Sign in

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

TI

TERAWULF INC. (WULF)·Q2 2025 Earnings Summary

Executive Summary

  • Q2 2025 revenue rose 34% YoY to $47.6M and 38% QoQ, driven by higher BTC price and expanded mining capacity, while cost of revenue fell sequentially as power prices normalized; non-GAAP Adjusted EBITDA improved to $14.5M from a Q1 loss, marking operational stabilization .
  • Management announced a 10-year, ~200+ MW hyperscale AI hosting lease with FluidStack, backed by a $1.8B Google support agreement, implying ~$3.7B contracted revenue and ~85% site-level NOI margins; WULF also secured an 80-year Cayuga ground lease for up to 400 MW capacity, reinforcing a >1 GW platform .
  • HPC hosting revenue has begun with WULF Den in July and is set to ramp in Q3/Q4 (CB-1/CB-2), shifting the financial profile to include high-margin hosting; SG&A guidance was raised to $50–$55M for 2025 to support accelerated HPC growth .
  • Street estimates via S&P Global were unavailable for Q2 2025; narrative catalysts include the FluidStack/Google structure lowering financing costs, HPC revenue commencement, and execution milestones at Lake Mariner and Cayuga (potentially positive for medium-term re-rating) ; S&P Global estimates unavailable.

What Went Well and What Went Wrong

What Went Well

  • Signed FluidStack AI hosting lease (~200+ MW) with Google backing ($1.8B credit support, warrants ~8%), adding ~$3.7B contracted revenue potential and ~85% site-level NOI margins; management: “extraordinary vote of confidence” from Google .
  • QoQ operational recovery: GAAP revenue +38% QoQ to $47.6M, cost of revenue exclusive of depreciation fell ~10% QoQ as NY power normalized; Adjusted EBITDA improved to $14.5M from -$4.7M in Q1 .
  • Platform expansion: 80-year Cayuga lease (up to 400 MW, 138 MW expected ready 2026) with strong zero-carbon power/fiber, broadening multi-site capacity for hyperscale/enterprise demand .

What Went Wrong

  • YoY mining metrics reflected halving headwinds: self-mined BTC fell to 485 from 699 YoY; power cost per BTC rose to $45,555 from $22,954 given halving, higher network difficulty, and power volatility .
  • Adjusted EBITDA declined YoY to $14.5M from $19.5M due to halving/difficulty and elevated operating costs as HPC investments ramped pre-revenue .
  • Hosting timing shifted: WULF Den revenue began in July (Q3), and CB-1/CB-2 revenue expected Q3/Q4, versus earlier target for initial revenue in Q2; SG&A guidance raised to support growth (near-term cost pressure) .

Financial Results

MetricQ2 2024Q1 2025Q2 2025
Revenue ($USD Millions)$35.6 $34.4 $47.6
Basic & Diluted EPS ($)$(0.03) $(0.16) $(0.05)
Cost of revenue as % of revenue39.1% 71.4% 46.4%
Adjusted EBITDA ($USD Millions)$19.5 $(4.7) $14.5

Estimates vs. Actuals

  • S&P Global consensus estimates for Q2 2025 (EPS/Revenue/EBITDA) were unavailable; cannot determine beat/miss. S&P Global estimates unavailable.

KPIs

KPIQ2 2024Q1 2025Q2 2025
Bitcoin Self-Mined (#)699 372 485
Value per Bitcoin Self-Mined ($)$65,984 $92,600 $98,219
Power Cost per Bitcoin ($)$22,954 $66,084 $45,555
Operational Hashrate (EH/s)8.0 (nameplate) 7.3 avg operating; 12.2 nameplate 12.2 operational; 12.8 nameplate

Additional Operating/Cash Metrics (Q2 2025)

  • Net loss: $(18.37)M; weighted avg shares 386.9M .
  • Cash, cash equivalents, and bitcoin: $90.0M as of June 30, 2025; debt ~$500.0M (2.75% converts due 2030) .
  • Cash and equivalents period-end: $91.4M (including restricted cash) .

