Sign in

You're signed outSign in or to get full access.

MH

MARA Holdings, Inc. (MARA)·Q3 2025 Earnings Summary

Executive Summary

  • Q3 2025 delivered record revenue of $252.4M (+92% YoY) and net income of $123.1M, aided by a $343.1M gain on fair value of digital assets; diluted EPS was $0.27 .
  • The company announced two strategic initiatives: (1) pending majority acquisition of Exaion (EDF subsidiary) to accelerate AI/HPC private cloud expansion, and (2) a collaboration with MPLX to develop integrated power generation and data center campuses initially ~400 MW, expandable to 1.5 GW in West Texas .
  • Bitcoin holdings rose to 52,850 BTC (incl. loaned/pledged/actively managed) with energized hashrate at 60.4 EH/s; purchased energy cost per BTC was $39,235 and cost per petahash per day improved 15% YoY to $31.3 .
  • Versus S&P Global consensus: revenue slightly missed, EPS missed, and EBITDA was below consensus; differences reflect non-GAAP adjustments and SPGI “Primary EPS” normalization vs company diluted EPS*.

What Went Well and What Went Wrong

  • What Went Well

    • Record revenue and energized hashrate: “Q3 2025 delivered record revenue and energized hashrate of 60.4 EH/s… the highest in the company’s history.”
    • Strategic energy and AI initiatives: “We are jointly announcing a collaboration with MPLX… initial capacity is expected to reach ~400 MW, with the option to expand up to 1.5 GW…,” and pending Exaion acquisition to extend private cloud/AI capabilities .
    • Operational efficiency: Cost/petahash/day improved 15% YoY to $31.3; cost per kWh remained $0.04 at owned sites .
  • What Went Wrong

    • Higher purchased energy costs: $43.1M vs $27.0M prior year, driven by expanded owned mining operations and 64% hashrate growth .
    • Operating and maintenance costs increased to $26.3M (vs $9.4M prior year) due to shipping/warehouse fees and labor costs .
    • Continued reliance on third‑party hosting and other energy costs of $75.7M (+$12.0M YoY); management aims to phase these out over time .

Financial Results

MetricQ1 2025Q2 2025Q3 2025
Revenue ($USD Millions)$213.9 $238.5 $252.4
Operating Income (Loss) ($USD Millions)$(541.1) $679.8 $47.6
Net Income (Loss) ($USD Millions)$(533.4) $808.2 $123.1
Diluted EPS ($USD)$(1.55) $1.84 $0.27
EBITDA ($USD Millions)$(493.4) $1,184.9 $326.4
Adjusted EBITDA ($USD Millions)$(483.6) $1,245.5 $395.6

Estimates vs Actual (S&P Global) – Q3 2025

MetricConsensusActualSurprise
Revenue ($USD)$254,472,020*$252,410,000 Miss*
Primary EPS ($USD)$0.675*$0.3747* vs company diluted EPS $0.27 Miss*
EBITDA ($USD)$327,611,750*$246,705,000* vs company EBITDA $326,398,000 Miss*

*Values retrieved from S&P Global. Note: SPGI “Primary EPS” and “EBITDA” reflect SPGI methodologies; company reports diluted EPS and EBITDA/Adjusted EBITDA per reconciliations, leading to differences .

KPIs and Operational Metrics

KPIQ1 2025Q2 2025Q3 2025
Blocks Won (Units)666 694 633
BTC Produced (Units)2,286 2,358 2,144
Avg BTC Produced per Day (Units)25.4 25.9 23.3
Share of Available Miner Rewards (%)5.5% 5.7% 5.0%
Energized Hashrate (EH/s)54.3 57.4 60.4
Purchased Energy Cost per BTC ($USD)$35,728 $33,735 $39,235
Cost per Petahash per Day ($USD)$28.5 $28.7 $31.3
Cost per kWh ($USD)$0.04 $0.04 $0.04
BTC Holdings (Units)47,531 49,951 52,850

Guidance Changes

MetricPeriodPrevious GuidanceCurrent GuidanceChange
Energized Hashrate TargetFY 2025“On track to reach 75 EH/s by year-end” “Advance toward target of 75 EH/s by year-end 2025” Maintained
Production Reporting CadenceStarting Q4 2025Monthly production reports Shift to quarterly reporting; MARA Pool visible on mempool Changed (reporting approach)
West Texas Integrated Power/Data CampusesMulti-yearN/AInitial capacity ~400 MW with option to expand to up to 1.5 GW across three sites New (strategic capacity plan)
AI Inference DeploymentNear-termPilots discussed prior First ten AI inference racks deployed at Granbury New milestone

