Back
SEC 10-K Annual Filing

ADVANCED MICRO DEVICES INC

AMD
Fiscal YearFY 2025
Filing DateFebruary 3, 2026
SourceSEC EDGAR
Section I

Business Overview

Overview

AMD drives innovation in high performance and AI computing to solve the world’s most important challenges. AMD technology powers billions of experiences across cloud and AI infrastructure, embedded systems, AI PCs and gaming. With a broad portfolio of AI-optimized CPUs, GPUs, networking and software, AMD delivers full-stack solutions that help customers turn data into breakthroughs, with the speed and scale needed for a new era of intelligent computing. Our high-performance product portfolios provide differentiated solutions, including semi-custom System-on-Chip (SoCs), adaptive SoCs and accelerated processing units (APUs). For data centers, we offer AI accelerators, microprocessors (CPUs), graphic processing units (GPUs), data processing units (DPUs), AI Network Interface Cards (AI NICs) and field programmable gate arrays (FPGAs). We offer ultra-low latency networking solutions. We also offer data center rack-scale platform designs that incorporate AMD data center products to meet the growing performance demands of AI supercomputers and machine learning workloads. In client computing, our CPUs, APUs and chipsets for desktops and notebooks deliver performance, efficiency, AI capabilities and modern security features for gamers, creators, consumers and enterprises. AMD was the first company to integrate a dedicated neural processing unit (NPU) on the same SoC as an x86 CPU for AI PCs. By bringing NPU-accelerated AI capabilities directly into mainstream x86 platforms, we established a differentiated technology footprint that supports a growing ecosystem of AI-enabled applications and lays the groundwork for the multi-generation Ryzen™ AI roadmap that continues to expand our leadership. Our GPUs, including discrete GPUs and semi-custom SoCs, are combined with software to power immersive gaming experiences across PCs, game consoles and cloud gaming services. Our embedded computing portfolio delivers high-performance, scalable solutions across CPUs, APUs, FPGAs, System-on-Modules (SOMs) and adaptive SoCs for markets such as automotive, industrial, healthcare, storage and networking with integrated AI processing capabilities.

We develop comprehensive software stacks that include development tools, compilers and drivers to enable our high-performance CPUs, APUs, GPUs and FPGAs. We have a comprehensive and open AI software stack to support the diverse set of AI ecosystems across frameworks, large and small language models and applications. By working closely with customers, we deliver customized solutions that meet their evolving needs, leveraging our broad IP portfolio and leadership in design, integration and advanced packaging. Our investments in technologies such as our custom-ready chiplet platform and AMD Infinity Fabric™ switch position us to maintain our leadership as a custom-design silicon provider of choice.

Our Strategy

We believe AI is shaping the next era of computing and its full potential will be realized when it becomes pervasive across cloud, edge and endpoint devices. With our compute engines, intellectual property, software enablement and deep expertise, AMD is positioned to lead in this next computing era. Our broad portfolio spans supercomputing, cloud, edge, embedded and end devices, providing the unique opportunity to make AMD the end-to-end AI leader.

In 2025, a key priority was accelerating growth in our Data Center segment. Demand for our data center AI accelerator products was strong as large hyperscale customers, OEMs and ODMs deployed our AMD Instinct™ MI350X Series GPUs. We advanced our AMD AI GPU roadmap to deliver an annual cadence of leadership for AMD Instinct solutions, beginning with the AMD Instinct MI350 Series GPUs in 2025. Beyond GPUs, we launched the 5th Gen AMD EPYC™ family of server processors in 2025, which deliver leadership performance and capabilities for a wide range of data center workloads, including AI. In addition, we previewed our “Helios” AI rack-scale platform solution that incorporates all of our data center products (CPUs, GPUs and Networking) to address the growing AI compute requirements.

We strengthened our AI leadership through a number of strategic acquisitions during the year. In March 2025, we acquired ZT Group Int’l, Inc. (ZT Systems), where we retained certain intellectual property and employees associated with the design operations (ZT Design Business). This acquisition enables us to deliver end-to-end AI solutions and accelerate the design and deployment of AMD-powered AI infrastructure at scale optimized for the cloud. In October 2025, we sold the ZT data center infrastructure manufacturing business (ZT Manufacturing Business) to Sanmina Corporation (Sanmina). We made other strategic acquisitions during the year to advance our software capabilities and enable highly optimized solutions across the stack, support co-packaged optics solutions for next-generation AI systems, and bring deep expertise in high-speed inference and reasoning-based AI technologies for large-scale deployments.

