Advanced Micro Devices, Inc.

Executive Summary

SELL, Low conviction. AMD rates a SELL, Low conviction. The business is materially stronger than it was two years ago, with Q2 2026 revenue up 50% YoY, data center up 107%, and management guiding Q3 revenue of $13.0B ± $0.3B, but the stock sits 2.3% below its all-time high and 72.9% above its 200-day average while trading far above peer valuation. The issue is not execution today, it is that the market value of $1.026T already assumes a long runway of AI growth with limited stumbles.

Investment Thesis

AMD's operating story is better than the valuation debate suggests. The company has turned itself into a real AI and server compute contender. Q2 2026 revenue rose 50% YoY to $11.5B, data center rose 107% to $6.7B, and management now says server revenue should grow more than 80% YoY in 2H26 and more than 70% for full-year 2027. That is not promotional fluff. It is backed by a strong beat streak, improved supply, a richer mix, and the start of Helios shipments later in Q3. If the only question were business momentum, the stock would deserve respect.

The problem is the entry price. At $630.63, AMD is worth $1.026T on FY2025 revenue of $34.6B and EPS of $2.67. That is 24.85x sales and 159.71x earnings, far above peer averages and hard to justify with 7.4% operating margin and 6.4% net margin. Bulls will correctly argue that FY2025 understates the forward earnings run rate because AI revenue is ramping fast. That is true. But the stock has already capitalized that argument aggressively, trading just 2.3% below the high after a 31.61% one-month surge. At this level, investors are no longer paying for AMD to win more AI business. They are paying for AMD to win enough high-margin AI business, quickly enough, with few execution slips, to support a trillion-dollar valuation.

The case now depends on three things. First, AMD must translate revenue growth into meaningfully higher margins, even though management admits data center AI gross margin remains slightly below corporate average during ramp. Second, the non-AI businesses cannot deteriorate too much, because Gaming fell 31% YoY and the PC market is expected to soften in 2H26. Third, the company must prove that its AI moat will strengthen, not just that demand is robust this year. EPYC share gains and Instinct traction are real. The open question is whether software, systems, and customer entrenchment are enough to make AMD a lasting AI platform rather than a valuable but secondary supplier.

Low conviction reflects respect for execution. A bearish call on AMD today is not a call against the company. It is a call against the price paid for the company.

Key Metrics

Financial Strength

Financial risk is low. AMD carries only $4.3B of debt, with Debt/Equity of 0.06 and Debt/EBITDA of 0.59, so the balance sheet is not the issue. The better debate is quality of earnings conversion: FCF was $1.6B in FY2025, respectable but not yet commensurate with a $1T market cap, and inventory has risen to $8.5B to support AI demand. Margin trajectory is improving at the quarterly level, with Q2 EBITDA margin at 29.1% and gross margin at 56%, but reported full-year profitability remains thin enough that valuation is doing most of the work in the equity story.

Competitive Position

AMD's moat rests on x86 server CPU share gains with EPYC, a credible alternative AI accelerator stack with Instinct, and deep engineering that is reflected in an Innovation Investment score of 100. That moat is real but not impregnable: it is strongest where customers want a second source against an incumbent, and weaker where software ecosystems and platform lock-in dominate. Over the next 3-5 years, durability depends less on raw silicon and more on whether ROCm, rack-scale systems, and customer support narrow the gap enough for AMD to be a standard AI deployment choice rather than a tactical supplement.

Management & Guidance

Management guided Q3 2026 revenue to approximately $13.0B plus or minus $300M, non-GAAP gross margin to 56%, and non-GAAP operating expenses to $3.65B. More importantly, Lisa Su lifted the medium-term slope, saying, "we now expect server revenue to grow more than 80% year-over-year in the second half of 2026 and more than 70% for the full year 2027, off a much higher base." Credibility is fairly high because AMD has beaten EPS in 7 of the last 8 quarters and has generally delivered on the product and supply ramp promised in the 2024 period, though investors should remember that AI gross margins are still described as slightly below the corporate average during ramp.

Segment Analysis

• Data Center (about 58% of Q2 2026 revenue): $6.7B, up 107% YoY, driven by EPYC and more than doubling Instinct revenue; management expects segment revenue to more than double YoY in 2027.
• Company total revenue ($11.5B in Q2 2026): Record quarter, up 50% YoY; Q3 guide is $13.0B ± $0.3B.
• Gaming (not disclosed in the provided data): Revenue fell 31% YoY on weaker semi-custom and softer graphics demand.
• Embedded (not disclosed in the provided data): Management cited design wins >$18B, putting embedded on track for another record year.
• Client / PC (not disclosed in the provided data): Ryzen remains a driver, but management expects the PC market softer in 2H26 as higher memory and component costs weigh on demand.
• Geography (not disclosed in the provided data): Geographic revenue split was not provided, so the revenue read is product-led rather than region-led.

