Position · Core / Quality Compounders
Advanced Micro Devices (AMD) Medium Conviction
Compare vs. competitors: AMD vs. Intel vs. NVIDIA →
Read the full Q1 2026 earnings report →
Subscribe to AMD updates (RSS) →
Download full 3-statement model (.xlsx) →
Thesis StatementAMD is the closest credible second source to NVIDIA for AI accelerators at scale, but the multiple already assumes years of Data Center execution.
Core Thesis
AMD is a leading designer of CPUs and GPUs for PCs and data centers, and is emerging as the closest thing to a credible second source for AI accelerators through its MI-series line, with traditional server CPU share gains compounding underneath that story.
Financial Metrics
- Market Cap$882B
- P/E (TTM)182.11
- EPS (TTM)$2.97
- Div. Yield0.00%
- Price$517.82
- ConvictionMedium Conviction
Bear Case
AI GPU share gains disappoint against a dominant incumbent, and at 182x trailing earnings the stock has essentially no margin for a growth miss — a single disappointing data-center quarter could compress the multiple sharply.
Investment Thesis
- The Thesis
- AMD is a leading designer of CPUs and GPUs for PCs and data centers, and is emerging as the closest thing to a credible second source for AI accelerators through its MI-series line, with traditional server CPU share gains compounding underneath that story.
- The Catalyst
- Every large AI buyer wants a credible alternative to the dominant GPU supplier, and continued MI-series adoption by hyperscalers would validate AMD as that real second source rather than a niche alternative.
- The Risk
- At 182x trailing earnings the stock has essentially no margin for a growth miss — if AI GPU share gains disappoint against the dominant incumbent, a single weak data-center quarter could compress the multiple sharply.
- The Connection
- Core to my semiconductor theme alongside SNDK, MU, and INTC — a direct bet on AI-accelerator diversification rather than the dominant incumbent.
Pre-Mortem Thesis Invalidation Parameters
Codified in advance, before any of these have happened, so a future decision to hold or exit isn't rationalized in the moment. If a condition below is met, the thesis as written is invalidated and the position gets re-underwritten from scratch — not automatically sold, but automatically questioned.
| Metric / Event | Automatic Review Trigger |
|---|---|
| Data Center Revenue Growth (YoY) | Decelerates below 30% YoY for 2 consecutive quarters. |
| AI-GPU Share Commentary | A hyperscaler customer publicly discloses pulling back MI-series adoption in favor of the incumbent supplier. |
I've built every PC I own around AMD parts
This isn't secondhand research — I've built my own computers, and AMD has been the component of choice in every build for one consistent reason: price-to-performance. Ryzen CPUs and Radeon GPUs have repeatedly delivered benchmark performance that competes with parts costing significantly more, which is exactly the kind of value proposition that turns a challenger into a real second source. Sourcing components myself means I've paid close attention to build guides, benchmark comparisons, and community sentiment across the DIY PC space for years, and AMD's reputation there has shifted from "budget alternative" to "legitimate high-end option" over that time — the same trajectory the data-center bull case depends on.
This is anecdotal, not financial data, and I'm not claiming enthusiast-PC sentiment predicts data-center AI-GPU share gains. But it's a real, first-hand data point on brand trust and product quality in AMD's original stronghold, and it's part of why I find the "credible alternative supplier" thesis plausible rather than purely narrative-driven.
| Market cap | $882B |
|---|---|
| P/E ratio (TTM) | 182.11 |
| EPS (TTM) | $2.97 |
| Dividend yield | 0.00% |
| Shares outstanding | 1.63B |
| Sector | Semiconductors & related devices |
| EBITDAEarnings Before Interest, Taxes, Depreciation, and Amortization — a measure of operating profitability before financing and accounting decisions. EV/EBITDA compares a company's full value (including debt) to this figure, often used to compare companies with different capital structures. | Quarterly EBITDA $2.23B (Q1 FY26) |
| PEG ratioPrice/Earnings-to-Growth: the P/E ratio divided by expected earnings growth. Below 1.0 is often read as cheap relative to growth; above suggests the market is pricing in a lot of future growth already. Different providers use different growth-rate assumptions, so figures vary by source. | 7.52 |
| Capex | $282M (quarterly) |
Valuation Logic
At 182.11x trailing earnings, this is a growth, not a value, multiple — the multiple becomes fragile when even a modest slowdown changes the expected downside faster than earnings can catch up; it only works if Data Center revenue keeps compounding near the 46%+ guided pace, since any deceleration compresses the multiple faster than earnings can catch up.
About the business
AMD designs digital semiconductors for PCs, gaming consoles, data centers (including AI), industrial, and automotive markets. Its traditional strength is CPUs and GPUs for PCs and data centers, and it's emerging as a real second source for AI GPUs and related hardware. It also supplies chips for the Sony PlayStation and Microsoft Xbox.
Why I own it
AMD is my bet that the AI accelerator market doesn't stay a single-vendor story forever — every large buyer wants a credible alternative supplier, and AMD is the closest thing to one. It's also my best-performing position, up nearly 160% on cost. The 182x trailing P/E is the honest cost of that bet: the market is pricing in a lot of AI-GPU share gain that hasn't fully shown up in earnings yet.
Risk/Reward Profile
| Bull Case | Bear Case |
|---|---|
| AMD's MI-series accelerators keep taking real share from the incumbent AI GPU leader as hyperscalers diversify their supply chains, data-center GPU revenue scales fast enough to catch up to the multiple, and traditional CPU share gains in servers keep compounding underneath it. | AI GPU share gains disappoint against a dominant incumbent, and at 182x trailing earnings the stock has essentially no margin for a growth miss — a single disappointing data-center quarter could compress the multiple sharply. |
Scenario Matrix
The single variable this position is most sensitive to is AI-accelerator share gains against the dominant incumbent. This maps each case against that variable directly.
| Scenario | AI accelerator share trajectory | Valuation implication | My estimated probability |
|---|---|---|---|
| Bull | MI-series share gains accelerate as hyperscalers actively diversify away from a single supplier. | Data-center revenue scales fast enough to catch up to today's multiple, supporting further upside. | ~25% |
| Base | AMD holds a solid #2 position with steady but unspectacular share gains. | Growth doesn't accelerate enough to justify the multiple on its own — forward return depends on earnings catching up, not further re-rating. | ~50% |
| Bear | Share gains disappoint against a dominant incumbent that defends its position. | At 182x trailing earnings, essentially no margin for a growth miss — a single weak data-center quarter compresses the multiple sharply. | ~25% |
Download this position's data
Fundamentals, scenario matrix, and risk/reward table — exported exactly as published on this page, no reformatting.