Slide 1: Renewable Computing
Power is the asset. Compute follows. Renewable Computing. Nasdaq: SLNH 1
Slide 2: Legal Disclosure & Disclaimer
Legal Disclosure & Disclaimer This presentation includes forward-looking statements within the meaning of the Private Securities Litigation Reform Act that reflect our current views with respect to, among other things, our operations, business strategy, interpretation of prior development activities,
Slide 3: Investment Thesis
INVESTMENT THESIS Long-term, behind-the- Power is the primary constraint in the AI era. Soluna meter power access at has secured long-term, behind-the-meter access to it at scale. We convert that access into contracted the speed and scale AI data center cash flows by building directly on site of dema
Slide 4: Company Overview
EXECUTIVE SUMMARY Company Overview + AI is driving Miners prefer Soluna bridges Clean energy goes to waste due to exponentially scalable, well- this gap curtailment growing demand managed, and for compute cost-advantaged hosting partners. Co-locating data centers behind the meter at renewable power
Slide 5: Current & Future Revenue Streams
Current & AI / High Performance Grid Ancillary Future Computing Services Revenue • Colocation and hosting services for companies that need AI-ready data centers. • Compensation to act as behind-the-meter flexible load for the grid Streams • Soluna develops data centers with JV partners and provides M
Slide 6: Soluna Organizational Structure
Soluna Organizational Structure Soluna Holdings (Nasdaq: SLNH) Soluna Energy Soluna Wind Soluna Digital Soluna HPC The development platform that The energy business that The Bitcoin Hosting business The AI business that builds and identifies and acquires acquires and operates wind that builds, operat
Slide 7: 2026 Corporate Focus
2026 Corporate Focus Develop Optimize Capital Grow AI Projects Formation Pipeline Advance Project Kati 2 with Energize Project Kati 1. Executing a disciplined capital Expand Soluna’s Renewable Metrobloks to shovel-ready Enhance profitability across strategy to fund pipeline Computing(™) pipeline by a
Slide 8: Growth. The next 1-3 years
Growth. The next 1-3 years Near-term Pipeline Interconnection Land Progression Progress Acquisition Kati 2 JV with Metrobloks, 29 commercial projects 9 projects totaling 914 15 projects totaling 1,719 350+ MW; Dorothy 3 and totaling 3,156 MW actively MW in active interconnection MW in active land ac
Slide 9: Soluna Development Pipeline as of August 1, 2026
Soluna Development Pipeline as of August 1, 2026 7000 MW 1.6GW 14MW 192MW ~6.3 GW+ 6000 MW 5000 MW 4.5GW 4000 MW 3000 MW 2000 MW 1000 MW 0MW Assessed Planning & Development Under Construction Operating Total Pipeline Curtailment assessments PPA negotiations or signed, Project Kati 1 is actively Proj
Slide 10: Our Data Center Projects
Our Data Center Projects We have nearly 2GW of data center capacity in operation, construction or development Project Location Power Source Size (MW) Model Status 1 Power Cost Partner(s) Dorothy 1A TX Wind 25 BTC Hosting Operating $35 N/A Dorothy 1B TX Wind 25 BTC Mining Operating $35 N/A Sophie KY
Slide 11: Our AI/HPC Projects in Development
Our AI/HPC Projects in Development We have over 1.6GW of data center capacity in operation, construction or development for AI/HPC Project Name Kati 2 Dorothy 3 Rosa Hedy Ellen Fei Gladys Total Size in MWs 350+ 300+ 242 198 145 240 150 1625+ Phase 1 / Development Site selection & land control Power