Guidance Changes

MetricPeriodPrevious GuidanceCurrent GuidanceChange
SG&A Expense ($M)FY 2025$40–$45 $50–$55 Raised
Power Price ($/kWh)2H 2025~$0.05 (Q2–Q4 guidance) ~$0.05 (2H guidance) Maintained
Mining EBITDA2H 2025N/APositive contribution expected New qualitative
Hosting Revenue Start2025First revenue Q2 (WULF Den 2.5 MW) WULF Den revenue commenced July; CB-1 in Aug; CB-2 Q4 Slight delay/phase shift
Capex per Critical MW (Core42)2025~$6.1M/MW (initial) ~$7.2M/MW (refined) Raised (design refinements)

Earnings Call Themes & Trends

TopicQ4 2024 (Q-2)Q1 2025 (Q-1)Q2 2025 (Current)Trend
AI/HPC hosting strategySecured Core42 72.5 MW; target 100–150 MW/yr Execution focus on WULF Den/CB-1/CB-2; project financing launch FluidStack lease (~200+ MW) with Google backstop; multi-year ramp Accelerating scale and quality
Financing & capital structure$500M converts; project finance intent Target ~$300M project finance mid-2025 Lower cost of funds expected via Google support; series of transactions Improved credit profile
Power prices/macroQ4 power spike; $0.059/kWh Q1 spike (weather); guide ~$0.05/kWh Q2–Q4 Normalized; guide ~$0.05/kWh 2H Normalization supportive
Site expansionLake Mariner to 500–750 MW roadmap Cayuga integration process underway Cayuga 80-year lease; 138 MW ready in 2026, up to 400 MW Platform expansion
Customer pipelineOption for Core42 expansion; colocation focus Active discussions with enterprises/hyperscalers FluidStack signed; CB5 exclusivity window; strong demand Demand broadening

Management Commentary

  • CEO on strategic inflection: “We signed a ten year 200 plus megawatt hyperscale AI hosting agreement with FluidStack… approximately $3.7B in contracted revenue… Google is providing a $1.8B backstop… an extraordinary vote of confidence” .
  • CFO on financial trajectory: “Our non GAAP adjusted EBITDA showed significant improvement in Q2 totaling $14.5M up from a negative $4.7M in Q1… we’ve updated our annual SG&A guidance to $50–$55M from $40–$45M reflecting accelerated growth in our HPC business” .
  • CEO on execution: “Wolf Den is fully operational and generating revenue. CB1 begins generating revenue within the next week, and CB2 is on track for Q4. We are hitting our milestones on time and on budget” .
  • CSO on Cayuga: “Independent Committee… approved moving forward at a discount to third party bids… equity rather than cash was the right way to structure the deal to keep everybody aligned” .

Q&A Highlights

  • Why FluidStack and Google structure: FluidStack’s deep enterprise engagements; Google backstop equals ~50% of payments over lease term, declining over 10 years; warrants ~8% equity stake—expected to materially lower financing costs .
  • Capacity and CapEx: Core42 total spend ~$430M with back-end heavy items (e.g., UPS); FluidStack build CapEx/MW higher (labor scale, customized design) versus Core42 .
  • Financing roadmap: Series of capital markets transactions in 2H 2025 leveraging Google support; shift from siloed project finance to more flexible, shareholder-friendly structures .
  • Growth cadence: Targeting 150–200 MW annual expansions; potential acceleration given improved credit profile and demand .
  • Operational lessons: Contractor continuity and refined liquid-cooled designs enable faster delivery; learnings transfer to Cayuga for efficient rollout .

Estimates Context

  • S&P Global consensus estimates (EPS, revenue, EBITDA) for WULF Q2 2025 were unavailable; comparative beat/miss vs. Street cannot be assessed. S&P Global estimates unavailable.

Key Takeaways for Investors

  • HPC hosting is now the core value driver: FluidStack/Google deal and Core42 ramp shift WULF’s mix toward high-margin, contracted NOI—expect narrative to tilt from mining cyclicality to infrastructure annuity streams .
  • Near-term financials benefit from normalized power and hosting revenue onset: QoQ revenue growth and EBITDA improvement should continue as CB-1 (Q3) and CB-2 (Q4) contribute .
  • Credit profile upgrade lowers cost of capital: Google support likely reduces financing rates and increases flexibility, supporting multi-asset build-out (Lake Mariner, Cayuga) .
  • CapEx intensity refined but yields preserved: Core42 CapEx/critical MW raised to ~$7.2M with rent per MW ~$1.6M and ~75% EBITDA margins; FluidStack terms also attractive (85% site-level NOI margins) .
  • Mining remains a cash engine, but optionality exists: Management expects positive mining EBITDA in 2H 2025; megawatts could migrate to HPC if economics favor long-term contracted returns .
  • Platform scalability is the edge: Lake Mariner approvals (up to 750 MW) plus Cayuga lease (400 MW) create line of sight to >1 GW capacity—key in a power-constrained AI cycle .
  • Watch execution milestones: CB-1/CB-2 energization timelines, financing prints in 2H, and potential CB5 option with Google-backed terms are the catalysts to re-rate .