WULF

1. Executive Summary

• TeraWulf accelerated its pivot toward AI/HPC infrastructure with major, contracted growth and balance sheet firepower to fund it. The company closed a fully upsized $1.0 billion 1.00% convertible due 2031 and expanded its Fluidstack relationship to 360 MW of contracted IT load at Lake Mariner, taking total contracted revenues to approximately $6.7 billion (with potential to ~$16 billion on extensions). Google increased its backstop to ~$3.2 billion and its pro forma equity stake to ~14%.
• Execution remains key near term: Q2 2025 Adjusted EBITDA softened year over year and Bitcoin mining economics remain pressured; however, management now guides initial HPC hosting revenue recognition beginning in Q3 2025, with a substantial ramp through 2026 supported by interconnection capacity and a newly secured 80‑year Cayuga site.

Key Takeaways

• Liquidity secured for build-out: Net proceeds of ~$975.2 million from the Aug-2025 convert (greenshoe fully exercised), with capped calls in place (cap price $18.76).
• Contracted growth locked in: Fluidstack expansion to ~360 MW IT load at Lake Mariner; contracted revenue ~$6.7 billion (potential up to $16 billion with extensions); Google backstop lifted to ~$3.2 billion and equity to ~14%.
• HPC revenue inflection: Updated guide to begin recognizing HPC hosting revenue in Q3 2025 (supersedes earlier Q2 guide); rapid deployment cadence through 2026.
• Scale and power access: Interconnection secured for 500 MW (plan to 750 MW), plus Cayuga (up to 400 MW) with ~138 MW expected ready in 2026 at <$0.05/kWh.

2. Financial Performance

Capital Raises & Proceeds

• Aug 2025: $1.0B 1.00% Convertible Senior Notes due 2031 (Rule 144A)
• Net proceeds ~$975.2M; capped calls added (cap price $18.76, a 100% premium).
• Initial conversion rate ~80.4602 shares/$1,000 (conversion price ~$12.43, ~32.5% premium to 8/18/25 close of $9.38).
• Use of proceeds: data center expansion and general corporate purposes; $100.6M to capped calls.
• Oct 2024: $500M 2.75% Convertible Senior Notes due 2030
• Net proceeds ~$487.1M; capped calls (cap price $12.80).
• Concurrent $115M stock repurchase; oversubscribed per management commentary.
• Shareholder returns: $200M repurchase authorization (Oct 2024) with $150M+ executed by YE 2024 and $33M repurchased YTD by May 2025.

Investor read-through: The Aug-2025 deal’s upsize and full greenshoe exercise indicate strong institutional demand; capped calls help mitigate dilution up to the cap.

Early Revenue Initiatives

• Core42 (G42): 72.5 MW HPC hosting delivery in 2025 remains on schedule; management now guides first HPC revenue in Q3 2025 (updated from prior Q2).
• Fluidstack:
• Aug 14, 2025: Two 10‑year HPC colocation agreements (>200 MW) anchoring ~$3.7B contracted revenue (potential ~$8.7B with extensions); indicative ~85% site NOI margin (~$315M annually for initial scope).
• Aug 18, 2025: CB‑5 (160 MW) option exercised on identical economics; total contracted IT load ~360 MW; contracted revenue ~$6.7B; potential ~$16B with extensions. Google increased backstop to ~$3.2B; warrants for 32.5M shares; ~14% pro forma equity.
• Pilot and path-to-scale: 2–2.5 MW HPC proofs-of-concept were completed earlier at Lake Mariner to de-risk deployments.

Expense Management & Cash Flow

• Q2 2025: Revenue $47.6M (vs. $35.6M Q2 2024); Adjusted EBITDA $14.5M (vs. $19.5M Q2 2024).
• H1 2025: Net loss $79.8M (vs. $20.5M loss H1 2024), driven by higher energy costs and the 2024 Bitcoin halving.
• Cost structure:
• Mining unit costs increased materially post-halving; management leans on demand response and miner efficiency programs to mitigate.
• SG&A guide unchanged post-Beowulf E&D acquisition: $40–45M SG&A and $20–25M operating expenses (2025).

3. Guidance and Future Outlook

Production Ramp–Up

• 2025: Deliver 72.5 MW HPC capacity to Core42; begin HPC revenue in Q3 2025.
• 2026: Lake Mariner ramp to 200–250 MW operational by YE 2026; Fluidstack CB‑5 (160 MW) operations commence 2H 2026; initial ~40 MW online in H1 2026, full deployment of initial Fluidstack scope by YE 2026.

Expansion Plans

• Lake Mariner (NY): Interconnection to 500 MW (target 750 MW). Ongoing discussions with Fluidstack for additional capacity beyond ~360 MW contracted.
• Cayuga (NY): 80‑year ground lease, 183 acres, scalable to ~400 MW; ~138 MW low-cost power expected ready 2026; plan for 67 MW solar and 800 MWh BESS.
• Corporate development: Beowulf E&D acquisition consolidates energy expertise and simplifies the structure to support financing and execution.

Operational Targets

• Economics: “Same economic terms” for CB‑5 suggests consistency with disclosed ~85% site NOI margins for earlier Fluidstack agreements.
• Execution cadence: Purpose-built, liquid‑cooled data halls; rapid deployment with long-lead items secured; high‑density readiness.

4. Strategic Positioning and Initiatives

Cost Management

• Power arbitrage and DR programs to reduce effective energy cost; average site power < $0.05/kWh at Cayuga enhances blended footprint cost.
• Capped calls limit dilution at higher share prices; convert coupons at 1.00%/2.75% lower cash interest burden vs. term debt.
• Vertical integration post-Beowulf E&D acquisition to optimize energy procurement and reliability.