Slide 12: Pipeline & Distinctives
Pipeline & Distinctives 12 | Nasdaq: SLNH
Slide 13: The Perfect Storm: Driving Energy Demand
The Perfect Storm: Driving Energy Demand Historic Load Infrastructure Grid Reliability Energy Arms Spike 1,2 Bottlenecks 3,4,5 Strained 6,7,8,9 Race 1,2,6 Explosive load growth: U.S. Interconnection queues exceed DOE (2025) and NERC (2024 LTRA) Unprecedented scale: CapEx is utilities now forecast 50
Slide 14: How We Source Power for Our Data Centers
How We Source Power for Our Data Centers CURTAILED Economic/Technical curtailment from negative LMP. ENERGY DC * Energy that Energy that is would have gone SUBTRACTIVE GRID sourced from the to grid without grid in excess of the load. ENERGY ENERGY plant production. * Soluna AI Data Center. 14 | Nasd
Slide 15: Rapid Time to Interconnection
Rapid Time to Interconnection Behind-the-Meter design delivers firm power for AI workloads. Renewable-first, with grid and on-site generation for firming. EXISTING RENEWABLE POWER PROJECT ASSETS GRID OPERATOR TELEMETRY PROJECT METER HV MV METER METER ERCOT GRID SUBSTATION GENERATION ASSET MDC METER SOL
Slide 16: How We Achieve Tier-3 Uptime Behind the Meter
How We Achieve Grid Energy Tier-3 Uptime 39.200% Behind the Meter DATA Resiliency CENTER Other Backup 2.000% UPTIME 1.982% 99.982% Downtime 0.018% Wind Energy 58.800% 16 | Nasdaq: SLNH
Slide 17: Our Power Cost Is Among the Lowest in the Industry
Our Power Cost Is Industry Soluna Among the Lowest $80 $80 In the Industry $70 $60 $50 $50 $40 $40 $30 $30 $20 $10 $0 Bitcoin 1 AI 2 (1) Luxor Research; Public filings from various miners (2) EIA.gov | https://www.eia.gov/electricity/monthly/epm_table_grapher.php?t=epmt_5_6_a 17 | Nasdaq: SLNH
Slide 18: Maestro OS Is Our Force Multiplier
Maestro OS Is Our TM Force Multiplier Our proprietary software streamlines site operations with intelligent monitoring, precise data, and full-stack automation—enabling faster decisions, higher uptime, and seamless grid integration. Control Data Collection Operations Power Extend equipment lifespan
Slide 19: Soluna's Distinctiveness
Soluna’s Distinctiveness Behind the Meter (BTM) BTM model enables Flexible power design Proprietary software and model captures accelerated access to with redundant options for proven execution stranded energy and low cost power power eases grid congestion Monetizing stranded power Using existing su
Slide 20: Artificial Intelligence
Artificial Intelligence 20 | Nasdaq: SLNH
Slide 21: AI Campus at Project Kati 2
AI CAMPUS AT PROJECTKATI 2 JOINT VENTURE A co-development JV combining Soluna's renewable-powered Texas campus METROBLOKS BRINGS with Metrobloks' hyperscale AI data center design and operations platform. • AI-ready data center design & operations • Pre-lease & customer engagement • Day-to-day site m
Slide 22: 22 | Nasdaq: SLNH
22 | Nasdaq: SLNH
Slide 23: 300MW+ Tier III AI Campus Plan
300MW+ Tier III AI Campus Plan Phased Onsite Power Renewable Grid Development Generation Clustering Stabilization Plan Strategy Infrastructure Kati-2 is planned as a two-phase, The campus power plan is Grid power for the campus is The campus design includes a AI-optimized data center campus expected
Slide 24: Scaling Renewable Computing to power the next generation of infrastructure
