Methodology
How AerInvest Actually Values a Stock
Worked example: NVDA, live as of September 5, 2026
Most stock-research tools give you a verdict — a rating, a color, a badge — without showing how they got there. AerInvest's fair value is built from three components, blended into one number. Here's exactly how, using the live analysis for NVDA (NVIDIA Corporation) as the example — these are real numbers pulled from the same engine every AerInvest user sees, not illustrative placeholders.
NVDA
NVIDIA Corporation · Stock
Current price
USD 230.36
12-month price scenarios
USD 217.00
USD 265.00
USD 323.00
1. Discounted Cash Flow (40% weight)
A 5-year free-cash-flow projection, discounted back at a WACC derived from the company's beta and debt/equity — not a flat assumption. NVDA's valuation framework here is: Stock range model based on live price, 52-week position, market cap, valuation metrics, profitability, balance sheet, income profile, and data-quality penalties.. Key assumptions feeding it:
- Current price: USD 230.36 — Primary anchor for scenario analysis.
- 52W position: 91/100 — Used as a momentum and drawdown context proxy.
- Market cap: $5.58T — Used as a rough liquidity and business-scale proxy.
- P/E ratio: Unavailable — Used as one valuation sanity check, not a complete model.
2. Relative Multiples (35% weight)
P/E, EV/EBITDA and P/FCF compared against sector peers, adjusted for quality — a company growing faster or carrying less debt than its peer group earns a premium multiple, not the peer median applied blindly. Peer context for NVDA: Peer context should be built around the Semiconductors industry. The comparison page now exposes side-by-side metrics; a larger peer universe can expand this further.
3. AI Quality Score (25% weight)
An AI model scores the company across several fundamental factors and adjusts the blended fair value by up to 15% based on quality signals a pure multiple can miss. The factor breakdown for this analysis:
- Fundamental Intelligence96/100 — Constructive contribution with 34% model weight and explicit data-quality penalties.
- Market Intelligence79/100 — Constructive contribution with 18% model weight and explicit data-quality penalties.
- Sentiment Intelligence70/100 — Constructive contribution with 12% model weight and explicit data-quality penalties.
- Macro Intelligence69/100 — Balanced contribution with 14% model weight and explicit data-quality penalties.
- Risk Intelligence82/100 — Constructive contribution with 22% model weight and explicit data-quality penalties.
The three scenarios, not one number
A single fair-value figure hides how much of the estimate depends on assumptions that could go either way. The bear/base/bull range above shows that spread explicitly — worst case USD 217.00, base case USD 265.00, best case USD 323.00. If the bear and bull cases are close together, confidence is higher; if they're far apart, that spread itself is information.
What could break this thesis
- Market cap was unavailable from the fallback chart endpoint.
- FMP stock news unavailable: FMP news/stock-latest returned 429.
- FMP ETF holdings unavailable: FMP etf/holdings returned 429.
Not investment advice
This is a live analysis pulled from the same engine every AerInvest user sees — it will change as NVDA's price, financials and estimates change. It illustrates the methodology, not a standing recommendation. See the full disclosure.