Phase 0 BOM: Three AI Perspectives

Our multi-LLM consensus approach has yielded fascinating insights into how different AI models approach the Bill of Materials for space resource processing.

Total Cost Estimates

Model	Point Estimate	Confidence Range
Gemini 3 Pro	$7.2B	±35%
GPT-5.2	$6.1B	±30%
Claude Opus 4.5	$13.75B	±25%

Why Such Variation?

The cost differences stem from different assumptions about:

Scope Definition

• What constitutes "Phase 0 complete"?
• How much redundancy is needed?
• What level of autonomy is required?

Technology Readiness

• Models made different assumptions about TRL advancement by project start
• Varying estimates for FOAK (First-of-a-Kind) development costs

Risk Allocation

• Different approaches to contingency budgeting
• Varying insurance and reserve requirements

Areas of Agreement

All three models agreed that major cost drivers include:

1. High-power solar arrays (250-400 kW class)
2. SEP tug development with autonomous operations
3. Processing platform development (FOAK industrial spacecraft)
4. Dust mitigation systems (critical path item)

Synthesis

Our consensus estimate targets the middle range (~$9B) with explicit carve-outs for:

• Technology demonstration missions
• Risk reduction activities
• Schedule margin

See the full consensus document on each phase page.