VOL. XCIV, NO. 247
★ WIDE MOAT STOCKS & COMPETITIVE ADVANTAGES ★
PRICE: 0 CENTS
Advanced Micro Devices, Inc.
AMD · The Nasdaq Stock Market LLC
Weighted average of segment moat scores, combining moat strength, durability, confidence, market structure, pricing power, and market share.
Request update
Spot something outdated? Send a quick note and source so we can refresh this profile.
Overview
Advanced Micro Devices, Inc. is a fabless semiconductor company supplying EPYC server CPUs, Instinct AI GPUs, Ryzen client CPUs, Radeon graphics/semi-custom gaming SoCs, and embedded processors/FPGAs/adaptive SoCs. FY2025 revenue mix was roughly 48% Data Center, 31% Client CPUs, 11% Gaming and 10% Embedded after splitting AMD's combined Client and Gaming disclosure by reported sub-revenue. AMD's strongest moat is design-in qualification around EPYC/Instinct at hyperscalers and embedded/custom SoC programs, reinforced by x86 compatibility, chiplet/product know-how, ROCm/open ecosystem work and Xilinx-derived adaptive-compute IP. Counter-pressures are severe: Nvidia's AI/CUDA and graphics dominance, Intel CPU recovery, Arm/custom silicon, console-cycle dependence, export controls, memory/foundry constraints and cyclical embedded inventory digestion.
Primary segment
Data Center
Market structure
Oligopoly
Market share
28.8%-41.3% (estimated)
HHI: 5,899
Coverage
4 segments · 8 tags
Updated 2026-05-01
Segments
Data Center
Global data-center compute semiconductors, centered on x86 server CPUs and merchant AI accelerators
Revenue
48%
Structure
Oligopoly
Pricing
moderate
Share
28.8%-41.3% (estimated)
Peers
Client CPUs
Global client x86 CPUs for desktops and notebooks
Revenue
30.7%
Structure
Duopoly
Pricing
moderate
Share
29.2%-31.2% (estimated)
Peers
Gaming
PC discrete gaming GPUs and semi-custom game-console SoCs
Revenue
11.3%
Structure
Oligopoly
Pricing
weak
Share
5% (estimated)
Peers
Embedded
Embedded processors, FPGAs, adaptive SoCs, and SOMs for industrial, communications, automotive, aerospace, healthcare, and data-center applications
Revenue
10%
Structure
Oligopoly
Pricing
moderate
Share
—
Peers
Moat Claims
Data Center
Global data-center compute semiconductors, centered on x86 server CPUs and merchant AI accelerators
FY2025 revenue share uses Data Center revenue of $16.635 billion divided by total revenue of $34.639 billion. Operating-profit share uses reportable segment operating income before All Other corporate items.
Design In Qualification
Demand
Design In Qualification
Strength
Durability
Confidence
Evidence
EPYC CPUs and Instinct GPUs require long technical qualification, roadmap alignment, software validation, and deployment work with hyperscalers, OEMs, and ODMs.
Erosion risks
- Nvidia CUDA and full-stack systems remain the dominant AI accelerator standard
- Cloud customers shift workloads to internal ASICs
- Export controls restrict China AI accelerator sales
Leading indicators
- EPYC server CPU unit and revenue share
- Instinct GPU revenue and named hyperscaler deployments
- ROCm framework compatibility and developer adoption
Counterarguments
- Intel still holds the majority of x86 server CPU units
- Nvidia has stronger AI accelerator software lock-in
Keystone Component
Supply
Keystone Component
Strength
Durability
Confidence
Evidence
AMD's data-center value proposition increasingly bundles EPYC CPUs, Instinct accelerators, Pensando networking, ROCm software, and rack-scale designs into core compute infrastructure.
Erosion risks
- Nvidia integrates GPUs, networking, CPUs, systems, and software faster
- Intel improves Xeon performance and packaging
- Merchant AI accelerators become commoditized on price
Leading indicators
- Rack-scale Instinct customer announcements
- EPYC attach rates in AI clusters
- Pensando networking adoption
Counterarguments
- AMD is still a challenger in AI accelerators
- The most differentiated system software remains controlled by Nvidia
Interoperability Hub
Network
Interoperability Hub
Strength
Durability
Confidence
Evidence
ROCm, open standards, UALink participation, and x86 compatibility position AMD as a lower-lock-in alternative for AI and HPC infrastructure, though adoption is still contested.
Erosion risks
- CUDA remains the default developer environment
- Open standards fragment across competing implementations
- Developers prioritize Nvidia first for model optimization
Leading indicators
- ROCm supported model and framework coverage
- Production workloads running on Instinct
- UALink ecosystem participation
Counterarguments
- Interoperability can be table stakes rather than a moat
- Open ecosystems often reduce rather than increase supplier pricing power
Client CPUs
Global client x86 CPUs for desktops and notebooks
FY2025 revenue share uses disclosed Client revenue of $10.640 billion divided by total revenue of $34.639 billion. Operating-profit share is allocated from combined Client and Gaming operating income by sub-revenue because AMD does not disclose separate Client operating income.
De Facto Standard
Network
De Facto Standard
Strength
Durability
Confidence
Evidence
AMD benefits from the entrenched x86 and Windows PC software base, which reduces friction versus Intel and raises barriers for Arm-based challengers.
Erosion risks
- Arm-based Windows PCs gain stronger application compatibility
- Microsoft or OEM support shifts toward non-x86 platforms
- Intel regains process and performance leadership
Leading indicators
- AMD client CPU unit and dollar share
- Windows-on-Arm unit share
- OEM notebook design wins
Counterarguments
- This standard is shared with Intel rather than controlled by AMD
- Intel still has deeper OEM channel relationships
Brand Trust
Demand
Brand Trust
Strength
Durability
Confidence
Evidence
Ryzen has gained credibility with gamers, enthusiasts, and higher-end PC buyers, allowing AMD to capture more dollar share than historical unit share would imply.
