Actionable insights — September 2026 Portfolio Update

21:01 1. Give every holding a five-year expected return and a hurdle

1. Estimate compounded EPS growth for five years (earnings, not revenue, if heavy investment will be amortized).
2. Pick a defensible exit P/E for the business quality.
3. Solve the implied annual return from today's price; also solve the price that would deliver your hurdle (here 15%).
4. Aim for "accurate, not conservative or aggressive" inputs; treat the output as directional.

• Multiples assumed above the business's history without a reason; inputs drifting toward whatever justifies holding.

1:11:24 2. Steer new money by expected return; don't churn the book

1. Rank all holdings by expected return.
2. Direct fresh cash to the top names; let below-hurdle names become holds.
3. Trim only when a great company's price far exceeds your fair value, and keep a foot in the door.
4. Respect friction, taxes, concentration and risk-factor overlap before any sale.

• A holding's return falling below ~10% even for top quality; one name growing into an outsized weight.

1:12:14 3. Run a clean-slate allocation to test your conviction

1. Pretend you own nothing and have a fixed sum.
2. Weight toward the highest expected returns but diversify across the next tier rather than going all-in.
3. Compute the weighted expected return and compare it with your actual portfolio's mix.

• Big gaps between the clean-slate weights and actual weights — candidates for where new money goes next.

8:37 4. Map each holding onto the phases of a technology cycle

1. Define the phases: build (suppliers), monetization (buyers turn it into revenue), distribution, commoditization, durable ownership of the customer.
2. Label each holding by the phase in which it earns the most.
3. Be early for the later phases (markets price them ahead) and be most careful holding build-phase suppliers into commoditization.

• Supplier pricing power fading as capacity catches up; hyperscaler AI revenue showing up in reported segments (e.g. Azure now disclosed).

56:51 5. Compare competitors' strategies for the same threat — aggregate vs build

1. For a disruptive technology, list what each incumbent is doing: partnering/aggregating or building in-house.
2. Weigh the aggregator's capital-light reach against the builder's control and cost.
3. Look for real deployments ("wheels on the ground"), not partnership announcements.

• Deployed vehicle/drone counts and share of orders automated, per quarter.

2:53 6. Measure yourself on time-weighted returns and multi-year win rates

1. Report all-in gains including every sale at a loss and dividends, not just current holdings.
2. Use time-weighted return so deposit timing doesn't flatter or punish results.
3. Judge underperformance years against a realistic win rate; check whether holdings' fundamentals (revenue growth) are still compounding.

• Fundamentals, not just price, lagging for a sustained period.