Project Budget

TL;DR

  • Total budget: $200–490 billion (with inflation $250–610 billion over 10 years)
  • Import from Earth: ~2,550 t electronics (~100 launches)

Currency and Inflation

All amounts are in 2026 US dollars (USD 2026).

Historical projects are adjusted for inflation. For a 10-year project we account for:

  • Average US inflation (2015-2026): ~2.5%/year
  • Over 10 years: (1.023)^10 = 1.255 ≈ +25%
  • This is a conservative estimate; actual inflation may be higher
Scenario 2026 Nominal With 10-year inflation
Optimistic ~$200 billion ~$250 billion
Baseline ~$355 billion ~$445 billion
Conservative ~$490 billion ~$610 billion

International References (2026)

Launch Costs ($/kg to LEO)

Country Vehicle $/kg Source
USA SpaceX Falcon 9 $2,700-5,000 Wikipedia
China Long March 5 $3,000 China-in-Space
China Long March 8 $3,550 NextSpaceflight
Russia Soyuz-2 $3,800-5,300 TASS
India PSLV $8,400-15,000 SpaceTechAsia
Future Starship, Long March 9, Yenisei $1,000-1,500 2030+ projections

For calculations: $2,500/kg (mix of current and future technologies)

Rocket Production Costs

Vehicle Unit Price Source
Starship ~$100M (estimate) SpaceX
Falcon 9 ~$50M SpaceX
Long March $30-50M CASC
Soyuz-2 ~$35M Roscosmos

For calculations: $50M/rocket (international fleet)

AI Development (2026 references)

Company Budget/year Employees Total spent Source
OpenAI $7 billion 1,500 ~$20+ billion DCD
DeepMind $12 billion ~3,000 ~$50+ billion Wikipedia
Waymo ~$2 billion 1,500 $11 billion over 15 years VentureBeat
Tesla FSD $10 billion 800+ ~$15 billion over 8 years CleanTechnica

AI specialist salaries: - OpenAI median: $875K/year - Top researchers: $10-20M/year - International teams (China/Russia): $100-200K/year

Construction in China (references)

Project Cost Area Source
BYD Shenzhen Phase 2 $2.87 billion 3.79 million m² SCMP
Tesla Gigafactory Shanghai ~$2 billion 860,000 m² Public data
Average construction cost ~$580/m² — Statista

Base Calculations (international prices)

1. R&D (excluding AI)

Item Estimate Rationale
Swarm management software 1-2 billion Distributed system
VR simulators and training 0.5-1 billion
Scientific research 1-2 billion Materials science, thermodynamics
Total R&D 3-5 billion

2. AI Development for Robots

Why easier than Waymo/Tesla: - No pedestrians, cyclists, unpredictable people - No traffic rules, signs, intersections — only factory tasks - Fixed operations (mining, assembly) vs infinite driving scenarios - Can fail and retry — not real-time safety-critical

Characteristics: - Autonomy due to 5-20 min communication delay - Extreme conditions require reliable hardware

Item Calculation Estimate
Team (4,000 people × $150K × 6 years) ML engineers, roboticists, labelers 3.6 billion
Compute (GPU, data centers) $0.75 billion/year × 6 years 4.5 billion
Robot prototypes for training Thousands of units for data collection 2 billion
Data labeling Labeling millions of scenarios 1 billion
Total AI 10-12 billion

Risks: With scaling (more data, complex models) budget may grow to $15-20 billion.

3. Earth Test Site (China)

Why China: - Construction 2-3× cheaper than USA - Faster construction (Tesla Shanghai in 1 year) - Developed industrial base and logistics

Location: Gansu Province — simulating Mercury conditions.

