Amazing value proposition.
— Paramendra Kumar Bhagat (@paramendra) April 24, 2026
Forget Mars, The Future Is Orbital AI https://t.co/rN0XnApaH5
So make it happen. Launch a company, and get it done.
I will be speaking at @Harvard and @MIT this weekend about why Nepal is the perfect destination for advanced compute clusters given:
— Pukar C. Hamal ๐๐ฝ ๐ (@pchamal) April 23, 2026
1️⃣ The geographic proximity to rising inference demand from 3.5 Billion people in APAC, IndoPacific and MENA and $35 Trillion+ in regional GDP!… https://t.co/PVX9JlaJzg pic.twitter.com/OUHxTQcOCQ
Nepal should be part of Pax Silica.
— Pukar C. Hamal ๐๐ฝ ๐ (@pchamal) April 23, 2026
I am hopeful our GREAT US Ambassador to India, The Honorable Sergio Gor @USAmbIndia is able to invite Nepal onto the efforts here.
It will bring lasting peace and stability to the region and it will ensure billions of people in the…
Below is a serious, unicorn-grade business plan built around the tweet’s core thesis:
Nepal can become a global destination for advanced compute clusters (AI data centers), powered by hydropower, positioned near massive inference demand, and supported by a growing labor force.
This is not a “small data center” idea. This is a geopolitical infrastructure play, like building the “Singapore of AI compute” for the Indo-Pacific.
BUSINESS PLAN: “Himalayan Compute”
The AI Power & Compute Hub of the Indo-Pacific
Goal: Unicorn valuation ($1B+) in ≤ 5 years
Mission: Build the world’s most cost-efficient, geopolitically aligned, hydropower-driven AI compute clusters in Nepal.
1. The Big Idea (What the Tweet Suggests)
The tweet is proposing that Nepal can become a strategic AI infrastructure node because:
AI inference demand is exploding across APAC, Indo-Pacific, and MENA
Nepal has cheap hydropower
Nepal has geography near billions of users
Nepal has a labor force that can scale quickly
So the business is:
Build advanced AI compute clusters (data centers optimized for GPUs) in Nepal and sell compute to:
US AI companies needing non-China regional compute
Indian AI startups and enterprises
Gulf sovereign AI initiatives
Global companies serving APAC markets
Government and defense contractors (US-aligned)
2. The Unicorn Angle (Why This Can Become $1B+)
Data center companies become unicorns quickly when they control:
power contracts
land + permits
network connectivity
GPU supply agreements
anchor customers
and can expand modularly
This is a “build once, rent forever” type of business.
If you lock in cheap energy + long-term customers, your revenues become predictable, and predictable infrastructure revenue trades at huge valuations.
A company doing $150M–$300M ARR with long-term contracts can easily justify unicorn valuation.
3. The Product
This is not one product. It’s a stack.
Product 1: AI Data Center Campuses
GPU-optimized infrastructure
Hydropower powered
Liquid cooling / immersion cooling systems
Modular container-based deployment for fast expansion
Product 2: “Compute-as-a-Service” (CaaS)
Rent GPU compute like AWS, but specialized.
Customers pay for:
inference compute
fine-tuning compute
private AI model hosting
Product 3: Sovereign AI Zones
Offer governments dedicated secure clusters:
“Nepal AI Zone”
“GCC AI Zone”
“India AI Zone”
“US-aligned compute zone”
This is where huge contracts come from.
Product 4: AI Workforce + Operations Layer
Train Nepali engineers and technicians to operate and maintain:
GPU clusters
cooling systems
cybersecurity
remote monitoring
This becomes a moat.
4. Market Opportunity
The AI compute market is exploding because:
ChatGPT-style apps require nonstop inference
every enterprise is building AI copilots
video AI, translation AI, autonomous systems require massive GPU cycles
Compute demand is not linear. It is exponential.
Even if Nepal captures 0.1% of APAC AI compute demand, that can mean hundreds of millions in annual revenue.
