.89. Shares fell 5.26% on Friday while the S&P 500 gained 0.61%.

Q2 2026 Results

• Revenue: $174.9 million, down 26.7% from $238.5 million in Q2 2025
• GAAP net income (loss): -$611.3 million, compared to +$808.2 million in Q2 2025
• Adjusted EBITDA: -$360.9 million, versus approximately $1.2 billion in Q2 2025
• Bitcoin mined: 2,422 BTC in the quarter; no Bitcoin was purchased
• Bitcoin holdings: 35,577 BTC (approximately $2.1 billion), down 29% year-over-year, including 9,270 BTC loaned or pledged as collateral as of June 30, 2026
• Energized hashrate: 70.3 EH/s, up 22% from 57.4 EH/s in Q2 2025

What Drove the Results

Both revenue and profitability deteriorated sharply despite a 22% increase in energized hashrate to 70.3 EH/s, reflecting a significant compression in Bitcoin mining economics. Adjusted EBITDA collapsed from approximately $1.2 billion in Q2 2025 to -$360.9 million in Q2 2026, a swing of more than $1.5 billion. Purchased energy cost per Bitcoin was $38,690 for owned sites, at a cost per kWh of $0.04. Total blocks won increased modestly to 700 from 694 in Q2 2025, while cost per petahash per day declined 4% year-over-year.

The GAAP net loss of $611.3 million diverged substantially from the adjusted EBITDA loss of $360.9 million, indicating meaningful non-cash charges layered on top of operating losses. The adjusted loss per share of $1.60, per Benzinga, stood in contrast to a Street consensus that had anticipated a $0.26 per share profit.

Strategic Pivot: Power Assets and AI Infrastructure

In its Q2 2026 shareholder letter, MARA framed its existing infrastructure as a platform for both Bitcoin mining and AI-oriented compute, arguing that the defining constraint on AI deployment is access to energized, permitted power, not capital. The company cited its 19 data centers across four continents as a competitive advantage in supplying power-ready infrastructure.

Subsequent to the quarter's end, MARA secured rights to a 2 GW powered land site in Matagorda County, Texas, subject to ERCOT and interconnect approvals. Combined with the pending Long Ridge acquisition, which is awaiting FERC approval, the two transactions are expected to expand MARA's total power portfolio to up to 4.8 GW. The Long Ridge acquisition is expected to immediately contribute positive EBITDA upon closing, with more than 70% of its power output under long-term contracts, according to the shareholder letter.

Wall Street View

Cantor Fitzgerald analyst Brett Knoblauch maintained an Overweight rating on MARA Holdings on August 7, 2026, while lowering his price target to $12 from $14, reflecting the weaker Q2 results. The target reduction acknowledges near-term mining economics without abandoning a constructive long-term view on the company's infrastructure repositioning.

Investor Takeaway

The Q2 miss was broad-based: revenue, net income, adjusted EBITDA, and EPS all deteriorated substantially versus year-ago levels despite operational expansion in hashrate and blocks won, highlighting that volume growth alone has not offset the compression in Bitcoin mining profitability. The central question for investors is whether the company's pivot toward owning large-scale power assets, particularly the Matagorda County site and the Long Ridge campus, can generate durable cash flow as Bitcoin mining economics remain under pressure. The pending regulatory approvals for both transactions represent the clearest near-term catalyst that could shift the financial trajectory.