We believe that AI systems will require not only powerful chips, but also full-stack innovation across compute, networking, systems architecture and software. AMD is uniquely positioned to deliver across this stack, combining industry-leading CPUs, GPUs and adaptive SoCs with networking, software and system integration expertise. We continue to invest in software capabilities and the open ecosystem through the AMD ROCm™ platform, delivering new features for high-performance AI training and inference. In 2025, we introduced key optimizations and expanded support for frameworks and libraries, improving performance for generative AI workloads and simplifying the developer experience.

In October 2025, we entered into a product purchase agreement with OpenAI OpCo, LLC, (OpenAI) to deploy 6 gigawatts of AMD GPUs, with the deployment of the first gigawatt of capacity powered by our AMD Instinct MI450 series products. This multiyear strategic partnership with OpenAI demonstrates our continued execution of hardware, software and full-stack solutions roadmaps.

Our Business

Beginning in the first quarter of fiscal year 2025, we combined the Client and Gaming segments into one reportable segment to align with how we manage our business. All prior period segment data were retrospectively adjusted.

Our three reportable segments are:

From time to time, the Company may also sell or license portions of its IP portfolio.

In addition to these reportable segments, we have an All Other category, which is not a reportable segment.

Data Center Segment

Data Center Market

The Data Center segment primarily includes server-class CPUs, GPUs, AI accelerators, DPUs, AI NICs, FPGAs, and adaptive SoC products. We leverage our technology to address the computational, visual data processing and AI workload acceleration needs in the data center market. Modern data centers require high performance, energy efficient, scalable and adaptable compute engines to meet the demand driven by the growing amount of data that needs to be stored, accessed, analyzed and managed. Different combinations of CPUs, GPUs, DPUs, AI NICs, FPGAs and adaptive SoCs enable the optimization of performance and power for a diverse set of workloads.

Data Center Products

Server CPUs.

Our CPUs for server platforms currently include the AMD EPYC™ Series processors. EPYC CPUs, which are based on the x86 architecture, are server-specific processors designed for high-performance computing, enterprise IT, supercomputing, and large data centers. Our 5th generation AMD EPYC family of server processors delivers improved performance and efficiency for AI, cloud and enterprise workloads.

Data Center GPUs.

Our AMD Instinct™ family of GPU products, including AMD Instinct MI200, MI300, MI325 and MI350 series, are based on AMD CDNA™ architecture and designed for AI training, inference and exascale-class scientific computing. We also announced next-generation AMD Instinct MI355X GPUs for large-scale AI deployments. Our visual cloud GPU offerings include products in the AMD Radeon™ PRO V families. Our visual cloud data center GPUs include a range of solutions tailored towards workloads requiring remote visualization, such as Desktop-as-a-Service, Workstation-as-a-Service and Cloud Gaming.

FPGAs and Adaptive SoCs.

We offer a wide range of FPGAs, adaptive SoCs and acceleration cards for the data center. Devices include the Virtex™, Kintex™, Artix™, and Spartan™ FPGA products, as well as Zynq™ and Versal™ adaptive SoCs. Our Alveo™ accelerator cards provide a platform for accelerating multiple data center workloads at the edge or in the cloud.

Networking Products.

Our AMD Pensando™ DPUs and comprehensive networking software stack offload data center infrastructure services from the host CPU and are used by large Infrastructure as a Service (IaaS) cloud providers to accelerate workload performance for hosted virtualized and bare-metal offerings. We introduced our AMD Pensando “Pollara” 400 AI NICs and “Vulcano” AI NICs, which deliver high-speed connectivity across GPU clusters providing high-performance, AI-ready, flexible solutions for scale-out networking.

The AMD Solarflare™ portfolio continues to offer low latency networking solutions for capital markets, including hardware adapters and the Onload™ user space networking libraries.