Item Calculation Estimate
Construction 2 million m² × $580/m² 1.2 billion
Vacuum chambers 10 large chambers × $50M 0.5 billion
Thermal-vacuum chambers (-180°C…+430°C) 0.3 billion
Radiation simulators 0.2 billion
Factory prototypes 3-5 full cycle iterations 2 billion
Robot prototypes Thousands of units, dozens of versions 1.5 billion
Housing and infrastructure 10,000-15,000 people 0.5 billion
Personnel 10,000 × $30K × 4 years 1.2 billion
Logistics and materials 1 billion
Total Earth Test Site 8-10 billion

4. Lunar Test Site

Item Estimate Rationale
Delivery of 60 t to Moon 4-6 billion 10-20 launches
Test site equipment 3-5 billion Factory, robots, concentrators
2 years of operations 2-4 billion Management, iterations
Total Lunar Test Site 9-15 billion

5. Production: Electronics for Self-Replication

Principle: On Mercury, factories and robots are built from local materials (Al, Fe, Si). Only electronics (chips, boards, sensors) are imported from Earth.

Item Calculation Estimate
Bootstrap factories (E1 × 2) Full import 280 t 2.8 billion
Electronics for 1649 factories ~90 t × $3K/kg 0.27 billion
Electronics for 60K robots ~765 t × $3K/kg 2.3 billion
Electronics for 1000 MDs ~275 t × $5K/kg 1.4 billion
Chips for 1.1B mirrors ~1,055 t × $10K/kg 10.6 billion
Vitamins (4 years) ~509 t × $10K/kg 5.1 billion
Total production ~2,974 t electronics ~22.4 billion

Price justification (international sources 2026):

Category Price Rationale Sources
Robots, factories $3K/kg ARM + LiDAR 🇬🇧 Raspberry Pi, 🇨🇳 Livox Mid-360 ($2.3K/kg), 🇺🇸 Velodyne ($4.8K/kg)
Mass Drivers $5K/kg IGBT power electronics 🇩🇪 Infineon, Semikron, 🇰🇷 SemiHow
Mirrors $10K/kg Space-grade rad-hard 🇺🇸 NASA, 🇪🇺 ESA
Vitamins $10K/kg Iridium ($200K/kg) + rare earths Strategic Metals, 🇧🇪 Umicore

Robot references (international):

Country Manufacturer Price Source
🇺🇸 USA Boston Dynamics Spot $2.3K/kg IEEE Spectrum
🇨🇳 China SIASUN, ESTUN 20-35% cheaper than Western TAdviser
🇷🇺 Russia Promobot <$30K service robot Promobot
🇮🇳 India Tata/Mahindra market +8.8% CAGR IMARC

Frames, structures, housings — local production from regolith. Details: Delivery.

6. Rocket Program

Item Calculation Estimate
Production of ~100 rockets 100 × $50M 5 billion
Spaceport upgrades 2-3 billion
Infrastructure 1-2 billion
Total rockets 8-10 billion

7. Launches and Delivery

Item Calculation Estimate
~2,550 tons (total) 2,550 t × $2,500/kg 6.4 billion
Mission control 11 years of operations 3-5 billion
Total launches 9-11 billion

Includes: Moon (80 t) + factories (~370 t) + MD boards (~275 t) + mirrors (1,055 t) + vitamins (765 t). NaS capacitors — local production. See Import Summary.

8. Energy Reception System (Microwave + Rectenna)

Technology: LSP stations on Moon receive Swarm light → PV → 2.45 GHz microwaves → rectenna on Earth.

References (2026): - Caltech SSPP: first space-to-Earth power transmission demo (2023) - JAXA SPS: orbital solar power station program - ESA SOLARIS: space power transmission research - NTT + MHI: 152W transmitted over 1 km via microwaves (world record) - Rectenna efficiency: 85% (current technology)

8.1 Receiver Technology R&D

Item Calculation Estimate
High-power klystron development 90% efficiency, scaling 2-3 billion
Rectenna panel development 2.45 GHz antenna-rectifiers 1-2 billion
Tracking/pointing system Phased arrays, beam steering 1-2 billion
Earth prototyping Test site testing 0.5-1 billion
Space demonstration Pilot Moon→Earth transmission 0.5-1 billion
Total Receiver R&D 5-9 billion