Erosion risks
- Intel restores high-end desktop and notebook competitiveness
- PC memory shortages or tariffs suppress upgrades
- Ryzen product cadence slips
Leading indicators
- Ryzen average selling price
- Desktop and mobile CPU revenue share
- Retail CPU bestseller rankings
Counterarguments
- Brand strength is not enough if Intel has better availability
- Client CPUs remain cyclical and promotion-driven
Gaming
PC discrete gaming GPUs and semi-custom game-console SoCs
FY2025 revenue share uses disclosed Gaming revenue of $3.910 billion divided by total revenue of $34.639 billion. Operating-profit share is allocated from combined Client and Gaming operating income by sub-revenue because AMD does not disclose separate Gaming operating income.
Design In Qualification
Demand
Design In Qualification
Strength
Durability
Confidence
Evidence
Semi-custom console SoCs involve deep co-design, long qualification, and multi-year platform cycles with console OEMs.
Erosion risks
- Sony or Microsoft choose alternative silicon for a future console cycle
- Console demand weakens late in the generation
- Cloud gaming and handheld devices alter form-factor demand
Leading indicators
- Next-generation console SoC awards
- Semi-custom revenue growth
- Console unit sell-through
Counterarguments
- Design wins can be lost at generation transitions
- AMD does not control console platform pricing or marketing
Keystone Component
Supply
Keystone Component
Strength
Durability
Confidence
Evidence
AMD's CPU/GPU SoCs are central components in major console platforms, but the Radeon PC GPU component of Gaming has much weaker competitive position.
Erosion risks
- Nvidia maintains near-monopoly PC AIB share
- Integrated graphics reduce low-end discrete GPU demand
- Memory prices and tariffs hurt consumer GPU demand
Leading indicators
- Radeon AIB share
- Console SoC shipment volumes
- Semi-custom gross-profit contribution
Counterarguments
- Nvidia dominates the PC gaming GPU profit pool
- Console SoC leadership is episodic and customer-dependent
Embedded
Embedded processors, FPGAs, adaptive SoCs, and SOMs for industrial, communications, automotive, aerospace, healthcare, and data-center applications
FY2025 revenue share uses Embedded revenue of $3.454 billion divided by total revenue of $34.639 billion. Operating-profit share uses Embedded operating income divided by reportable segment operating income before All Other corporate items. No numerical aggregate market share is included because AMD's Embedded segment spans multiple markets and current public share data are not sufficiently attributable.
Design In Qualification
Demand
Design In Qualification
Strength
Durability
Confidence
Evidence
Embedded processors, FPGAs, and adaptive SoCs are designed into long-life systems where qualification, reliability specs, and customer-specific tailoring create replacement friction.
Erosion risks
- Customers use ASICs or ASSPs instead of FPGAs
- Altera or Lattice wins next-generation sockets
- Industrial and communications customers reduce inventory
Leading indicators
- Embedded revenue growth and backlog quality
- New FPGA and adaptive SoC design wins
- Customer inventory commentary
Counterarguments
- AMD's Embedded revenue declined in FY2025
- The market is fragmented across many verticals and substitutes
Ecosystem Complements
Network
Ecosystem Complements
Strength
Durability
Confidence
Evidence
AMD's Xilinx-derived FPGA/adaptive SoC portfolio is supported by development boards, tools, IP, reference designs and SOMs that make customer development easier.
Erosion risks
- Competing FPGA toolchains improve developer productivity
- Open-source hardware flows reduce proprietary tool dependence
- Low-cost embedded AI modules commoditize edge compute
Leading indicators
- Adaptive SoC attach to development kits and SOMs
- Toolchain adoption and developer activity
- Versal and Kria design wins
Counterarguments
- Tool ecosystems are weaker moats than hard IP or exclusive sockets
- Many embedded customers prioritize cost and longevity over vendor ecosystem
Evidence
core strategic compute partner
Shows a major hyperscaler-scale AI customer committing to roadmap-level AMD deployment.
large hyperscale customers, OEMs and ODMs deployed our AMD Instinct MI350X Series GPUs
Supports the existence of scaled data-center GPU qualification and deployment.
designed for AI training, inference and exascale-class scientific computing
Describes Instinct as a core accelerator component for high-value data-center workloads.
GPUs, CPUs, networking and open software
Supports AMD's broader full-stack data-center component strategy.
continued growth of the AMD ROCm ecosystem
ROCm is the main software-layer evidence for AMD's interoperability-based AI thesis.
Showing 5 of 25 sources.
Risks & Indicators
Erosion risks
- Nvidia CUDA and full-stack systems remain the dominant AI accelerator standard
- Cloud customers shift workloads to internal ASICs
- Export controls restrict China AI accelerator sales
- HBM or leading-edge foundry shortages delay product ramps
- Nvidia integrates GPUs, networking, CPUs, systems, and software faster
- Intel improves Xeon performance and packaging
Leading indicators
- EPYC server CPU unit and revenue share
- Instinct GPU revenue and named hyperscaler deployments
- ROCm framework compatibility and developer adoption
- AI accelerator export-license updates
- Rack-scale Instinct customer announcements
- EPYC attach rates in AI clusters
Curation & Accuracy
This directory blends AI‑assisted discovery with human curation. Entries are reviewed, edited, and organized with the goal of expanding coverage and sharpening quality over time. Your feedback helps steer improvements (because no single human can capture everything all at once).
Details change. Pricing, features, and availability may be incomplete or out of date. Treat listings as a starting point and verify on the provider’s site before making decisions. If you spot an error or a gap, send a quick note and I’ll adjust.