5. Why Nepal Specifically Works
A) Hydropower = the killer advantage
AI data centers are power-hungry.
Nepal has one of the best “renewable baseload” profiles in Asia.
Hydropower also helps cooling.
B) Geopolitical advantage
The world is decoupling from China.
US companies want “friendly compute zones” in Asia.
Nepal can position itself as:
neutral geography but aligned infrastructure
C) Proximity advantage
Latency matters for inference.
Nepal can serve:
North India (Delhi region)
Bangladesh
Pakistan
Gulf via fiber routes
Southeast Asia
D) Cheap land and labor
Compared to Singapore, Dubai, Tokyo, Seoul.
6. Competitive Landscape (Who You’re Fighting)
Direct competitors:
AWS, Azure, Google Cloud (dominant but expensive)
Oracle Cloud (cheap but smaller)
CoreWeave (AI-specific)
Gulf sovereign compute initiatives
India’s upcoming GPU clusters
Your advantage:
You’re not trying to “beat AWS globally.”
You are creating:
the cheapest AI compute node for Indo-Pacific inference workloads powered by renewable energy
That’s a unique wedge.
7. Business Model (How You Make Money)
Revenue Stream 1: Long-term GPU leasing
Sell compute contracts like:
$5M/year
$20M/year
$100M/year
These are common in AI infrastructure.
Revenue Stream 2: On-demand compute
A cloud-like marketplace:
per GPU-hour pricing
inference pricing
reserved capacity pricing
Revenue Stream 3: Colocation for AI companies
Customers bring their own GPUs, you provide:
power
cooling
physical security
network connectivity
Revenue Stream 4: Government / Defense secure clusters
Highest margin contracts.
Revenue Stream 5: Energy arbitrage + carbon credit monetization
Hydropower-powered AI compute can earn carbon advantage.
8. Go-To-Market Strategy
This is a B2B infrastructure company.
You don’t grow by ads. You grow by anchor contracts.
Step 1: Secure one US “flagship” partner
Target:
OpenAI partners
Anthropic partners
Microsoft ecosystem companies
US defense contractors
AI inference providers
Step 2: Secure India enterprise contracts
Target:
Reliance Jio
Tata
Infosys/Wipro
Indian banks
Indian government cloud
Step 3: Secure Gulf sovereign AI demand
Target:
UAE AI initiatives
Saudi Vision 2030 compute demand
Qatar AI ecosystem
This is the highest-budget buyer group on earth right now.
9. The “Unicorn Moat” (What Stops Copycats)
If someone says “India can do this too,” yes, but your moat is:
Moat 1: Locked hydropower contracts
Secure 15–25 year power purchase agreements.
Moat 2: First-mover permits + land bank
Acquire and pre-permit sites near hydropower substations.
Moat 3: Fiber + connectivity deals
Negotiate high-bandwidth fiber links to India and global routes.
Moat 4: GPU supply contracts
Partner with NVIDIA ecosystem, AMD, and hyperscaler supply chains.
Moat 5: Talent pipeline
Build “Nepal AI Infrastructure Academy” producing technicians at scale.
Moat 6: Political alignment + trust
Market Nepal as “trusted neutral compute zone.”
10. Execution Roadmap (5-Year Plan)
Year 1: Foundation + Flagship Build
raise $50M–$150M seed + Series A (infrastructure-heavy)
secure hydropower PPAs
acquire land + permits
sign 1 anchor customer
build Phase 1 cluster (10–30 MW)
Target: $10M–$30M revenue run-rate
Year 2: Expand + Sell to India
expand to 50 MW
sign Indian enterprise customers
deploy GPU marketplace platform
hire 300+ Nepali technical workforce
Target: $50M ARR
Year 3: Regional Compute Hub
expand to 100–150 MW
build second site
add sovereign AI zones
secure Gulf contracts
Target: $150M–$250M ARR
Year 4: Become the “AI Singapore” of the Himalayas
expand to 300 MW+
offer AI model hosting + inference products
partner with major cloud vendor for hybrid offering
Target: $400M ARR
Year 5: IPO-ready infrastructure unicorn
500 MW+ planned pipeline
multi-site campus system
government-grade contracts
predictable long-term revenues
Target: $700M–$1B ARR potential
At that point, valuation can exceed $5B–$15B depending on contracts.