8.2 Ground Infrastructure

Item Calculation Estimate
Rectenna panel production ~100 stations, GW scale 8-15 billion
Station construction Site preparation, foundations 10-20 billion
Land rights and approvals Lease/purchase, frequency regulation 2-5 billion
Environmental reviews EIA, compensations 1-3 billion
Control system Tracking, distribution 1-3 billion
Grid integration HVDC, operator coordination 3-6 billion
Energy storage (buffer) For night interruptions 3-6 billion
Total ground infrastructure 28-58 billion

8.3 Total Reception System

Subcategory Minimum Maximum
Receiver R&D 5 9
Ground infrastructure 28 58
TOTAL energy reception 33 67

9. Additional Items

Item Calculation Estimate
Communication system (DSN) Expansion of existing networks 2-4 billion
Propellant production 450,000 t LOX+methane 5-8 billion
Personnel training 50,000 people × $50K × 5 years 12-15 billion
International coordination Lawyers, diplomats 1-2 billion
Mission insurance 2-4 billion
Total additional 22-33 billion

10. LSP Stations on Moon

Instead of an orbital Hub at L1, LSP (Lunar Solar Power) stations on the Moon’s surface are used. Details: Energy Reception Hub.

Why LSP, not Hub?

Parameter Orbital Hub (L1) LSP (Moon)
Efficiency 10% 18%
Radiators 7,900 km² in space 0 (heat to ground)
Mass 15 million t in orbit ~178,000 t on surface
Resources Delivery from Moon In-situ

LSP Characteristics

Parameter Value
Stations 40 on lunar limbs
PV area per station ~160 km² (6,400 km² total)
Total mass ~178,000 t (per data.json HUB-001)
Materials Al 60%, Cu 28%, Si 9%, Fe 0.3% — local
Electronics from Earth ~200-300 t (2-3 launches)

LSP Cost

Item Calculation Estimate
Production (ISRU) Robots and equipment already counted in “Lunar Test Site” 0
Electronics from Earth ~250 t × $10,000/kg 2-3 billion
Management and logistics Construction coordination of 40 stations 2-4 billion
Total LSP 4-7 billion

Note: LSP stations are essentially free because production is from lunar materials by robots already counted in other budget items

Summary: Base Estimates

Category Minimum Maximum
1. R&D (excluding AI) 3 5
2. AI Development 10 12
3. Earth Test Site (China) 8 10
4. Lunar Test Site 9 15
5. Production (electronics) 25 32
6. Rocket program 4 6
7. Launches 6 8
8. Energy reception (rectenna) 25 49
9. LSP stations on Moon 4 7
10. Additional items 22 33
BASE TOTAL ~119 ~183

Structure by Location

Grouping by project phases and implementation locations.

Summary (Level 1)

# Location Min Max %
1 Earth (R&D, AI, test site) 34 44 24%
2 Moon (test site + LSP) 13 22 11%
3 Mercury (electronics, rockets, launches) 38 43 26%
4 Energy receiver (rectenna) 33 67 32%
5 Additional items (communications, propellant, training) 22 33 ~8%
TOTAL ~140 ~209 100%

Breakdown (Level 2)

1. Earth (~17%)

R&D, AI development, test site, personnel training.