Client and Gaming Segment

Client Business

Client Market

Our CPUs and APUs power PCs that are an integral part of how customers work, learn and play. The Client market encompasses the global personal computing ecosystem, including desktop and notebook PCs sold primarily through OEMs, distributors and select direct customers. Our products consist mainly of x86 CPUs and APUs marketed under the AMD Ryzen™ and AMD Ryzen™ AI brands for consumer, commercial and enthusiast segments. Our customers focus on a combination of performance, efficiency, reliability, and value that aligns with their specific use cases. We built a leadership roadmap across these areas and are continuing to further differentiate end user experiences by infusing AI across our portfolio.

Client Products

Desktop CPUs.

Our desktop CPU and APU offerings include the AMD Ryzen™ and AMD Ryzen Threadripper™ processors. The Ryzen 9000 Series processors feature “Zen 5” cores, along with X3D models featuring 2nd generation AMD 3D V-Cache™ technology for leadership gaming performance. In 2025, we launched processors for gaming and content creation with Ryzen 9 9950X3D and Ryzen 9 9900X3D. In January 2026, we announced the Ryzen 7 9850X3D, in addition to our Ryzen 7 9800X3D series. Our Ryzen Threadripper™ 9000 Series processors for professional workstations feature increased core counts, expanded memory bandwidth and updated platform capabilities. Our portfolio also includes a broad range of Ryzen 9000 and Ryzen series desktop processors featuring a wide range of capabilities and price points for the growing Socket AM5 platform. In 2025, we released additional AMD Ryzen G Series processors with integrated advanced graphics to extend the longevity of the Socket AM4 platform.

Notebook CPUs.

Our latest mobile processors are designed to deliver premium laptop experiences with local AI. In 2025, we launched AMD Ryzen AI 300 Series processors featuring a next generation NPU supporting Microsoft Copilot+ PCs, our latest “Zen 5” architecture and our AMD RDNA™ 3.5 graphics architecture. In January 2026, we launched our Ryzen AI 400 series processors, the next generation of processors for AI PCs. Building on our AI PC momentum, we expanded our portfolio by introducing Ryzen AI-powered platforms across notebooks and mini desktops and introduced a new reference platform for AI developers in January 2026, the Ryzen AI Halo, that features up to 128GB of memory for large AI model support. We also expanded the Ryzen AI Max processor lineup with the Ryzen AI Max+392 and Ryzen AI Max+388 to enable wider price points for our maximum GPU configurations. Our AMD Ryzen 8000 Series mobile processors, built on the “Zen 4” architecture feature our first generation NPU, and our Ryzen 10 and 100 series mobile processors, powered by both our “Zen 2” and “Zen 3+” core architectures, all address mainstream consumer and commercial markets. For handheld gaming systems, our Ryzen Z1 Series continues to deliver immersive experiences, and we introduced the Ryzen Z2 Series, bringing higher efficiency and improved graphics performance for portable gaming devices including the new Asus ROG Xbox Ally.

Commercial CPUs.

AMD Ryzen PRO solutions deliver enterprise-class security, manageability, and reliability for commercial client PCs including business notebooks, desktops, and workstations. In 2025, we reinforced our leadership in commercial computing with new Ryzen PRO offerings that integrate AI capabilities, delivering performance, mobility, and enterprise features for our customers. Our AMD Ryzen AI PRO 300 Series, brings world-class security and manageability, with leadership productivity, battery life and AI capabilities to business notebooks and mobile workstations. With the announcement of our Ryzen AI 400 PRO Series, we continued to expand our next gen AI PC leadership and offerings. In 2025, we also launched the AMD Ryzen PRO 200 Series mobile processors that expand our commercial CPU portfolio enabling high performance PCs spanning a range of price points. For desktops, the AMD Ryzen PRO 8000G Series desktops offer an integrated AI solution for enterprises, and our AMD Ryzen Threadripper PRO CPU enables leadership performance, expanded memory bandwidth and platform reliability for premium workstations.

Chipsets.

We offer a full suite of chipset products to support our AMD Ryzen and AMD Ryzen Threadripper platforms, for entry level through professional workstation desktop systems. These chipsets complement our latest CPU innovations, supporting robust connectivity and scalability for diverse client computing needs.