11. Funding Strategy (How You Finance This)
This cannot be funded like a normal startup.
You need a hybrid structure:
Phase 1: Venture + Strategic Capital
Raise from:
US VC funds with infrastructure focus
sovereign wealth funds
defense-tech aligned investors
Indian conglomerates
Phase 2: Project Finance
Once contracts exist, use:
debt financing
infrastructure funds
export credit agencies
World Bank-style development financing
This allows you to scale massively without dilution.
12. Key Partnerships Required
Hydropower producers (Nepal’s core asset)
Fiber connectivity providers
NVIDIA / GPU suppliers
Cooling technology companies
US AI cloud platforms
Indian telecom operators
Government of Nepal (fast-track permits, tax incentives)
13. Biggest Risks (And How You Solve Them)
Risk 1: Nepal political instability
Solution: Create a legally protected “AI Special Economic Zone.”
Risk 2: Connectivity bottlenecks
Solution: Invest early in fiber redundancy to India.
Risk 3: GPU supply shortage
Solution: secure long-term supply agreements, use diversified GPU sources.
Risk 4: Perception issue (“Nepal isn’t stable enough”)
Solution: credibility via US and MIT/Harvard network + global board members.
Risk 5: India builds its own cheaper clusters
Solution: sell Nepal as “hydropower + neutrality + export compute node,” not competing directly with India.
14. What the Company Should Actually Be Called
Brand matters for geopolitics.
Strong name options:
Himalayan Compute
Everest Cloud
HydroAI Grid
NepalCompute
Sagarmatha AI Infrastructure
15. The Killer Pitch (One Paragraph)
Nepal is sitting on the most underpriced strategic asset in Asia: renewable hydropower. We will convert that power into advanced AI compute clusters serving the fastest-growing AI markets in the Indo-Pacific and MENA. By locking in long-term power contracts, building GPU-optimized data center campuses, and selling compute to US-aligned AI companies and regional enterprises, we will become the Singapore of AI infrastructure—an export engine for Nepal and a global node in the US-led AI stack.
16. What Makes This a “Win-Win-Win” (as the Tweet Says)
Win for Nepal
billions in export revenue
jobs and tech ecosystem
global strategic relevance
Win for the US
trusted compute base in Asia
alternative to China-linked infrastructure
strengthens Indo-Pacific alignment
Win for the World
cheaper AI compute
more decentralized AI infrastructure
clean-energy-driven AI growth
17. The Unicorn Milestone Metric
The metric that matters most is:
Megawatts deployed under contract
If you have:
100 MW deployed
with long-term contracts
with 70% utilization
You are already on a unicorn trajectory.
Final Summary
This tweet is essentially proposing:
Nepal should become an AI energy-to-compute exporter.
The unicorn company is the execution vehicle:
lock hydropower
build GPU campuses
sell compute to APAC + MENA markets
scale via project finance
become the Indo-Pacific’s AI infrastructure hub
If executed properly, this is not merely a unicorn.
This is a $10B+ infrastructure empire in the making.
Investor Pitch Deck Outline (15 Slides)
Himalayan Compute: Nepal’s Hydropower-Powered AI Compute Hub
Slide 1 — Title
Himalayan Compute
The AI Infrastructure Hub of the Indo-Pacific
Clean energy → AI compute → global export revenue
Raising: Seed / Series A ($X)
Founder + key partners
Tagline: “Hydropower into intelligence.”
