NVIDIA moat duration and fair value under margin compression scenarios
These are the assumptions the argument rests on, each one argued rather than asserted: what it is, where our number came from, how sure we are, and what it is worth to disagree with us. Move them. If the conclusion survives only at our settings, that is worth knowing before you accept it.
The model's central output is an implied share price: the fair value of NVIDIA equity under the chosen scenario, expressed in dollars per share. It answers the question, 'What should this stock trade at if our assumptions about earnings growth and market sentiment are correct?' The house number is the Base scenario result, currently $245–$278/share, which sits +9% to +24% above the $224 market price as of 7 August 2026.
Base $261 · Thesis $322 · Antithesis $178 · Panel, pre-verification (their house case) $322 — the house number is the price of the Base case being right.
The market's $224 (as of 2026-08-07) sits at roughly FY2027 EPS base 8–9 × FY2028 EPS growth 5–17.5 — closer to the Antithesis case than the Panel, pre-verification (their house case).
Market data for reconciliation only — not house evidence · NASDAQ close via Bloomberg
- implied share price
- $261.40
The model projects FY2027 EPS forward one year at a specified growth rate, applies an exit multiple to the resulting FY2028 EPS, and discounts the terminal value by 16.5% if the market-share-erosion lever is activated. The output is an implied share price in $/share, representing fair value under the chosen scenario assumptions.
You are standing exactly where we are. Move an assumption and this panel prices the disagreement.
Swing in implied share price across each assumption's full range, holding your other settings. Recomputed where you stand, never cached.
- $261.40
- $261.40
- not published yet
No reader distribution is published for this model yet. What you can see is our number and yours. Save your assumptions and your implied share price to the record and you become part of the distribution the moment it reaches 25 readers.
0 of 25 readers on record · nothing is shown until 25.
One position per browser. No account, no email: a random token in this browser's storage. Saving again replaces your previous position.
A switch that applies a 16.5% discount to the fair value estimate if you believe NVIDIA is losing market share to competitors like AMD, Intel, or custom hyperscaler chips. When activated (set to 1), it models the scenario where dual-sourcing and in-house accelerators are materially eroding NVIDIA's dominance.
Why we hold our number · Off by default. Cites e2 and e3 for the share-erosion claim, not the superseded margin figure — retained under Editor memo 2026-08-07. Switched on, it applies the lever's stated 15–18% fair-value haircut (midpoint 16.5%): the case where hyperscaler dual-sourcing erodes NVIDIA's share faster than the thesis allows.
Downstream: moving it changes every model-derived figure in this workspace for you alone. It does not change the published article, the house case, or what any other reader sees.
WHAT AM I DECIDING?
A contested lever is a number the adversary attacked and we declined to move; our reason is printed beside it.
Downstream: moving it changes every model-derived figure in this workspace for you alone. It does not change the published article, the house case, or what any other reader sees.
We treat reader behaviour on this page as a research input. Every signal we recorded from this browser is listed below, with what we read into it. No account, no email, no cross-site tracking: a random token in this browser's storage, which you can clear here.
Nothing recorded yet from this browser on this page.