Product Development

• Purpose-built HPC/AI data halls (CB series) with closed-loop water cooling, dual 345 kV transmission lines, and ultra‑low‑latency fiber.
• Modular deployment approach demonstrated (POC to multi‑hundred MW scale).

Market Expansion

• Anchor customers: Core42 and Fluidstack with multi‑year, high-visibility contracts; Google as financial backstop partner to Fluidstack.
• Geographic diversification via Cayuga to support incremental MW and customer mix.

5. Competitive Positioning and Market Trends

Market Positioning

• One of the largest U.S. HPC campuses under development with ~360 MW contracted at Lake Mariner; unique combination of power availability, cooling, and speed to market.

Competitive Strengths

• Low-cost, predominantly zero‑carbon power at scale; secured interconnection capacity.
• Execution credibility: rapid scaling, customer wins, and fully subscribed capital raises.
• Balance sheet flexibility to fund multi‑year growth.

Emerging Industry Trends

• AI compute scarcity: GPU supply constraints and power bottlenecks favor well-sited, ready‑to‑build operators.
• Shift to high‑density liquid cooling data halls; customers seek long‑term, stable TCO and sustainability.
• Capital intensity rising across AI infrastructure; strong sponsor and capital markets access is a core differentiator.

6. Technology and Innovation Strategy

Technological Advancements

• Closed‑loop water cooling and high‑density hall designs tailored for GPU clusters.
• Power system redundancy and grid interconnection design enabling rapid MW step-ups.

New Product Developments

• CB‑3/CB‑4/CB‑5 Lake Mariner program for Fluidstack; ongoing Core42 dedicated buildings.
• Planned solar + BESS at Cayuga to enhance resiliency and green credentials.

Alignment with Market Needs

• Facilities engineered for AI/HPC workloads (cooling, power density, fiber latency), matching hyperscaler and AI lab requirements.
• Contract structures (10‑year terms, extensions) align to customers’ long‑duration compute roadmaps.

7. Risk and Reward Analysis

Growth Catalysts

• Contracted backlog: ~$6.7B contracted revenue (potential to ~$16B with extensions) underpins multi‑year growth visibility.
• Capital stack: ~$1.5B in low‑coupon converts (2030/2031) and strong market access support continuous build-out.
• Power access: Lake Mariner (to 750 MW) + Cayuga (to ~400 MW) address industry-wide power scarcity.

Downside Risks

• Execution/timing risk: Build schedules (e.g., Fluidstack CB‑5 in 2H 2026) and the updated Q3 2025 HPC revenue start.
• Customer concentration: Material exposure to Core42 and Fluidstack; counterparty risk partly mitigated by Google backstop.
• Energy price volatility and weather-driven curtailments impacting mining and site-level economics.
• Financing/dilution: Future capital needs and convert dilution if shares trade above $12.43; capped calls mitigate dilution up to $18.76.

Valuation Metrics

• Framework (backlog-driven):
• At full deployment of the ~$6.7B contracted base, implied annual revenue is ~$670M (10‑year term). Using disclosed ~85% site NOI for initial agreements and “same terms” for CB‑5 implies potential site NOI of ~$570M at steady state.
• Applying AI colocation EV/EBITDA multiples of 10–14x suggests an infrastructure value range of ~$5.7–$8.0B for the HPC segment at maturity (ex‑Bitcoin).
• Balance sheet considerations:
• Debt-like instruments: $500M (2030) + $1.0B (2031) converts outstanding; low coupons reduce cash burden.
• Dilution mechanics: Initial conversion price ~$12.43; capped calls offset dilution up to $18.76.
• Near-term EBITDA is mining-weighted and volatile; medium-term EBITDA mix shifts sharply toward HPC as 2025–2026 capacity is commissioned.

8. Investment Thesis

Investment Rationale

• Contracted, high‑margin growth anchored by blue‑chip partners (Core42, Fluidstack; backstopped by Google) with long‑duration revenue and favorable unit economics.
• Power and scale moat: Rare combination of interconnection capacity, low-cost power, and liquid‑cooling infrastructure positions TeraWulf to capture AI demand.
• Funded to execute: Recent $1.0B convert upsized with full greenshoe plus prior $500M convert provide runway for multi‑site expansion.

Price Target Justification

• Base approach: 2027E steady-state HPC EBITDA/Site NOI ~ $450–$600M (allowing for ramp and operating overhead), at 10–12x EV/EBITDA implies EV $4.5–$7.2B for HPC, with upside from additional contracts and Cayuga.
• Equity value depends on net debt/cash at maturity and fully diluted shares (consider converts and warrants). Capped calls mitigate dilution up to $18.76. Investors should bridge to per-share outcomes using current FD share count and pro forma net debt at commissioning.

Influencing Market Dynamics

• AI infra scarcity (power + GPUs) sustains pricing and utilization; policy support for clean energy is a tailwind.
• Bitcoin price/power costs drive near-term cash generation variability but become less material as HPC mix grows.

9. Macroeconomic and Industry Trends

Regulatory Changes

• Continued emphasis on sustainable, zero‑carbon power aligns with state and federal policy priorities; long‑term ground leases (Lake Mariner and Cayuga) add siting certainty.

Supply Chain Dynamics

• GPU and electrical gear lead times remain a gating factor; TeraWulf has secured long‑lead items for near‑term builds and uses modular designs to compress schedules.

Technology Adoption Trends

• Rapid AI/GenAI adoption drives demand for high‑density, liquid‑cooled capacity with robust fiber; multi‑year contracts and backstops reflect customers’ need for committed, large‑scale capacity.