AMD — Advanced Micro Devices, Inc.
Advanced Micro Devices, Inc. (AMD), established in 1969 and headquartered in Santa Clara, California, operates as a global leader in the semiconductor industry. The company organizes its extensive operations into two primary segments: Computing and Graphics, and Enterprise, Embedded and Semi-Custom.
Price chart & technical analysis
Loading chart…
As of 2026-09-13, AMD trades at $516.13 (market cap ~$841.6B), up 166.9% over six months and roughly 12% below its late-June record high of $584.73, after Q2 2026 record revenue of $11.54B (+50.1% YoY) with Data Center revenue doubling to $6.7B. The fundamentals are genuinely accelerating, but at 131.0x trailing earnings and 20.4x trailing sales (vs. a 122.96x P/E and 8.7x P/S five-year average), this is a great business at a demanding price.
As of 2026-09-13 the tape and the sell-side agree. Q2 2026 (reported Aug 4, 2026) delivered record revenue of $11.54B, +50.1% YoY, EPS beat (~$1.66 vs ~$1.61 consensus) and Data Center revenue doubled to $6.7B. The late-August Anthropic partnership - up to 2 gigawatts of Instinct GPUs on AMD's Helios rack-scale platform - plus management's raised AI accelerator TAM of roughly $3T by 2030 (up from ~$2T) reinforced AMD as the credible merchant-silicon alternative to Nvidia. Raymond James upgraded to Strong Buy with a $641 target on Aug 25; the ratings mix is 1 strong buy / 49 buy / 20 hold / 0 sell, with a $598.24 consensus target and $654.63 average of recent targets. Price confirms the mood: +8.07% in the week and +6.87% in the month through 2026-09-13, RSI at 57.96. The one blemish is that shares dipped post-earnings on in-line gross-margin guidance and sit 11-17% below the June peak - bullish, but not euphoric.
Revenue has run from $5.80B (Q3 2023) to $11.54B (Q2 2026) and the growth rate is accelerating, not fading: YoY was +17.6% (Q3 2024), +24.2% (Q4 2024), +31.7% (Q2 2025), +35.6% (Q3 2025), +34.1% (Q4 2025), +37.9% (Q1 2026) and +50.1% (Q2 2026). Gross margin has structurally stepped from 46-47% across FY2023 to 50-51% through FY2024/early FY2025 and 52.8-54.3% over the last three quarters; the ugly 39.8% print in Q2 2025 is a one-off (an export-control charge that pushed operating income to -$134M on $7.69B of revenue), not a trend break. Operating margin tells the same story: 3.9-11.4% through 2023-2024 versus 13.7%, 17.1%, 14.4% and 17.2% over the last four quarters. R&D is up 68% since Q3 2023 against 99% revenue growth, so the margin expansion is operating leverage, not just pricing. Diluted EPS went from $0.18 (Q3 2023) to $1.38 (Q2 2026). Cash backs the earnings: TTM free cash flow of $8.40B is roughly 3x the $2.75B of the prior four quarters and the accruals ratio is -0.08% (cash slightly ahead of reported net income). Caveats: Q2 2026 net income of $2.297B exceeded operating income of $1.990B, so non-operating/tax items flattered GAAP EPS, and capex jumped to $808M from $222M in Q4 2025 as capacity ramps. Net: an inflection, not a blip.
Expensive, and not only in absolute terms. As of 2026-09-13: 131.0x trailing P/E vs a 122.96x five-year average; 20.4x trailing sales vs an 8.7x five-year average; 12.5x book; 79.6x EV/EBITDA; roughly 1.0% FCF yield on an $841.6B market cap. The bull case rests on forward math - next-FY consensus EPS of $7.619 implies a ~67.7x forward P/E and the quoted PEG of 1.058 - but that PEG is arithmetic flattery, since the denominator is a one-year earnings spike off a charge-depressed base rather than a durable growth rate. Even if AMD hits the $50.87B next-FY revenue consensus, it still trades around 16.5x forward EV/sales. Verdict: expensive relative to its own history and to its current 15.7% operating margin / 5.4% ROIC, though not irrational if the data-center ramp compounds for years.
