Skip to content

NVDA Real Walkthrough (From Thesis to Transcript to KPI Verification)

Follow a stock all the way through — using NVIDIA as an example (because we have 8 quarters of transcripts + 7 thesis versions + actual verifier data).

Step 1: What's NVDA's Current State

Ticker: NVDA (NVIDIA Corp), NASDAQ
Market Cap: ~$3.3T (formerly world #1)
Business: AI GPU + Data Center Networking + Autonomous Driving Platform
Latest Thesis Version: v7 (plus v91 v8, including CRWV circular financing red_flag)
Latest Financials: Q4 FY26 (reported Feb 2026), revenue $39.3B, gross margin 75%, EPS $0.89

Next catalyst: Q1 FY27 Earnings 2026-05-20 (tomorrow) — this is the perfect time for a real walkthrough.

Step 2: Read the Q4 FY26 Transcript (Last Earnings)

Our corpus has the Motley Fool NVDA Q4 2026 transcript.

Beginner's Reading Method (1 hr):

  1. Skip the opening: The CEO always says "we had a record quarter" — no information value. Go straight to the finance section of prepared remarks.
  2. Look at segment breakdown + YoY %:
  3. Data Center: $35.6B (+93% YoY) ← This is the main driver
  4. Gaming: $2.5B (+13%)
  5. Auto: $0.6B (+103%)
  6. Management guidance for next quarter:
  7. Q1 FY27 revenue: ~$43B (consensus $44B)
  8. Gross margin: 73.5-74% (down from 75%, due to Blackwell ramp costs)
  9. Read Q&A (this is the highest-density information section):
  10. Analysts' real questions reveal what they care about
  11. Management's unscripted answers reveal the truth (vs. prepared remarks marketing spin)

Key Q&A Questions Example (NVDA Q4 FY26 Real)

Q (Morgan Stanley): "Vera Rubin shipment timing?" A (Jensen): "Production samples shipping H1 calendar 2026, full ramp H2" → Signal: Rubin is 1-2 quarters earlier than expected, bull for FY27 revenue.

Q (Bernstein): "How are you managing Samsung HBM3e qualification?" A: "We work with all three [Hynix/Micron/Samsung], qualifying Samsung is an ongoing process" → Signal: Samsung not qualifying is bear (but NVDA has backup, not fatal).

Step 3: Fill Transcript Data into Thesis Framework

Refer to Thesis 4-Dimension Framework:

Field NVDA Current
view bull
confidence medium
core_thesis "NVDA benefits from AI infrastructure capex; Q1 FY27 earnings will validate demand strength; watch customer quality risk (CRWV)"
support 1 "2026 hyperscaler capex $725B+, NVDA main beneficiary"
support 2 "CRWV $99.4B backlog (NVDA holds $36.6B)"
support 3 "327 institutions' 13F still have NVDA as AI core holding"
red_flag 1 "Samsung strike → HBM supply disruption" (trigger: strike >7 days)
red_flag 2 "CRWV-NVDA circular financing concerns" (trigger: NVDA exposure to CRWV expands)
catalyst "2026-05-20 Q1 FY27 earnings"
price_outlook base $215-235 / bull $245-260 / bear $190-205

Step 4: Derive Verifiable KPI Predictions from Thesis

Our system forces the thesis to output ≥3 KPI predictions (Layer 5 teaches how to write). NVDA v91 real predictions:

predictions:
  - kpi_name: revenue_quarterly
    expected_value: 43
    expected_range: [42.5, 44]
    unit: USD billion
    reporting_quarter: "Q1 FY27"
    verify_at: 2026-05-20
  - kpi_name: data_center_segment_revenue
    expected_value: 36
    expected_range: [34, 38]
    unit: USD billion
    reporting_quarter: "Q1 FY27"
    verify_at: 2026-05-20
  - kpi_name: gross_margin_non_gaap
    expected_value: 74.0
    expected_range: [73.5, 75]
    unit: percent
    reporting_quarter: "Q1 FY27"
    verify_at: 2026-05-20

Tomorrow (5/20) after earnings, the KPI verifier will automatically run via cron, compare predictions vs. actuals, and tag each prediction as confirmed / partial / wrong.

Step 5: Multi-Perspective Sanity Check (Layer 3)

Don't look at NVDA fundamentals in isolation — look at it simultaneously:

Multi-PM Perspective: - Value PM (Buffett-style): "NVDA P/E 30x isn't cheap, but ROE 100%+ is historically unprecedented; hold for now" - Growth PM (Druckenmiller-style): "FY27 guide $200B+, growth still 50%+, add position" - Macro PM (Soros-style): "AI capex is the macro theme, but Fed hawkishness + 10Y 4.5% compresses valuations; hold, don't add"

Anti-Thesis: - If you were an NVDA short, how would you argue? - "Hyperscaler capex is over-investment; $725B is a bubble peak; ROI disappoints" — this is a real angle written by independent analysts at SemiAnalysis - Your thesis must have an explicit reply: "We think it's not a bubble because X / Y / Z"

Macro Overlay: - Current macro regime = mixed, near-term neutral — neither risk_off nor risk_on - Impact on NVDA: rising rates (compresses valuations) + improving liquidity (supports) → net neutral, no change to NVDA bull thesis

Technical Overlay: - NVDA current trend=uptrend, momentum=neutral, ADX 30.9 (strong trend) - Entry zone: $210-218 (20d SMA 211.3 + Bollinger middle) - Exit/stop: $236-240 resistance / stop loss below $200 - → "Fundamental bull + Technical bull but momentum neutral = don't chase highs, wait for pullback to 20d MA to add"

Step 6: How to Update After 5/20 Earnings

Earnings Night (After-Hours):

  1. Read management commentary + Q&A (Motley Fool usually publishes transcript 1-2 days later)
  2. Compare against our KPI predictions:
  3. Actual revenue vs. our $43B median prediction
  4. Actual Data Center segment vs. $36B
  5. Actual gross margin vs. 74%
  6. Look at next quarter guidance:
  7. Q2 FY27 guide >$50B? = bull confirmation
  8. <$45B? = warning

7-14 Days After Earnings:

  1. 13F reports disclose (Q1 holdings around 5/15) — verify against our 13F support
  2. Sell-side analysts raise/lower target prices — watch for differentiation
  3. Our thesis decision: maintain v8 / upgrade to v9

How to Move to the Next Stock

Completing NVDA = you've mastered a full cycle: thesis → transcript → KPI predictions → multi-view sanity → post-earnings verification → reflection

Pick your most familiar ticker (e.g., TSM / AMD / GOOGL) and repeat:

  • Find its Motley Fool transcript
  • Write a 5-dimension thesis
  • 3-5 KPI predictions
  • Wait for the next earnings to verify

After 3-5 stocks, you'll be more systematic than 90% of retail investors.

Next Layer → Layer 3 · Multi-Perspective