Scaling Renewable Computing to power Renewable Computing is Soluna’s advantage — aligning energy and the next generation of infrastructure compute to scale AI infrastructure faster and more efficiently. TOTAL PIPELINE 6.3GW+ = AI PIPELINE Kati 2 Dorothy 3 Greenfield Sites Up to 300 MW+ Up to 300 MW+ M
Slide 25: Project Dorothy 3
FUNDAMENTALS Project Dorothy 3 TARGET CAPACITY POWER SOURCE 150-300 MW Wind LINE OF BUSINESS LAND AI 300 Acres Expansion of Renewable-Powered AI Infrastructure With full ownership of the 150 MW wind farm, ERCOT grid access, and additional contracted sources, Dorothy 3 is expected to support 150 MW+
Slide 26: Company Overview
Company Overview 26 | Nasdaq: SLNH
Slide 27: Overlapping challenges
3 1 Constrained Grid overlapping 2 3 Exploding challenges Compute Imbalanced Demand Energy Supply- Demand (1) The Computational Limits of Deep Learning, Thompson N et Al - 2022 27 | Nasdaq: SLNH
Slide 28: Constrained Grid
1 / CONSTRAINED GRID Historic Infrastructure Grid Reliability Load Spike 1,2 Bottlenecks 3,4,5 Strained 6,7,8,9 National interconnection queues exceed DOE (2025) and NERC 2.6TW (2024 LTRA) warn of rising 50GW reliability risks without of new generation Permitting and build times MAJOR NEW needed by
Slide 29: Exploding Compute Demand
2 / EXPLODING COMPUTE DEMAND 100M The energy required to run AI tasks is accelerating, Llama 3 GPT-4 10M with an annual growth rate between 26% and 36% GPT-3 Total power draw required 1M AI models 100k Power draw 10k 1k 100 2012 2014 2016 2018 2020 2022 2024 2026 Source: Epoch AI. https://epoch.ai/d
Slide 30: Imbalanced Energy Supply-Demand
3 / IMBALANCED ENERGY SUPPLY-DEMAND 30 40 % of energy produced by renewable plants goes unused. Source: Soluna Curtailment Assessments of IPPs in Pipeline. Curtailment estimates from ISO/RTO websites. Wood Mackenzie. 30 | Nasdaq: SLNH
Slide 31: Renewable Energy Has a Wasted Energy Opportunity
RENEWABLE ENERGY HAS A WASTED ENERGY OPPORTUNITY 31 | Nasdaq: SLNH
Slide 32: Design Compute Around Power Availability
DESIGN COMPUTE AROUND POWER AVAILABILITY — NOT THE OTHER WAY AROUND 32 | Nasdaq: SLNH
Slide 33: All forms of renewable energy are growing faster than ever
GW All forms of 12,000 renewable energy are growing faster than ever. 6,000 Tripling in 2030 Accelerated Case 2030 Main Case 2030 0 2022 2023 2024 2025 2026E 2027E 2028E 2029E 2030E Source: IEA data - https://www.iea.org/reports/renewables-2025/executive-summary 33 | Nasdaq: SLNH
Slide 34: Soluna is the power-compute convergence
Soluna is the power<>compute convergence – using curtailed renewable energy to power advanced computing infrastructure. 34 | Nasdaq: SLNH
Slide 35: Renewable Computing
To reach their Renewable full potential AI Energy has a & HPC need wasted energy a sustainable opportunity energy source RENEWABLE COMPUTING: Using compute itself to bridge the supply gap needed to power massively increased compute. 35 | Nasdaq: SLNH
Slide 36: Renewables are a perfect match for AI
Renewables are a perfect match for AI Renewables are a perfect match for AI Many AI workloads can be paused Renewables AI or time-shifted, allowing compute to run when wind or solar is producing instead of forcing constant grid demand. 36 | Nasdaq: SLNH
Slide 37: Case Study: Texas Wind Farm Curtailment
CASE STUDY Texas Wind Farm Curtailment 150 MW Wind Generation Generation Curtailed 80% 60% 73% 66% 57.3% 59.7% 40% 42.7% 40.3% 20% 34% 27% 0% Source: Soluna Data Analysis, Wind Farm Data 2021 2022 2023 2024 37 | Nasdaq: SLNH