The market is paying now for growth that has partly arrived and partly not. Half the thesis is in the numbers already (50.1% revenue growth, Data Center doubling to $6.7B, TTM FCF tripling); the other half is still a promise - MI450 volume production, Helios rack-scale execution, and the multi-gigawatt Meta/Anthropic commitments converting into revenue in 2027+, against a TAM management now pegs at ~$3T by 2030. At ~67.7x forward earnings, that promise is heavily discounted. Practically: earnings beats alone won't move the stock much (Q2 beat, shares still dipped on gross-margin guidance). What re-rates or de-rates the multiple is gross-margin trajectory and firm, dated revenue commitments from the gigawatt-scale deals. If Q3 (Nov 3, 2026) shows revenue above the ~$12-13B sequential run-rate implied by FY consensus with GM pushing toward the mid-to-high 50s, the multiple holds; if revenue beats on lower-margin rack-scale mix, expect multiple compression even as EPS rises. At this multiple, price is a function of the multiple, not the estimate.
As of 2026-09-13, RSI(14) stood at 57.96 - firm but not overbought - with price 7.07% above the 20-day SMA (~$482), 3.94% above the 50-day (~$497) and 48.78% above the 200-day (~$347). Returns through that date: +2.49% on the day, +8.07% over one week, +6.87% over one month, +0.89% over three months, +166.89% over six months, against a 52-week range of $149.85-$584.73. The key read is that three-month flat patch inside a six-month melt-up: AMD consolidated sideways after the late-June peak of $584.73 and has just broken back up on strengthening RSI, so momentum is intact and there is room before overbought. But the 200-day sits 49% below spot, so there is no nearby structural support and the immediate test is the $584.73 high. Entry implication: chasing here is high-beta work; the cleaner entries were the flat three-month base just exited, or a pullback toward the 50-day near $497.
- Hyperscaler AI capex cycle: Microsoft, Meta, Anthropic and OpenAI data-center builds are the single biggest swing factor for AMD's Data Center line; any ROI-driven capex digestion would hit AMD first and hardest.
- US export controls on high-performance chips to China: already caused a real gross-margin hit and inventory charge in Q2 2025, and remains a recurring, hard-to-timing risk to both revenue and reported margins.
- Rates and duration risk: with a 2.476 beta and a 131x trailing P/E, AMD is among the most rate-sensitive large caps - its value is back-end loaded, so rising yields compress the multiple mechanically.
- Competitive intensity: Nvidia's entrenched share in AI accelerators and its push into custom PC processors, plus Intel's CPU roadmap, frame AMD as the challenger rather than the incumbent in its most valuable market.
- Supply chain: advanced packaging (CoWoS), HBM availability and TSMC foundry allocation gate the MI450/Helios ramp - visible in the capex step-up to $808M in Q2 2026 from $222M in Q4 2025.
- Q3 2026 earnings on Nov 3, 2026 - the first clean read on Helios/MI450 supply execution and forward gross-margin trajectory, the two things that drove the post-Q2 selloff.
- Conversion of the Anthropic (up to 2 GW) and Meta multi-gigawatt commitments into firm, dated revenue, plus MI450 production velocity through the back half of 2026.
- Continued server CPU share gains against Intel, the explicit rationale behind Raymond James' Aug 25 upgrade to Strong Buy and $641 price target.
- Further upward revisions to management's AI accelerator TAM (~$3T by 2030) and any additional large-scale AI infrastructure wins that validate the Helios rack-scale platform.
- Sell-side target drift: consensus sits at $598.24 with recent targets averaging $654.63 and a Street high of $1,250 - upgrades and target raises remain a mechanical tailwind.
- Valuation risk is the dominant one: 20.4x TTM sales versus an 8.7x five-year average and 79.6x EV/EBITDA means a 20% multiple compression erases roughly $165B of market cap even if earnings fully deliver.
- Gross-margin execution and mix: Q2 2026 already sold off on in-line GM guidance, and heavier rack-scale/Helios system mix (lower margin than stand-alone GPUs) could structurally cap gross margin well below the 60% bulls expect.
- Nvidia's entrenched dominance and its entry into custom PC processors - AMD is the challenger in AI accelerators, and share gains are not guaranteed at any price.
- Geopolitical/export-control risk: the 2025 MI308 episode shows how quickly a licensing change converts into charges and forgone revenue; recurrence would hit both the P&L and the multiple.