Slide 2 — The Problem
AI demand is exploding, but the world lacks:
affordable GPU compute
clean reliable power
low-latency regional inference hubs
geopolitically trusted infrastructure in Asia
scalable data center sites outside China-linked ecosystems
Compute is the new oil. Power is the bottleneck.
Slide 3 — The Market Opportunity
AI compute demand is growing exponentially across:
APAC + Indo-Pacific + MENA
3.5B+ population region
$35T+ GDP footprint
Key drivers:
inference at scale (copilots, translation, search, video AI)
sovereign AI initiatives
enterprise fine-tuning adoption
TAM: Multi-trillion-dollar AI infrastructure buildout.
Slide 4 — The Solution
Build GPU-optimized advanced compute clusters in Nepal powered by hydropower.
Deliver:
low-cost AI compute at scale
AI colocation + leasing
sovereign secure clusters
high-bandwidth regional inference
Nepal becomes the “Singapore of AI Compute.”
Slide 5 — Why Nepal Wins
Nepal has a rare convergence of advantages:
Cheap hydropower (renewable baseload)
Cooling efficiency (geography + hydrology)
Proximity to India + APAC inference demand
Lower land and labor cost than regional hubs
Strategic neutrality + global trust potential
Energy + geography + timing = once-in-a-century advantage.
Slide 6 — Product Offering
Four products under one platform:
AI Data Center Campuses (10–500 MW modular scale)
Compute-as-a-Service Marketplace (GPU-hour + inference pricing)
Sovereign AI Zones (government / defense-grade secure clusters)
AI Operations Workforce Pipeline (training + staffing)
Slide 7 — Business Model
Revenue streams:
Long-term compute leases (multi-year contracts)
GPU colocation (bring-your-own-GPU)
On-demand cloud GPU marketplace
Sovereign AI clusters (premium pricing)
Managed inference hosting
Carbon advantage / clean compute premium
Recurring infrastructure revenue + high utilization flywheel.
Slide 8 — Competitive Landscape
Competitors:
AWS / Azure / Google (expensive, congested)
CoreWeave (US-centric)
Gulf sovereign clusters
Indian GPU initiatives
Our wedge:
cheaper power
clean compute
regional latency
geopolitical trust
first-mover infrastructure + land + permits
Positioning: AI compute export economy.
Slide 9 — Moat / Defensibility
Six barriers competitors cannot easily replicate:
Locked 15–25 year hydropower PPAs
Pre-permitted AI Special Economic Zone
Fiber redundancy into India and global routes
GPU supply partnerships
Operational expertise in high-density GPU cooling
Nepal-based talent pipeline + workforce academy
We are not a data center. We are an energy-to-compute platform.
Slide 10 — Go-To-Market Strategy
Enterprise-first, anchor-contract strategy:
Phase 1: US-aligned AI companies
Phase 2: Indian enterprises + telecoms
Phase 3: Gulf sovereign AI initiatives
Phase 4: Regional marketplace (SMEs + developers)
Distribution:
direct enterprise sales
strategic partnerships with cloud + telecom operators
government-to-government facilitation
Slide 11 — Traction & Pipeline (What We Will Show Investors)
Include metrics such as:
signed LOIs for power and land
government approvals underway
preliminary anchor customer discussions
engineering and EPC partner onboarded
early MoUs with fiber providers
advisory board (ex-hyperscaler / defense / infra experts)
This slide becomes the credibility builder.
Slide 12 — Technology & Infrastructure Plan
Core technical architecture:
NVIDIA/AMD GPU clusters
liquid cooling / immersion cooling
modular container deployment
Tier III+ reliability
cybersecurity + secure enclave offerings
AI workload scheduling platform (marketplace layer)
Build fast, scale modularly, operate like hyperscalers.
Slide 13 — Financial Model (5-Year Snapshot)
Show investor-ready high-level numbers:
MW deployed year-by-year
utilization assumptions
revenue per MW
gross margin targets
ARR targets
Example targets:
Year 2: $50M ARR
Year 3: $150–250M ARR
Year 5: $700M+ ARR potential
Key message: high recurring revenue + infrastructure multiples.
