Himalayan Compute: Grand Solara Vision: Appendix B: Financial Projections

Himalayan Compute: Grand Solara Vision

 

Appendix B: Financial Projections 

Himalayan Compute: Detailed 10-Year Financial Projections (2026–2035)These projections are ambitious yet grounded in the vision from the blog posts, current AI data center market dynamics, and industry benchmarks (e.g., high CapEx intensity, rising revenue per MW for AI-optimized capacity, premium pricing for green/strategic compute, and infrastructure-style multiples). Key Assumptions

• Capacity Ramp: Starts small (permits, anchors, Phase 1) and scales aggressively via modular builds, project finance, and Nepal’s hydropower potential. Reaches ~120 GW by 2035 (ambitious but plausible if Nepal unlocks tens of GW and the company becomes a major global node; global AI capacity needs are in the hundreds of GW).
• Utilization: Ramps from 60% (early ramp-up) to 90% (mature long-term contracts).
• Revenue per MW/Year: Starts ~$8M (conservative for mixed workloads) and rises to $30M+ with premium inference, sovereign deals, managed services, and AI value-adds. Benchmarks reflect high-density AI clusters and recurring contracts.
• CapEx: ~$10–20M+ per MW all-in (power, land, cooling, GPUs, fiber; higher for AI). Use $15M/MW average. Early years more equity/VC-heavy; later debt/project finance.
• Margins: Gross 60–78% (power cost advantage + scale); EBITDA after opex ~50–60% of gross profit.
• Valuation Multiples: High early (growth/AI premium: 30–50x ARR) then moderate (20–40x) as a mature infrastructure leader. Trillion-dollar target requires massive scale + sustained high multiples driven by strategic moat and ecosystem value.
• Other: Nepal revenue retention, foundation/gov equity (dilution impact modeled at high level), no major forex shocks. All figures in USD billions unless noted.

Projected FinancialsYear | MW Deployed | Utilization | Rev/MW ($M) | Revenue ($B) | Cum. CapEx ($B) | Gross Margin % | EBITDA ($B) | Valuation ($B, illustrative)

• 2026: 50 MW | 60% | 8 | 0.24 | 0.75 | 60 | 0.10 | 7
• 2027: 200 MW | 70% | 9 | 1.26 | 3.75 | 65 | 0.57 | 44
• 2028: 500 MW | 75% | 10 | 3.75 | 11.25 | 70 | 1.84 | 150
• 2029: 1,200 MW | 80% | 12 | 11.52 | 29.25 | 72 | 5.81 | 518
• 2030: 3,000 MW (3 GW) | 85% | 15 | 38.25 | 74.25 | 75 | 20.08 | 1,913
• 2031: 7,000 MW (7 GW) | 85% | 18 | 107.10 | 179.25 | 77 | 57.73 | 4,284
• 2032: 15,000 MW (15 GW) | 90% | 22 | 297.00 | 404.25 | 78 | 162.16 | 10,395
• 2033: 30,000 MW (30 GW) | 90% | 25 | 675.00 | 854.25 | 78 | 368.55 | 20,250
• 2034: 60,000 MW (60 GW) | 90% | 28 | 1,512.00 | 1,754.25 | 78 | 825.55 | 37,800
• 2035: 120,000 MW (120 GW) | 90% | 30 | 3,240.00 | 3,554.25 | 78 | 1,769.04 | 64,800

Notes on the Table:

• Revenue grows via capacity × utilization × pricing. Early revenue from anchors/colocation; later dominated by high-margin sovereign/CaaS.
• Cumulative CapEx assumes phased buildout (total ~$3.5T by 2035 reflects enormous scale; real-world would use heavy leverage, JVs, customer pre-pays, and government support to reduce equity burden).
• Valuation: ARR multiples start premium (AI hype + growth) and normalize. By 2030–2035, $1T+ is achievable with proven execution, contracts, and Nepal’s national backing (comparable to hyperscaler valuations today, adjusted for pure-play compute upside).

Revenue Breakdown (Illustrative Evolution)

• Years 1–3: 60–70% long-term leasing/colocation; 20–30% on-demand; 10% services.
• Years 4–7: Balanced with growing sovereign zones (high-margin) and ecosystem (training, carbon credits).
• Years 8–10: Diversified recurring (70%+ contracted), value-added AI platforms, regional/secondary sites.

EBITDA Margin Path: Improves with scale, power cost advantages (hydropower), and operational efficiency (local talent + automation).
Cash Flow & Funding:

• Early years (2026–2028): Negative free cash flow due to CapEx → funded by equity raises ($100M–$few billion rounds) + strategic investors (Gulf, US, India).
• Mid-term: Project finance/debt kicks in as contracts de-risk (Power Purchase Agreements, customer anchors).
• Later: Highly cash-generative; dividends, buybacks, or reinvestment into expansion/ecosystem.
• Total funding needs: Tens to hundreds of billions equity + trillions leveraged/project (global AI infra CapEx context supports feasibility).

Sensitivity & Risks:

• Upside: Faster Nepal policy (One Desk), better GPU access, higher pricing ($40M+/MW), or global AI boom → exceed $1T earlier.
• Downside: Delays in permits/power/fiber, lower utilization (geopolitics), or competition → slower ramp (e.g., halve capacity → valuation significantly lower but still massive).
• Breakeven/Profitability: Likely by 2028–2029 at scale.
• Nepal Impact: Billions in exports, taxes, jobs; foundation 10% funds direct transfers.

These projections position Himalayan Compute as a national economic multiplier while delivering outsized investor returns. They assume flawless execution of the One Desk Policy, meritocratic talent strategy, and global demand tailwinds. Actuals would be refined with detailed modeling, audits, and phased milestones. For custom scenarios (e.g., different capacity ramps or pricing), provide additional parameters.
Himalayan Compute: Grand Solara Vision