Earnings Call Themes & Trends

TopicPrevious Mentions (Q1/Q2 2025)Current Period (Q3 2025)Trend
AI/HPC & InferencePositioning for edge inference; partnerships (TAE Power Solutions, PADO AI); R&D on cooling/ASICs First AI inference racks deployed at Granbury; focus on cost per token; modular/containerized approach Scaling from vision/pilots to initial deployment
Energy Ownership & Vertical IntegrationAcquired 114 MW wind farm; flare gas buildout; shift to owned sites (70% capacity) MPLX collaboration for fuel-to-compute power plants; emphasis on owning generation assets Deepening energy control; larger scale
International Expansion & Sovereign AIGoal to reach 50% revenue from international by 2028; partnerships with energy entities Pending majority stake in Exaion (EDF), advisor Gérard Mestrallet; Europe/Middle East focus Formalizing entry via M&A; credibility building
Cooling Tech (2PIC) & Operations2PIC pilots underway Exited near-term investment in two-phase immersion; favor direct-to-chip cooling Tactical shift to near-term ROI
Grid Flexibility & Modular DCsData center curtailment/value as flexible loads Modular non-water-cooled data centers; behind-the-meter operations; ERCOT connection optionality Continued emphasis
Bitcoin Treasury StrategyBTC actively managed/loaned; SMA; yield per share KPI ~33% holdings activated; continued lending/trading; opportunistic monetization Active treasury management sustained

Management Commentary

  • “Electrons are the new oil. Energy is becoming the defining resource of the digital economy…” .
  • “Our guiding metric is simple: profit per megawatt hour… whether in Bitcoin, AI inference, or grid stability.” .
  • “Initial capacity is expected to reach ~400 MW, with the option to expand up to 1.5 GW across three planned sites” (MPLX collaboration) .
  • “As of September 30, 2025, we held a total of 52,850 Bitcoin, including 17,357 Bitcoin that were loaned, actively managed, and pledged as collateral” .
  • “We have exited near-term investment in two-phase immersion… focus resources on opportunities with more immediate and higher return potential.” .

Q&A Highlights

  • Strategic emphasis: Owning energy assets and integrating power generation with compute to deliver lowest cost per token for AI inference; modular/containerized builds for flexibility and capital efficiency .
  • Regulatory/process: Build gas-fired generation first (air permits), operate behind-the-meter initially; ERCOT engagement upon grid interconnect .
  • Exaion rationale: Access to tier III/IV private cloud expertise, secure data operations, and enterprise customers; expand internationally leveraging EDF subsidiary’s capabilities .
  • Business model optionality: Reluctance to commit to colocation-only models; preference for operating inference AI without mandating colocation deals .
  • Hashrate growth and diversification: Duty to grow global hashrate to support network security; deployment mix across wind, flare gas, and low-cost assets; Auradine miner adoption for load balancing .

Estimates Context

  • Revenue: Actual $252.4M narrowly missed SPGI consensus $254.5M* .
  • EPS: Company diluted EPS $0.27 vs SPGI Primary EPS consensus $0.675* and SPGI Primary EPS actual $0.375*; difference reflects SPGI normalization vs company diluted EPS methodology .
  • EBITDA: Company EBITDA $326.4M vs SPGI EBITDA consensus $327.6M* and SPGI actual $246.7M*; SPGI definitions differ from company’s EBITDA/Adjusted EBITDA reconciliations .
  • Implication: Street likely adjusts models for (i) volatility from fair value gains/losses on digital assets, (ii) evolving mix between owned power and third-party hosting, and (iii) AI/HPC ramp timing.
    *Values retrieved from S&P Global.

Key Takeaways for Investors

  • Near-term: Expect stock sensitivity to BTC price and fair value marks; watch for execution milestones on MPLX power campuses and additional AI inference deployments as catalysts .
  • Medium-term thesis: Vertical integration into owned generation should structurally lower power costs, improve margins, and provide optionality between BTC mining and AI inference workloads .
  • Efficiency focus: Despite network difficulty highs, MARA held cost/kWh at $0.04 and improved cost/petahash YoY; continued migration away from third-party hosting should aid unit economics .
  • Liquidity/Capital: $1.025B zero-coupon converts issued (2032) and ATM proceeds bolster strategic flexibility; combined liquid assets ~$7B at quarter end .
  • International expansion: Exaion (EDF) accelerates sovereign AI/private cloud strategy in Europe; management targets 50% international revenue by 2028 .
  • Reporting/Transparency: Shift to quarterly production reporting beginning Q4; MARA Pool remains observable on mempool .
  • Watch KPI trajectory: EH/s toward 75 by year-end 2025, BTC holdings activation ratio, cost per token for AI inference, and “profit per MWh” narrative adoption by management .

Appendix: Q3 2025 Additional Data Points (from 8-K & Shareholder Letter)

  • Purchased energy costs: $43.1M vs $27.0M prior year, driven by owned operations and 64% growth in hashrate .
  • Operating & maintenance: $26.3M vs $9.4M prior year; shipping, warehouse, and labor .
  • Third‑party hosting and other energy costs: $75.7M vs $63.7M prior year .
  • Depreciation & amortization: $167.3M vs $101.9M prior year; increased fleet deployment .
  • Gain on fair value of digital assets (including BTC receivable): $343.1M in Q3 .
  • Combined unrestricted cash and cash equivalents and BTC: ~$6.8B as of 9/30/2025 .
  • September production update: 218 blocks (+5% M/M), 736 BTC (+4% M/M), EH/s 60.4 (+2% M/M) .
  • August production update: 208 blocks, BTC holdings 52,477; wind farm containers/miners connected .