Slide 14 — Funding Ask & Use of Funds
Raising: $X Seed / Series A
Use of funds:
land acquisition + permits
hydropower PPA deposits
first 10–30 MW buildout
fiber connectivity build
security + compliance
engineering team + operations academy
GPU procurement strategy
Goal: Phase 1 operational within 12–18 months.
Slide 15 — The Vision + Exit
Vision: Nepal becomes a global AI export powerhouse.
Outcomes:
the Indo-Pacific’s trusted AI compute hub
carbon-friendly compute leader
sovereign AI partner to multiple nations
Exit paths:
IPO as infrastructure/AI cloud company
acquisition by hyperscaler
strategic merger with global data center operator
Closing line:
“Hydropower is Nepal’s oil. We turn it into intelligence.”
Below is a detailed, investor-grade financial projection model for Himalayan Compute over 5 years, with assumptions clearly stated. These are not “fantasy numbers”—they are structured the way infrastructure investors expect.
I’ll give you a full 5-year projection, including capex, revenue, gross margin, EBITDA, cash needs, and valuation logic.
1. Key Assumptions (Base Case)
A) Capacity Buildout (MW Deployed)
We assume modular expansion.
| Year | IT Load (MW) Deployed | Total Facility Power (MW) |
|---|---|---|
| Y1 | 20 MW | 30 MW |
| Y2 | 60 MW | 90 MW |
| Y3 | 150 MW | 225 MW |
| Y4 | 300 MW | 450 MW |
| Y5 | 500 MW | 750 MW |
IT Load MW = actual compute power for GPUs
Facility MW = IT load + cooling + overhead
Typical ratio: PUE ~ 1.4–1.6 (we assume 1.5)
B) Capex Cost per MW (GPU-Optimized Facility)
Data centers vary widely depending on tier, land, cooling, and power systems.
We assume:
$8M per IT MW for full buildout (data center + electrical + cooling)
plus incremental fiber and security costs
This is aggressive but plausible in Nepal due to low land and labor costs.
So:
Capex = IT MW × $8M
C) Revenue per IT MW
Revenue depends on GPU density and contract pricing.
We model two revenue streams:
Stream 1: Long-Term Enterprise Contracts (70% of capacity)
3–7 year contracts
stable utilization
premium pricing
Stream 2: Spot / Marketplace Compute (30% of capacity)
higher margins
variable utilization
We assume blended annual revenue per IT MW:
$2.2M revenue per MW per year at full utilization
This is reasonable because 1 MW of GPU compute can generate $2M–$4M annually depending on pricing and utilization.
D) Utilization Ramp
| Year | Utilization |
|---|---|
| Y1 | 45% |
| Y2 | 60% |
| Y3 | 70% |
| Y4 | 75% |
| Y5 | 80% |
E) Power Cost
Hydropower advantage is central.
Assume:
$0.035 per kWh average industrial rate (very competitive)
effective all-in delivered power cost: $0.045/kWh after transmission, redundancy
F) Gross Margin
Compute/data center operators often run gross margins 40–70%.
We assume:
Year 1: 35% (ramp inefficiency)
Year 2: 45%
Year 3: 55%
Year 4: 60%
Year 5: 62%
G) Operating Expenses (Opex)
Opex includes:
staff
security
compliance
sales
network
insurance
admin
We assume opex as % of revenue:
| Year | Opex % of Revenue |
|---|---|
| Y1 | 35% |
| Y2 | 25% |
| Y3 | 18% |
| Y4 | 15% |
| Y5 | 13% |
This is realistic because infrastructure scales well.
2. Capex Plan (5 Years)
Capex = IT MW added × $8M
| Year | IT MW Added | Capex ($M) |
|---|---|---|
| Y1 | 20 | 160 |
| Y2 | 40 | 320 |
| Y3 | 90 | 720 |
| Y4 | 150 | 1,200 |
| Y5 | 200 | 1,600 |
| Total | 500 MW | $4.0B |
Yes, the capex is enormous. That’s why the model requires project finance and debt after early traction.