Subcategory Calculation Amount
R&D (excluding AI) 3-5 billion
— Swarm management software Distributed system 1-2 billion
— VR simulators Operator training 0.5-1 billion
— Scientific research Materials science, thermodynamics 1-2 billion
AI Development 10-12 billion
— AI team 4,000 people × $150K × 6 years 3.6 billion
— Compute $0.75 billion/year × 6 years 4.5 billion
— Robot prototypes Thousands of units for training 2 billion
— Data labeling Millions of scenarios 1 billion
Earth Test Site (China) 8-10 billion
— Construction 2 million m² × $580/m² 1.2 billion
— Test equipment Chambers, simulators 1 billion
— Factory prototypes 3-5 iterations 2 billion
— Robot prototypes Dozens of versions 1.5 billion
— Test site personnel 10,000 × $30K × 4 years 1.2 billion
— Housing and logistics 1.5 billion
Personnel Training 13-17 billion
— Staff training 50,000 people × $50K × 5 years 12-15 billion
— International coordination Lawyers, diplomats 1-2 billion
Total Earth 34-44 billion

2. Moon (~5%)

Technology validation and LSP station construction.

Subcategory Calculation Amount
Test Site 9-15 billion
— Delivery of 60 t to Moon 10-20 launches 4-6 billion
— Test site equipment Factory, robots, concentrators 3-5 billion
— 2 years of operations Management, iterations 2-4 billion
LSP Stations 4-7 billion
— Electronics from Earth ~250 t × $10,000/kg 2-3 billion
— Management and logistics 40 stations 2-4 billion
Total Moon 13-22 billion

3. Mercury (~30%)

Electronics for self-replication, rocket program, launches.

Self-replication principle: Factories and robots are built on Mercury from local materials. Only electronics (chips, boards, sensors) are imported from Earth — ~2,550 t over 11 years.

Subcategory Calculation Amount
Electronics (import) ~21 billion
— Bootstrap factories (E1 × 2) Full import 280 t 2.8 billion
— Electronics for 1649 factories ~90 t × $3K/kg 0.27 billion
— Electronics for 60K robots ~765 t × $3K/kg 2.3 billion
— Electronics for 1000 MDs ~275 t × $5K/kg 1.4 billion
— Chips for 1.1B mirrors ~1,055 t × $10K/kg 10.6 billion
— Vitamins (4 years) ~509 t × $10K/kg 5.1 billion
Rocket Program 4-6 billion
— Production of ~50-100 rockets ~75 × $50M 3.75 billion
— Spaceport upgrades 2 sites 1-2 billion
Launches and Delivery 6-8 billion
— ~2,550 t to Moon + Mercury 2,550 t × $2,500/kg 6.4 billion
— Mission control 4 years of operations 0-2 billion
Total Mercury ~38-43 billion

Support (propellant, communications, insurance) — in “Additional Items” section.

4. Energy Receiver (~32%)

Ground infrastructure for energy reception.

Subcategory Calculation Amount
Receiver R&D 5-9 billion
— Klystron development 90% efficiency 2-3 billion
— Rectenna development 2.45 GHz antenna-rectifiers 1-2 billion
— Pointing system Phased arrays 1-2 billion
— Prototypes and demos 1-2 billion
Ground Infrastructure (rectenna) 28-58 billion
— Rectenna (~100 stations × 100 km²) 10,000 km² × $0.8-1.5M/km² 8-15 billion
— Station construction Site preparation, foundations 10-20 billion
— Land rights and approvals Lease/purchase, frequency regulation 2-5 billion
— Environmental reviews EIA, compensations 1-3 billion
— Energy storage Buffers 3-6 billion
— Grid integration HVDC and coordination 3-6 billion
— Control system Tracking, distribution 1-3 billion
Total Receiver 33-67 billion

Analysis: Which Items Are Underestimated?

Comparison with real projects shows some estimates may be optimistic.

1. AI Development (was: $15-20 billion → revised: $25-40 billion)

References:

Company Spent Result Source
OpenAI $20+ billion GPT-4, no AGI DCD
Waymo $11 billion over 15 years No full autonomy VentureBeat
Tesla FSD $15 billion over 8 years Level 2 autopilot CleanTechnica
DeepMind $50+ billion Research Wikipedia

Problem: Helios requires fully autonomous robots in conditions where Waymo/Tesla would fail: - Vacuum, radiation, +430°C/-180°C - 5-20 minute communication delay (no GPS, no real-time) - Unpredictable regolith and conditions

Adjustment for baseline scenario: $18 → $30 billion (+$12 billion)

2. Rocket Program (was: $8-10 billion → revised: $10-14 billion)

References:

Program Cost Result
SLS $35 billion 1 rocket (expendable)
Starship $5-10 billion Development (no serial production)
Long March 9 (planned) $10-15 billion Chinese super-heavy

Thanks to 99.998% localization: Only ~100 rockets needed (not ~54,000). $50M/rocket × 100 = $5 billion — realistic.

