The course's example: Nvidia, 20 Nov 2026, from Modules 9 and 10 as they stand. An example of form, not a recommendation.
Steps 1–5, in short. Price $145.98 (19 Nov 2026 close), market cap $3,531B, EV $3,510B; down 34% from $221.97 on 21 Oct, and the filings give no cause, so neither does the memo. Peer multiples applied to Nvidia's own figures: about $79 to $990 a share, about $400 on Broadcom's EV/operating income (Module 9 lesson 06). DCF (Module 10 lesson 08): Story A $267.10, base $157.65, Story B $83.45; weighted 25/50/25, $166.46. Reverse DCF (Module 10 lesson 09): at a 60% year-5 margin and 10%, the price needs 8.0% growth a year, year-5 revenue about $532B. Unknowns go in as questions, with no invented probabilities.
Step 6. The memo (Module 17 lesson 09's template).
MEMO — Nvidia (NVDA), 20 Nov 2026, price $145.98 (19 Nov 2026 close). Sources: FY2026 10-K; 10-Q to 26 Jul 2026; 8-Ks of 17 Aug–17 Nov 2026; Module 9 lesson 06; Module 10 lessons 07–09.
| Heading | Content |
|---|---|
| DECISION | Research more. Five questions from the filings could change every number below. |
| THESIS | Data-centre sales keep growing because cloud companies keep buying, and CUDA keeps them from switching (16-07). |
| FACTS | Twelve months to 25 Oct 2026: revenue $361.8B, operating margin 66.1%. Q3 revenue $115.8B (+103%). Two customers 36% of FY2026 revenue. |
| VALUATION | P/E 15.7. Priced like a peer: $79–$990 a share, about $400 on Broadcom's EV/operating income; relative values only. DCF $83.45–$267.10, weighted $166.46 on my weights. Run on the S&P 500, the template sits about 15% below it, so my opinion: the base case, 8% above the price, is near it. |
| WHAT THE PRICE ASSUMES | In my model, 8.0% growth a year for five years, margin slipping to 60%: less than the company's own guidance ($130.0B ± 3% for the quarter to January; four such quarters ≈ $520B) and the industry's expected 11.71%. It needs the boom not to reverse: Story B gives $83.45. |
| VARIANT VIEW | None yet. Differing would need evidence on whether the boom lasts, and I have none. |
| CATALYSTS | The 10-Q (due by 4 Dec 2026). Q4 results. A permanent CEO (an 8-K). Hugging Face closing (H1 2027). |
| RISKS | (1) Customer concentration; customers' own chips. (2) Commitments if demand turns: the guarantee capped at $105B; $279B of supply commitments. (3) A leadership change with no stated reason. |
| WHAT WOULD MAKE ME WRONG | Data Center revenue falls year-on-year two quarters running, or a top customer says it's moving to its own chips (sell card row 1). |
| SIZE | None: not acting. |
| OPEN QUESTIONS | What could (1) the guarantee cost if OpenAI can't pay, and (2) Hugging Face add, if approved? (3) What does the unfiled 10-Q show? (4) Who leads the company? (5) How concentrated are customers now? |
| REVIEW DATE | Mon 7 Dec 2026, after the 10-Q. Then act only if I can write a variant view with a number that survives the base-rate check; do nothing if I still can't. |
Step 7. Size, only if acting: risk limit ÷ distance to your exit or review point (Module 16 lesson 03), under your own caps; then the five-row sell card (lesson 07).
Step 8. Journal (Module 17 lesson 04), three of its fields: What I expect, by when: Q4 revenue inside $126.1B–$133.9B, on results day, 70% sure. What would make me wrong: revenue outside it. How I feel right now: unsettled by the fall; tempted to act. Pre-mortem: a year on, it went badly because the research never ended; so on 7 Dec, choose.
Keeping it updated: a worked update
That memo is itself an update. Three of its lines as written on 27 Aug 2026, after the quarter to 26 Jul (price $209.66, 26 Aug close):
FACTS Q2 revenue $96.2B (+106%); Data Center $89.0B (+117%)
CATALYSTS Q3 results in Nov, against guidance of
$108.0B ± 2% ($105.84B–$110.16B)
REVIEW DATE Q3 results day
The dated update, 20 Nov 2026:
| Update | |
|---|---|
| What changed | 17 Nov: Q3 revenue $115.8B, above the range, 7.3% over its midpoint; Data Center $107.6B (+110%); Q4 guidance $130.0B ± 3%. 8-K of 12 Nov: on 11 Nov the board accepted Huang's resignation as President, CEO and director; CFO Colette Kress is interim CEO; no reason given. |
| Lines it touches | FACTS; VALUATION and WHAT THE PRICE ASSUMES, re-run at the new price; CATALYSTS (Q4 guidance, a permanent CEO); RISKS and OPEN QUESTIONS (who leads). |
| Does the decision move? | No. Data Center grew year-on-year, so WHAT WOULD MAKE ME WRONG isn't met. The price fall is a prompt to re-read, not a reason (16-07). Still research more. |
| Next review | 7 Dec 2026, after the 10-Q. |
The habit: date, what changed, which lines, whether the decision moves, next date. Ten minutes per filing.