- Cyclicality the trailing numbers cannot flag: AI capex and semis are cyclical, and the deterministic value score (built entirely on trailing YoY growth, margins and ROIC) looks best precisely when the cycle is strongest - it has no mechanism to warn of a demand air pocket.
- Customer concentration: a handful of hyperscalers and AI labs drive the incremental growth; one delayed or rescoped deployment materially moves the outlook.
- Earnings quality nuance: Q2 2026 net income of $2.297B topped operating income of $1.990B on non-operating/tax items, so GAAP EPS growth is flattered relative to core operating momentum.
AMD is a genuinely improving business - accelerating 50.1% revenue growth, gross margin up from ~47% to ~54%, operating margin from ~4% to ~17%, net cash with 0.064 debt/equity and a 2.61 current ratio - and it is the only credible merchant-silicon challenger to Nvidia in AI infrastructure. The problem is entirely price: at $516.13 you pay 20.4x trailing sales and ~67.7x next-year earnings for that, and my own fair-value work lands near $429. The mandatory lean reads Cautiously Bullish because the fundamental band is good while the valuation overlay is rich, and I agree directionally - but the honest read is that the shares are running ahead of what the current numbers support. Own it as an AI-infrastructure growth position, not for valuation support; scale in on weakness toward the low $400s or a retest of the 50-day rather than chasing the current re-acceleration. This is analysis, not personalized advice.
Revenue growth has run from $5.80B (Q3 2023) to $11.54B (Q2 2026) and the growth rate is accelerating, not fading: YoY was +17.6% (Q3 2024), +24.2% (Q4 2024), +31.7% (Q2 2025), +35.6% (Q3 2025), +34.1% (Q4 2025), +37.9% (Q1 2026) and +50.1% (Q2 2026). Gross margin has structurally stepped from 46-47% across FY2023 to 50-51% through FY2024/early FY2025 and 52.8-54.3% over the last three quarters; the ugly 39.8% print in Q2 2025 is a one-off (an export-control charge that pushed operating income to -$134M on $7.69B of revenue), not a trend break. Operating margin tells the same story: 3.9-11.4% through 2023-2024 versus 13.7%, 17.1%, 14.4% and 17.2% over the last four quarters. R&D is up 68% since Q3 2023 against 99% revenue growth, so the margin expansion is operating leverage, not just pricing. Diluted EPS went from $0.18 (Q3 2023) to $1.38 (Q2 2026). Cash backs the earnings: TTM free cash flow of $8.40B is roughly 3x the $2.75B of the prior four quarters and the accruals ratio is -0.08% (cash slightly ahead of reported net income). Caveats: Q2 2026 net income of $2.297B exceeded operating income of $1.990B, so non-operating/tax items flattered GAAP EPS, and capex jumped to $808M from $222M in Q4 2025 as capacity ramps. Net: an inflection, not a blip.
The 74.6 / 'Good Fundamentals' band is a fair read of the trailing fundamentals: growth at 85 (accelerating revenue, tripling FCF), financial health at 85.1 (net cash, 0.064 D/E, 2.61 current ratio, strong FCF trend) and a slightly-below-par quality score of 51.5 dragged by a 15.7% operating margin and 5.4% ROIC - both of which are artifacts of a goodwill-heavy balance sheet and the trailing window rather than deteriorating operations. If I adjusted anything, I would nudge quality up a few points (not 10-25) because the trailing four quarters include Q2 2025's export-control charge, which knocked gross profit by roughly $0.8B and pushed that quarter's operating margin to -1.7%; ex-charge, TTM operating margin is closer to ~17.7% and the band would only strengthen within 'Good Fundamentals.' I deliberately do not dock the score for richness, because the valuation overlay ('Rich vs. its own history') is excluded from the weighted number by design and already carries that signal. Note the score's one blind spot here is the mirror image of the cyclical-trough warning: AMD is near a cyclical high in AI capex, so a trailing-only model flatters it - the fundamentals band is right, but the score says nothing about the price paid. No deviation from the mandated band/valuation mapping is warranted; no dated corporate event (acquisition, restatement, fraud, bankruptcy) exists in the brief to override it.
AI-generated analysis for informational purposes only, not financial advice.