If a new rocket is needed (Russia+China): +$5-8 billion for contribution to joint launcher development.

Adjustment for baseline scenario: $55 → $75 billion (+$20 billion)

3. Factory Production (was: $46-49 billion → revised: $60-80 billion)

References:

Factory Cost
Tesla Gigafactory Shanghai ~$2 billion
Tesla Gigafactory Nevada ~$5 billion
Semiconductor fab (Taiwan) $12-20 billion

Counterargument: Helios factories are simple metallurgy, not semiconductors. At 1650-unit series, price drops.

However: These are autonomous space factories on Mercury. $30M/unit at 1650 series — optimistic.

Adjustment: $30M → $50M/factory = $82.5 billion (×1.7)

4. Launches ($55 billion = $2,500/kg) ✓

This estimate is realistic assuming operational 100+ t heavy rockets: - SpaceX Falcon 9: $2,700-5,000/kg (today) - Heavy rockets (target): $1,000-1,500/kg - Long March: $3,000/kg

With developed infrastructure and launcher mix — $2,500/kg is achievable.

Adjustment Summary

Item Was (baseline) Became Δ
AI Development 18 30 +12
Rocket program 9 12 +3
Production 68 100 +32
Total adjustment +47

This increases the baseline scenario from ~$227 to ~$274 billion, and with ×1.5 reserve — to ~$430 billion.

Three Scenarios (revised)

Optimistic Scenario (~$200 billion)

Conditions: - Heavy rockets operational by 2028, price $1,000/kg - China and Russia take on 50% of production - AI reaches required level in 4-5 years (not 6-7) - Series effect reduces prices by 30% - Self-replication works — only electronics imported

Category Estimate ($ billion)
R&D (excluding AI) 3
AI Development 8
Earth Test Site 6
Lunar Test Site 8
Electronics (self-replication) 25
Rockets 3
Launches 5
Energy reception (rectenna) 25
LSP stations on Moon 4
Additional 18
Reserve (+20%) 21
TOTAL ~126 billion
With ×1.6 margin ~202 billion

Baseline Scenario (~$355 billion, or ~$445 billion with inflation)

Conditions: - Launcher mix (Starship + Long March + Soyuz) - International cooperation works - AI requires 6-7 years of development - Moderate delays and overruns - Self-replication works — only electronics imported

Category Estimate ($ billion)
R&D (excluding AI) 4
AI Development 11
Earth Test Site (China) 9
Lunar Test Site 12
Electronics (self-replication) 30
Rockets 5
Launches 8
Energy reception (rectenna) 40
LSP stations on Moon 5
Additional 28
Reserve (+30%) 53
TOTAL (USD 2026) ~224 billion
With ×1.6 margin ~$358 billion
With +25% inflation ~$448 billion

Conservative Scenario (~$490 billion, or ~$610 billion with inflation)

Conditions: - Heavy rockets delayed until 2032 - AI requires 8-10 years and additional iterations - Political complications increase costs - Technical problems require rework - Need to develop own super-heavy launcher

Category Estimate ($ billion)
R&D (excluding AI) 6
AI Development 20
Earth Test Site 14
Lunar Test Site 20
Electronics (self-replication) 40
Rockets 8
Launches 12
Energy reception (rectenna) 58
LSP stations on Moon 7
Additional 40
Reserve (+50%) 128
TOTAL (USD 2026) ~380 billion
With ×1.3 margin ~$494 billion
With +25% inflation ~$617 billion

Why Not a Trillion?