Capital Allocation

Capital allocation looks disciplined. AMD generated $1.6B of FCF on $34.6B of FY2025 revenue, spent $0.8B on capex, and returned only $0.2B through buybacks with no dividend, which is sensible given the product and inventory ramp. The balance sheet is lightly levered, so the main capital question is not solvency, it is whether inventory and AI working-capital needs translate into sustainably higher cash earnings rather than just revenue growth.

Management Execution & Track Record

Management's execution since the 2024 Q3 call has been strong. Two years ago, AMD guided Q4 2024 revenue to $7.5B ± $300M, 54% gross margin, and said Data Center GPU revenue would exceed $5B in FY2024 while supply would remain tight but sufficient for significant 2025 growth. By Q2 2026, quarterly revenue reached $11.5B, quarterly gross margin expanded to 56%, EBITDA margin improved to 29.1% from 22.8% in the 2024 Q3 period, and the company moved from data center revenue of $3.5B in 2024 Q3 to $6.7B in 2026 Q2. Supply also appears to have improved as promised: server CPU supply was tight in 1H26, but management now says improved supply supports a raised 2H26 outlook, while HBM allocation visibility for 2027 is described as good. The earnings record supports credibility, with 7 beats in the last 8 quarters, including +15.9%, +6.2%, and +2.5% surprises in the last three reports. The one caveat is that the AI ramp is still margin-dilutive at the gross margin line versus corporate average, so execution has been excellent commercially, but the full earnings power is not yet proven.

Positioning & Flows

Ownership is institutional and crowded, with the top three institutions holding 66.9% and insiders owning just 0.4%. Short interest is modest at 2.46%, or 39.98M shares, with 1.5 days-to-cover, which limits classic squeeze fuel and suggests most skeptics have already stepped aside rather than leaned in. Options positioning is supportive near term, with a positive GEX regime and zero-gamma at $591.58, but that matters more for trading path than for fundamental value after a 31.61% one-month move.

Bull Case

• Q2 2026 revenue rose 50% YoY to $11.5B and EBITDA reached $3.35B, lifting EBITDA margin to 29.1%, up 1,970 bps YoY.
• Data center revenue grew 107% YoY to $6.7B and now represents roughly 58% of company revenue, showing AMD has become a real AI and server compute platform, not just a PC CPU story.
• Management raised the forward AI slope materially, with Q3 2026 revenue guided to $13.0B ± $0.3B, server revenue growth >80% YoY in 2H26, and data center segment revenue expected to more than double YoY in 2027.
• Balance sheet risk is low, with Debt/Equity 0.06, Debt/EBITDA 0.59, and Current ratio 2.61, giving AMD room to fund inventory and product ramps.

Bear Case

• FY2025 reported profitability does not support the stock price, with EPS of $2.67, operating margin of 7.4%, and net margin of 6.4% against a $1.026T market cap.
• AI mix is helping growth but not peak economics yet, because management said data center AI gross margin remains slightly below corporate average during ramp, even as inventory rose to $8.5B.
• Non-AI businesses are not clean, with Gaming down 31% YoY and management expecting a softer PC market in 2H26 due to higher memory and component costs.
• The stock has become technically stretched, up 12.65% in 1 week, 21.79% in 3 months, and only 2.3% below the high, which raises downside if any guide merely meets instead of beats.

Valuation & Price Target

Engine price target $336.34 (-46.7% vs current), Low confidence.

• P/E Multiple: $322.11 (14% weight)
• PEG (growth-adjusted): $103.23 (20% weight)
• P/S Multiple: $305.2 (20% weight)
• Quality-adjusted: $676.54 (20% weight)
• DCF: $46.05 (5% weight)

This view aligns with the quant baseline of SELL, though for a narrower reason than business quality. The quantitative engine sets a $336.34 price target, -46.7% from the current $630.63, with low confidence; that direction is hard to ignore when the stock trades at 159.71x P/E and 24.85x P/S against a peer average 39.4x P/E. The disagreement is with the severity implied by some valuation methods, not the direction: AMD is not a broken story, because Q2 revenue grew 50% YoY, data center grew 107%, and management is guiding $13.0B Q3 revenue plus a stronger 2027 outlook. Sell-side consensus remains Strong Buy (26 buy / 4 hold / 0 sell), but that is a low-signal context item here. The stock can remain expensive for a while, yet the combination of trillion-dollar market cap, modest FY2025 margins, and a rate backdrop of 5.18% on the 10-year keeps the risk/reward skewed down.

Risk Assessment

Risk score 37/100 (Low). The biggest risk to a bearish view is that AMD converts its 2027 data center revenue expected to more than double YoY into a far larger earnings base quickly enough to make today's 159.71x P/E look temporarily misleading.

The Bottom Line

AMD rates a SELL, Low conviction. The company is executing at a high level, and the AI ramp is real, but the stock already reflects a near-best-case path with little tolerance for slower margin conversion or weakness outside data center. A move toward the quant base case of $336.34 would not require business deterioration, only valuation normalization. The call would change with either a materially lower entry price or clear evidence that AI revenue is converting into sustainably higher company-wide earnings power faster than the current FY2025 margin profile suggests.