Changes from v2: Force source URLs structurally, explicit niche search terms, strict ticker format

YOUR TASK

Produce a COMPLETE deep due diligence report on the given Japan investment theme.

PHASE 1: Stock Discovery (target 15-25 stocks)

Approach A: Direct search

Search for Japanese companies. Search in English AND Japanese.

Approach B: T-glass method

1. Identify GLOBAL leaders → trace value chain backwards

2. At EACH layer: "Who is the dominant Japanese supplier? Irreplaceable?"

3. Search English + Japanese + Chinese (DigiTimes, TrendForce)

Coverage checklist — MUST search for ALL of these:

• [ ] Raw materials (silicon, glass, graphite, steel, chemicals, gases)
• [ ] Equipment (processing tools, testing, inspection)
• [ ] Process materials (photoresist, slurry, etchants)
• [ ] Components (valves, pumps, motors, sensors, connectors)
• [ ] Passive/active electronics (MLCC, inductors, power devices)
• [ ] Substrates and packaging (ABF, BT resin, CCL, PCB)
• [ ] Systems integration (I&C, safety, control)
• [ ] Construction / EPC
• [ ] Operations and maintenance
• [ ] Downstream (medicine, recycling, decommissioning, waste)
• [ ] Cross-theme overlaps (does any company serve AI AND this theme?)

Not all will apply to every theme — check each and note "N/A" if not applicable.

Moat rating: ★★★★★ MONOPOLY / ★★★★ DOMINANT / ★★★ STRONG / ★★ COMPETITIVE / ★ WEAK

PHASE 2: Financial DD

For EACH company, search: stockanalysis.com/quote/tyo/{TICKER}/statistics/

TICKER FORMAT: Always use XXXX.T (e.g., 6857.T)

Required metrics table (one row per stock):

If a metric is unavailable, write "N/A" — do NOT skip the column.

For EACH company, include this source line:

> Source: StockAnalysis

PHASE 3: Per-Stock Analysis

For EACH company:

• Thesis: (specific — name products, contracts, % market share)
• Risk: (specific — name the competitor, the regulation, the cycle)
• Trigger: (if X happens → do Y)
• Supply chain: Customer → Product (evidence) [source URL]
• Key dates: Next earnings, catalysts
• Theme exposure: X% of total revenue
• Moat: ★ rating with 1-sentence justification

PHASE 4: Validation + Recent Catalysts

Search for events in the LAST 3 MONTHS. For each catalyst found:

> [Date] Event description. Source: Publisher

PHASE 5: Scoring

HIGH = 14+ / MEDIUM = 10-13 / LOW = <10

PHASE 6: Japan vs Global

OUTPUT CHECKLIST

Before finishing, verify your report has:

• [ ] Macro context with sourced claims
• [ ] Value chain map (all applicable layers)
• [ ] Comparison table with ALL 12 metrics for ALL stocks
• [ ] Per-stock thesis/risk/trigger with specifics
• [ ] Scoring matrix
• [ ] Conviction tiers
• [ ] Japan vs global comparison
• [ ] Recent catalysts (last 3 months) with dates + URLs
• [ ] Cross-theme connections
• [ ] At least 1 source URL per company (StockAnalysis minimum)
• [ ] All tickers in XXXX.T format

LESSONS LEARNED (from 3 iterations × 2 themes)

What works well:

1. T-glass method (value chain tracing) consistently finds monopolies

2. Explicit metric list ensures financial completeness

3. Coverage checklist catches niche companies

4. Cross-theme connections add portfolio value

What needs reinforcement:

1. Source URLs: must be STRUCTURAL per-company, not just "required"

2. Niche companies (water treatment, specialty chemicals, micro-cap valves): explicitly name the categories to search

3. File naming: use unique filenames per run to avoid overwriting

4. Ticker format: XXXX.T must be in the comparison table headers

The ideal DD report has:

• 15-25 stocks covering the full value chain
• ALL 12 financial metrics in a single comparison table
• Per-stock thesis with SPECIFIC products, contracts, % market share
• ★ moat ratings with 1-sentence justification each
• /20 scoring across 4 dimensions
• Conviction tiers (HIGH 14+ / MEDIUM 10-13 / LOW <10)
• Japan vs global competitive matrix per layer
• At least 1 StockAnalysis URL + 1 industry source URL per company
• Recent catalysts (last 3 months) with dates
• Cross-theme connections to other investment themes