Slide 38: Case Study: Soluna Converts ~50% of Curtailed Energy to Compute
CASE STUDY Soluna Converts ~50% of Curtailed Energy to Compute 50 MW Data Center – Project Dorothy 1 Converted by Soluna Curtailed 80% 60% 40% 42.7% 40.3% 20% 34% 27% 22.6% 20.6% 16.9% 14% 0% 2021 2022 2023 2024 Source: Soluna Data Analysis, Wind Farm Data 38 | Nasdaq: SLNH
Slide 39: With its concentration of renewables, Texas is the beginning
With its concentration of renewables, Texas is the beginning… 39 | Nasdaq: SLNH
Slide 40: Renewable power plants generate TWh/yr of wasted energy
Renewable power plants in the U.S. alone generate 20 TWh/yr of wasted renewable energy. Grid-scale Renewable Power Plant Source: https://www.spglobal.com/market-intelligence/en/news-insights/research/2025/10/grid -congestion-remains-key-issue-as-data-center-load-growth-stresses-system 40 | Nasdaq: S
Slide 41: Soluna's solution for grid and data center decarbonization
Soluna’s solution for grid and data center -25% decarbonization As new data-heavy applications drive massive energy demand, most data centers rely on carbon-intensive grids to power these workloads. While RECs are widely used to offset emissions, they often fail to reflect real-time energy usage and
Slide 42: The Soluna Way
Source The low-cost curtailed Soluna power from IPPs Way Grow Build, own, pipeline operate data We tackle wasted energy centers through digital infrastructure. As we optimize the grid and serve our customers, we fuel our growth, funding further expansion to make renewable energy a superpower. Grow A
Slide 43: What Our Customers Are Saying
What Our Customers Are Saying “Our partnership with Soluna is another “Partnering with Soluna has been a “Blockware is proud to continue significant step in advancing our North game-changer for Compass. Their building alongside Soluna, expanding American self-mining strategy. Each of commitment to op
Slide 44: Meet the Soluna Leadership Team
Meet the Soluna Leadership Team 150 years of combined experience in starting, managing, and leading companies John Belizaire Michael Toporek Michael Picchi Jessica Thomas Mary O’Reilly Ryan Carver Chief Executive Officer Executive Chairman Chief Financial Officer Chief Accounting Officer Chief People O
Slide 45: Learn more at solunacomputing.com
Learn more at solunacomputing.com 45 45 | Nasdaq: | Nasdaq: SLNH SLNH
Slide 46: Connect With Us
Connect With Us X / Twitter LinkedIn Newsletter @SolunaHoldings Soluna Holdings bit.ly/solunasubscribe 46 | Nasdaq: SLNH
Slide 47: Appendix
Appendix 47 | Nasdaq: SLNH
Slide 48: For Soluna's Q2 2026 results and update, click here
For Soluna’s Q2 2026 = results and update, click here 48 | Nasdaq: SLNH
Slide 49: Existing Debt by Entity (as of June 30, 2026)
Existing Debt by Entity (as of June 30, 2026) $27.1M subsidiary (no Soluna Holdings guarantee) $6M GreenCloud debt (with Soluna Holdings guarantee) Soluna Holdings, Inc. (SLNH)1 Soluna Wind Soluna Digital, Inc (SDI)2 Soluna Cloud, Inc (SCI) Holdings, Inc. (SWH) Soluna DVSL Soluna DV Wind Soluna SW,
Slide 50: AI / HPC Data Center Project ROIC
1 AI / HPC Data Center Project ROIC 2 YR 1 YR 2 YR 3 ● Capital is returned at exit 2,100,000 ● 34.8% IRR3 34.8% ● 145% ROIC2 IRR3 CAPEX 158,000 158,000 145% Net Cash Flow4 ROIC2 (1,400,000) (1) Key assumptions: (a) Assumes ~15-month build period with 3-year hold and exit at YR3, (b) Returns shown on