—
Top Site
—
Best 15-Yr TCO
—
🇺🇸 US Rank

Global Economic Dials

⚡ Non-Conventional Solutions
Model Note

Baseline physics from T. Lambert's foundry economics: equipment = 70–80% of build, materials ≈ 40–45% of run cost, electricity ≈ 30% (US) vs 5% (low-cost grids), labor ≈ 25% (US) vs 10%, waste cheapest. 15-year TCO horizon. Dials span 0–200%+, the full range the optimizer may need.

Target Site Selection — Global Hex Cartogram

HOSTILE
OPTIMAL ◈ dashed = USA  •  click a tile to inspect & target
Selected Site — Deep Scan

Site Ranking ● LIVE

🎯 Auto-Optimizer

Target: United States. The optimizer searches every global dial for the smallest change from your current settings that makes the target #1 — then explains, in plain language, what would have to happen in the world.

0 evalsidle
// solver log — binary-search traces appear here
📖 Dial Glossary (plain English)
  • Fab construction cost — the all-in price to build one leading-edge fab. Terry's benchmark: ~$50B.
  • Equipment share — how much of the build is lithography & fab machinery vs. buildings and hookups. Terry: 70–80%. Higher share = more exposure to the ASML chokepoint.
  • Labor cost pressure — a worldwide multiplier on skilled wages. 100% = today; 50% = wages halve everywhere; 200% = they double.
  • Materials cost pressure — same idea, for wafers, photoresist, wet chemicals and specialty gases.
  • Energy price shock — a worldwide multiplier on electricity prices. Below 100% = cheap-power glut; above = power crunch.
  • Discount rate — how heavily investors punish money tied up during the 3–5 year wait for first silicon. High = expensive capital.
  • Subsidy — the share of the build cost the government pays. 40% ≈ a maximum-effort CHIPS-style package.

🧮 Change Tabulator — What Would It Take For Every Country?

Session buffer: 0 / 20 country solutions
Tip: solving countries one-by-one with the Auto-Optimizer pre-fills this table — each buffered solution is subtracted from the big compute. Buffer persists for this session.
Idle. Estimated full compute: ~7,300 evaluations.

Field Notes — Source Material & Model Assumptions

The Capital Problem

"About $50 billion per foundry… 3 to 5 years before first silicon… return on investment tends to be between 10 to 15 years."
Modeled as the construction dial plus a discounted 15-year TCO. The discount rate captures the pain of money not making money.

Run-Cost Anatomy

"70–80% of build cost is photolithography & fabrication machinery… 40–45% of run cost is materials… 30% electricity in the US, 5% elsewhere… labor 25% US, 10% elsewhere. Waste disposal is cheapest — re-processing is common."
These ratios are the engine; each country re-prices each component locally, and each optimizer explanation is written in terms of them.

The Materials Chokepoint

"Historically, about 90% of this quartz has been mined in North Carolina… deposits elsewhere save a little on transportation."
Captured in each country's materials index.

The Lithography Chokepoint

"SVG was acquired by the Dutch company ASML in 2001, effectively marking the end of domestic U.S. production for high-end optical lithography."
Each country has a litho-access score; the equipment share dial controls how much that score matters.

The SMR Gambit

"If you could get cheap electric — say if you had a couple of SMRs running the place — you could probably knock 25% out of the run cost."
Toggle SMR Power Park — or let the optimizer show you that a global energy price shock in reverse does the same job for high-power-cost nations.

How To Read The Optimizer

Score = 55% cost + 45% feasibility. The optimizer can only move world price dials — it cannot fix a weak ecosystem, thin talent pool, or blocked lithography. When a country can't reach #1 at any dial setting, the verdict tells you the gap is structural, and which national levers would help instead.