Is Advanced Micro Devices Inc. (AMD) overvalued?

How we measure value for AMD

Our fair-value range for Advanced Micro Devices Inc. is triangulated across multiple discounted-cash-flow lenses, peer multiples, scenario-weighted earnings, and — where the archetype permits — owner earnings or reverse DCF. Because AMD is classified as a mature compounder stock, we calibrate every parameter (discount rate, terminal growth, deceleration curve, terminal multiple, scenario probabilities) to that archetype rather than applying generic defaults. The result is the $67.5–$160 band, not a single point estimate, because point estimates collapse a real range of possibilities into false precision.

The discount rate that anchors the model is 13.31% (cost of equity, CAPM with adjusted beta) for earnings-based DCFs, and 13.26% for any free-cash-flow-to-firm work. We do not mix these — using WACC to discount EPS double-counts capital structure, a common amateur mistake — and the bear case is built first, before any bullish triangulation, to counteract anchoring.

Where the gap to fair value comes from

AMD currently trades at a forward P/E of about 35.3× with a PEG of 1.76, against $742.2B of market capitalization. The composite fair-value bridge from current price ($455) to midpoint ($113) reflects the difference between what the market implies about future earnings, growth durability, and capital efficiency — and what our analyst-grade models say those numbers should look like under base-case assumptions. The further the live multiple drifts from the multiple our models would tolerate, the wider the upside-or-downside number gets.

On a relative basis the peer table inside the full report compares AMD with its closest sector neighbours on P/E, EV/EBITDA, return on equity, and operating margin. PEG-adjusted peer multiples — forward P/E divided by integer-percent growth — are the cleanest single-screen "is it expensive?" lens for mature compounder archetypes; the report shows where AMD sits on that axis.

What would change the verdict

The 5×5 sensitivity matrix in the full report stress-tests AMD's fair value across plausible cost-of-capital and terminal-growth combinations. Two reasons we publish the matrix: first, no analyst is right about Ke or terminal growth to two decimal places; second, the model output is more sensitive to those two inputs than to almost anything else. If the company's adjusted Ke creeps higher or terminal growth decays faster than our base case, the midpoint compresses; if either is too pessimistic, it expands.

On top of the matrix we publish five formal stress tests with quantified fair-value impact (recession, margin compression, capital-cycle late stage, regulatory shock, key-product cannibalization, etc., calibrated to mature compounder risk), and for tickers with earnings inside 60 days we add an earnings decision tree with beat/inline/miss branches. Together they answer the practical question: how much would have to go wrong for the bull case to evaporate, and how much would have to go right for the bear case to break.

Bottom line — is AMD overvalued today?

Our current rating for AMD is Sell, with a 52/100 confidence score and a 6.5/10 moat assessment. The valuation evidence looks meaningfully overvalued, but the rating is not a price call — it is a function of the gap between price and fair value, the durability of the moat, and the dispersion of the bull/base/bear distribution. Position-sizing rules in the full report quantify how much of that view to express at the current price versus waiting for a wider margin of safety.

This page summarises the fair-value question; the canonical report at /stocks/amd/analysis expands the same evidence into 14 sections — bear case first — including the full sensitivity grid, scenario tree, and the assumption ledger.

Frequently asked questions