This is not a pure venture model.
It becomes an infrastructure roll-out.
3. Revenue Projections
Revenue formula:
Revenue = IT MW deployed × revenue/MW × utilization
Where revenue/MW at 100% utilization = $2.2M
Year-by-Year Revenue
Year 1
MW: 20
Utilization: 45%
Revenue = 20 × $2.2M × 0.45
= $19.8M
Year 2
MW: 60
Utilization: 60%
Revenue = 60 × $2.2M × 0.60
= $79.2M
Year 3
MW: 150
Utilization: 70%
Revenue = 150 × $2.2M × 0.70
= $231.0M
Year 4
MW: 300
Utilization: 75%
Revenue = 300 × $2.2M × 0.75
= $495.0M
Year 5
MW: 500
Utilization: 80%
Revenue = 500 × $2.2M × 0.80
= $880.0M
4. Gross Profit Projections
| Year | Revenue ($M) | Gross Margin | Gross Profit ($M) |
|---|---|---|---|
| Y1 | 19.8 | 35% | 6.9 |
| Y2 | 79.2 | 45% | 35.6 |
| Y3 | 231.0 | 55% | 127.1 |
| Y4 | 495.0 | 60% | 297.0 |
| Y5 | 880.0 | 62% | 545.6 |
5. Operating Expense & EBITDA Projections
Opex = revenue × opex %
| Year | Revenue ($M) | Opex % | Opex ($M) |
|---|---|---|---|
| Y1 | 19.8 | 35% | 6.9 |
| Y2 | 79.2 | 25% | 19.8 |
| Y3 | 231.0 | 18% | 41.6 |
| Y4 | 495.0 | 15% | 74.3 |
| Y5 | 880.0 | 13% | 114.4 |
EBITDA = Gross Profit – Opex
| Year | Gross Profit ($M) | Opex ($M) | EBITDA ($M) |
|---|---|---|---|
| Y1 | 6.9 | 6.9 | 0.0 |
| Y2 | 35.6 | 19.8 | 15.8 |
| Y3 | 127.1 | 41.6 | 85.5 |
| Y4 | 297.0 | 74.3 | 222.7 |
| Y5 | 545.6 | 114.4 | 431.2 |
By Year 5, EBITDA is $431M which is absolutely unicorn-level infrastructure profitability.
6. Cash Flow and Financing Requirements
This business is capex-heavy, so it must evolve into project finance quickly.
Total Capex over 5 years = $4B
You cannot raise $4B in VC equity.
Instead, you structure it as:
10–20% equity
80–90% debt/project finance once contracts are signed
Capital Stack Assumption (Typical Infrastructure)
| Source | % |
|---|---|
| Equity | 20% |
| Debt / Project Finance | 80% |
So required equity over 5 years:
$4.0B × 20% = $800M equity
Required debt:
$4.0B × 80% = $3.2B debt
This is realistic if contracts are in place.
7. Funding Roadmap (Equity Rounds)
Seed / Series A (Year 1)
Raise: $75M–$150M
Use:
land + permits
engineering
first 20MW facility
fiber backbone + redundancy
Series B (Year 2)
Raise: $200M–$300M
Use:
expand to 60MW
GPU procurement partnerships
secure anchor contracts
Series C (Year 3)
Raise: $400M–$600M
Use:
expand to 150MW
sovereign AI zone buildouts
After this point, most expansion becomes debt-financed.
8. Unit Economics Per MW (Investor Critical)
Revenue per IT MW at 75% utilization:
$2.2M × 0.75 = $1.65M per MW per year
Gross profit per MW at 60% gross margin:
$1.65M × 0.60 = $0.99M per MW per year
Payback Period
Capex per MW = $8M
Gross profit per MW = ~$1M/year
Payback = 8 years (gross profit basis)
But once utilization rises and pricing improves, payback becomes 5–6 years, which is strong for infrastructure.