American projects are inflated due to:

  1. Cost-plus contracts — contractor benefits from spending more
  2. Bureaucracy — approvals, reviews, reports
  3. Contractor monopoly — Boeing, Lockheed without competition
  4. Legal costs — lawsuits, patents, insurance

Examples of budget inflation:

Project Initial budget Final Growth
James Webb $1 billion $10 billion ×10
SLS $9 billion $23+ billion ×2.5
F-35 $233 billion $400+ billion ×1.7

Why Helios is different:

  1. International competition — China, Russia, India are cheaper
  2. Mass production — 1650 factories, 60,000 robots, economies of scale
  3. Fixed-price contracts — not cost-plus
  4. Common goal — fewer political games

Comparison with Real Projects

All amounts converted to USD 2026 (with inflation adjustment for historical projects).

Project Nominal USD 2026 Duration Result Source
Manhattan $1.9 billion ~$30 billion 3 years Nuclear weapons Brookings
Apollo $25-28 billion ~$280 billion 11 years 12 people on Moon Planetary Society
GPS (initial) — ~$12 billion ongoing 31 satellites Time
James Webb $10 billion ~$10 billion 24 years 1 telescope Planetary Society
SLS (development) $29-35 billion ~$35 billion 13+ years 1 rocket Planetary Society
ISS $100-150 billion ~$150-250 billion 26 years 1 station Wikipedia
Tiangong ~$8 billion ~$8 billion ~10 years 1 station Estimate
Artemis $93 billion ~$93 billion (by 2025) 14+ years 3 flights Space.com
Three Gorges ¥200 billion ~$28-37 billion 17 years 22.5 GW Wikipedia
Starship $5-10 billion ~$5-10 billion 10+ years Rocket SpaceNews
Tesla Gigafactories $15+ billion ~$15+ billion ~10 years 6 factories Wikipedia
Helios (baseline) — ~$355-445 billion 10-20 years 10,000 TW Our estimate

Conclusion: Helios costs approximately 1.5× Apollo but delivers 4000 times more energy than current planetary consumption. In terms of cost/benefit ratio — the most efficient project in history.

Year-by-Year Distribution (baseline scenario)

Year Phase Main expenses Amount (billion)
1-2 R&D, AI start AI team, design 35-45
3-4 Earth Test Site, AI Construction in China, model training 65-80
5-6 Lunar Test Site, production Lunar missions, mass production 85-100
7-10 Mercury: deployment First expeditions, Mass Drivers 80-100
11-15 Mercury: scaling Vitamins, energy reception 65-80
Total (USD 2026) 330-405
With inflation ~410-510

Known Limitations

Attention: This section describes problems requiring solutions.

High-Priority Issues

1. Testing and Debugging Underestimated

Stage Current estimate Realistic estimate Rationale
Factory prototypes 3–5 iterations 5–10 iterations 3-5 ground (Gansu) + 2-5 lunar
Robot prototypes $2 billion $5–10 billion Hundreds of versions, thousands of units
Lunar Test Site 2 years 1.5–2.5 years 3-5 iterations × 3-5 months + deployment

Reason: The ground test site (Gansu) allows completing major iterations before going to the Moon (TRL 4→6). The Moon validates in the real environment (TRL 6→7-8). Lunar iteration cycle with dedicated launches: ~2-4 months (3-day flight; main time is analysis and fix manufacturing).

Detailed methodology: Risks → Testing Estimation Methodology

Require Research

2. Mission Insurance

  • $2–4 billion — optimistic for 10,000 tons of cargo
  • At real risk rates, may be 5–10% of cargo value
  • Realistic: $5–10 billion for insurance

3. Political Risks

  • Sanctions between countries → cooperation breakdown
  • Government changes → priority shifts
  • Competition instead of cooperation → cost duplication

Cannot be estimated, but potentially +50–100% to budget.

See Also