Investing Notes.md

Personal investing notes — principles, company research, and the reasoning behind each decision, in one source file.

• Don't buy a business you can't explain. If you can't answer "how does this company make money?" in one sentence, pass.
• Look for the moat (why competitors can't copy this) first. Without a moat, even great results don't last.
• Set a valuation ceiling in advance. Even a great company bought too expensive returns little.
• Max single-position weight = 20% of the portfolio. Cap concentration regardless of conviction.
• Write the sell conditions when you buy. Deciding "why I'll sell" in advance prevents emotional trading.

• Core: data-center AI accelerators. The real moat is the CUDA software ecosystem — switching costs beat raw hardware.
• Bull case: AI training/inference demand is a structural shift in computing, not a passing boom. Backed by hyperscaler capex.
• Risks: (1) high valuation — small misses hit hard. (2) customer concentration — a few large clouds are a big share of revenue. (3) in-house ASIC competition. (4) cyclicality — inventory/macro sensitive.
• Monitor: data-center revenue growth, cloud capex guidance, inventory levels, competitor accelerator shipments.
• 2025-02 decision: added to position. Assumption = data-center capex lasts 2-3 years.

• Core: global beverage brand + distribution. Moat = brand + bottler network + shelf dominance.
• Bull case: 60+ years of dividend increases (Dividend King). Defensive, predictable cash flow. Dampens portfolio volatility.
• Risks: (1) low growth — single-digit revenue growth. (2) sugar/health headwinds. (3) FX losses when the dollar is strong. (4) dividend appeal falls relatively when rates rise.
• Monitor: organic revenue growth, payout ratio, FX impact, share of still/health categories.
• 2025-06 decision: new position. Assumption = dividend-growth record holds + need a defensive anchor.

• 2025-02 added to NVIDIA.
• 2025-06 opened Coca-Cola.
• 2025-09 passed on chasing high-valuation growth name A — valuation exceeded the checklist ceiling. "Good company != good price."