9. Scenario Model (Bear, Base, Bull)
Bear Case
Revenue per MW: $1.5M
Utilization capped at 60%
slower demand
Year 5 revenue:
500 × 1.5 × 0.60 = $450M
Still strong.
Base Case (Model Above)
Year 5 revenue: $880M
Bull Case
Revenue per MW: $3M
Utilization 85%
Year 5 revenue:
500 × 3 × 0.85 = $1.275B
That becomes a multi-unicorn.
10. Valuation Projection (How You Hit Unicorn)
Infrastructure compute companies trade on:
ARR multiples (8x–20x)
EBITDA multiples (15x–30x)
or MW deployed + contracted revenue
Year 3 Valuation Potential
Revenue: $231M
If valued at 8x revenue:
Valuation = $1.85B
You hit unicorn by Year 3 if execution is credible.
Year 5 Valuation Potential
EBITDA: $431M
If valued at 20x EBITDA:
Valuation = $8.6B
Or if valued at 10x revenue:
$880M × 10 = $8.8B
This becomes a $5B–$15B company depending on contract quality.
11. Cost Breakdown (What Drives Gross Margin)
Key expenses:
power (largest variable cost)
cooling systems
maintenance contracts
network bandwidth
staffing + security
insurance + compliance
depreciation
Hydropower makes your cost structure structurally superior.
If competitors pay $0.08–$0.12/kWh and you pay $0.04–$0.05, you win permanently.
12. What Investors Want to See to Believe This Model
To justify these projections, the company must prove:
Power Purchase Agreements (PPAs) locked for 15–25 years
Government support: AI Special Economic Zone
Fiber redundancy to India + global networks
Anchor contracts signed (even 1–2)
EPC partners with data center track record
GPU supply strategy (NVIDIA ecosystem partnerships)
Clear compliance posture (US trust + security)
13. Final 5-Year Projection Summary Table
| Year | IT MW | Utilization | Revenue ($M) | EBITDA ($M) | Capex ($M) |
|---|---|---|---|---|---|
| Y1 | 20 | 45% | 19.8 | 0.0 | 160 |
| Y2 | 60 | 60% | 79.2 | 15.8 | 320 |
| Y3 | 150 | 70% | 231.0 | 85.5 | 720 |
| Y4 | 300 | 75% | 495.0 | 222.7 | 1,200 |
| Y5 | 500 | 80% | 880.0 | 431.2 | 1,600 |
14. The One-Line Investor Conclusion
If Himalayan Compute secures cheap hydropower PPAs and signs multi-year AI compute contracts early, it becomes a predictable infrastructure cash machine—capable of reaching unicorn valuation by Year 3 and $5B–$10B by Year 5.
BUSINESS PLAN: “Himalayan Compute” https://t.co/qyGKeC9iqh @SwarnimWagle@shisir @DrSJaishankar @MEAIndia @ShahBalen @narendramodi @hamrorabi @DilBhusanPathak@kathmandupost @palkisu @republic @PM_nepal_@PMOIndia @JM_Scindia @NitishKumar @BJP4Bihar ๐
— Paramendra Kumar Bhagat (@paramendra) April 24, 2026
๐️ Himalayan Compute: Nepal’s Global AI Infrastructure Strategy https://t.co/a9Apk3txb4
— Paramendra Kumar Bhagat (@paramendra) April 24, 2026
๐️ เคนिเคฎाเคฒเคฏเคจ เคเคฎ्เคช्เคฏुเค: เคจेเคชाเคฒเคो เคเคเค เคชूเคฐ्เคตाเคงाเคฐ เคฐเคฃเคจीเคคि https://t.co/BPfTXRt7xk
— Paramendra Kumar Bhagat (@paramendra) April 25, 2026
