Key Highlights
- MARA Holdings experienced a 27% year-over-year revenue decrease to $174.9 million in Q2 2026
- CleanSpark’s third fiscal quarter revenue plummeted 30.5% to $138 million
- Both mining operations are pivoting toward AI and high-performance computing sectors
- Market enthusiasm for AI-related announcements has diminished, with typical stock increases falling from 24% to approximately 10%
- The TEM AI Infrastructure Growth Index has declined approximately 28.5% since its June high
Leading cryptocurrency mining firms MARA Holdings and CleanSpark have disclosed disappointing quarterly financial results, reflecting the sector’s ongoing transformation toward artificial intelligence and high-performance computing capabilities.
MARA’s second-quarter 2026 revenue reached $174.9 million, representing a 27% contraction from the $238.5 million generated during the corresponding period in 2025. Meanwhile, CleanSpark’s third fiscal quarter brought in $138 million in revenue, marking a 30.5% reduction compared to the $198.6 million recorded in the previous year.
Financial Losses Expand for Mining Operations
MARA’s net loss expanded significantly to $611.3 million, a stark contrast to the $808.2 million net income achieved in the comparable year-earlier quarter. The company recorded a substantial $343 million fair-value impairment on its digital asset holdings. Adjusted EBITDA deteriorated from a $1.2 billion profit to a $360.9 million loss.
CleanSpark recorded a net loss of $239.8 million, reversing from the $257.4 million net income reported in the previous year’s quarter. The company’s adjusted EBITDA shifted from a positive $377.7 million to a negative $113 million.
During the quarter, MARA successfully mined 2,422 Bitcoin at an average market price of approximately $71,325. The company’s energized hashrate increased 22% year-over-year to 70.3 EH/s. MARA currently maintains a treasury of 35,577 Bitcoin, valued at roughly $2.1 billion, securing its position as the fourth-largest corporate Bitcoin holder globally.
CleanSpark’s holdings stood at 13,924 Bitcoin as of June 30, placing it eleventh among publicly traded companies. The firm also maintained $202.6 million in cash reserves and reported total assets valued at $2.7 billion.
Market Response to AI Infrastructure Moves Weakens
Notwithstanding the financial setbacks, both organizations are advancing their AI infrastructure initiatives. MARA is progressing toward finalizing its Long Ridge acquisition, anticipated to contribute positive EBITDA immediately while enhancing capacity at its Hannibal facility. The company also secured an additional 2 GW through a new location in Matagorda County, Texas.
CleanSpark announced a substantial 20-year lease agreement valued at $6.6 billion at its Sandersville facility with a tenant characterized as having high investment-grade creditworthiness.
Nevertheless, investor reactions to AI-related announcements have moderated considerably. Research conducted by Blocksbridge Consulting examined 25 AI and HPC transactions announced between June 2024 and August 2026. The analysis revealed that average stock price increases on announcement days declined from approximately 24% for initial deals to roughly 10% for more recent transactions.
Initial announcements generated substantial market reactions. Core Scientific’s inaugural CoreWeave hosting arrangement propelled its stock price upward by more than 40%. TeraWulf’s first agreement with Fluidstack resulted in a nearly 60% surge.
Recent announcements have produced more modest market responses. TeraWulf’s 401-megawatt lease arrangement with Anthropic generated approximately 5% stock appreciation. CleanSpark’s $6.6 billion AI hosting transaction resulted in roughly 9% gains.
The TEM AI Infrastructure Growth Index, which monitors companies developing AI data center operations, has declined approximately 28.5% from its June peak. Similarly, the Philadelphia Semiconductor Index has fallen nearly 17% from its July maximum.
MARA CEO Fred Thiel indicated the company intends to engage across various segments of the AI infrastructure value chain while maintaining rigorous capital allocation discipline.