Gaming Business

Graphics Market

Graphics processing is a fundamental component across many of our products. Our customers generally use our graphics solutions to enable immersive visualization and to process AI/ML based workloads. We develop our graphics products for use in various computing devices and entertainment platforms, including data centers, desktop PCs, notebook PCs, handheld devices, All-in-Ones and professional workstations. In addition, we leverage our core IP, including graphics and processing technologies, to develop semi-custom solutions deeply co-engineered with our customers, which has enabled many of today’s leading gaming consoles and handheld gaming products.

Semi-Custom Products.

Our semi-custom products are tailored, high-performance, customer-specific solutions based on CPU, GPU and multi-media technologies. We work closely with our customers to define solutions precisely matching their system requirements. AMD semi-custom SoC products power the Sony PlayStation® 5, the Microsoft® Xbox Series S™ and X™ game consoles, as well as the recently revealed Valve Steam Machine PC.

Discrete Desktop and Notebook GPUs.

Our AMD Radeon RX discrete GPU processors for PCs power the latest gaming and creation platforms. In 2025, we released our Radeon RX 9000 Series graphics cards, based on a new AMD RDNA™ 4 architecture, designed for increased AI performance, new levels of ray tracing capabilities and enhanced performance per watt to deliver new levels of performance for gamers and creators. All of our graphics products are supported by the AMD software: Adrenalin Edition™ application, which is regularly enhanced to provide the latest in performance, features, and stability. We also released FSR “Redstone” gaming technologies to integrate Machine Learning for improved performance and visuals on our Radeon RX 9000 graphics cards.

Professional GPUs.

Our AMD Radeon PRO family of professional graphics products are designed for integration in desktop workstations, optimized through hardware and software for demanding use cases such as 3D rendering, design and manufacturing for Computer-Aided Design (CAD), and media and entertainment for broadcast and animation pipelines on high resolution displays. In 2025, we launched our Radeon AI PRO R9000 Series of graphics cards, intended for local AI inference and creative workloads enabling professionals to accelerate AI-driven content creation and advanced visualization tasks directly on their workstation.

Embedded Segment

The Embedded Market

The Embedded segment primarily includes embedded CPUs, APUs, FPGAs, and adaptive SoC products. Embedded products address computing needs in aerospace and defense, automotive, industrial, vision and healthcare, communications infrastructure, test and measurement, emulation and prototyping, audio, video and broadcasting, and data center. Typically, our embedded products are used in applications that require varying levels of performance, where key features may include relatively low power, small form factors, and 24x7 operations. High-performance graphics are important in some embedded systems. Support for Linux®, Windows® and other operating systems as well as for increasingly sophisticated applications are also critical for some customers. Other requirements may include meeting rigid specifications for industrial temperatures, shock, vibration and reliability. The embedded market has moved from developing proprietary, custom designs to leveraging industry-standard instruction set architectures and processors to help reduce cost and speed time to market.

Embedded Products

Embedded CPUs and APUs.

Our embedded portfolio expanded in 2025 with the introduction of three new AMD EPYC embedded processor series: AMD EPYC Embedded 9005 Series, EPYC Embedded 4005 Series processors and EPYC Embedded 2005 Series. These new series deliver enhanced performance and extended lifecycles for networking, storage, and edge server applications. We also launched the AMD Ryzen Embedded 9000 Series processors, offering performance-per-watt and low latency for automation systems and machine vision applications. In January 2026, we introduced a new family of Ryzen AI Embedded series processors with two new series: the Ryzen AI Embedded P100 and Ryzen AI Embedded X100 Series. These additions complement our existing AMD EPYC Embedded CPUs and AMD Ryzen Embedded Series processors including V-Series APUs and CPUs and R-Series APUs and CPUs.

Our embedded solutions are designed to support high performance compute, high-bandwidth network connectivity and security, high-performance storage requirements for enterprise and cloud infrastructure, 3D graphics performance and 4K multimedia requirements of automotive infotainment systems.

FPGAs and Adaptive SoCs.

Our FPGA products are hardware-customizable devices that can be tailored to meet the specific needs of each customer, enabling them to differentiate their products and accelerate time to market. Our FPGA families include AMD UltraScale+™, UltraScale™ 7 Series, and other earlier series. Adaptive SoC products include the AMD Zynq™ SoC and Zynq UltraScale+ Multi-Processing System-on-a-Chip (MPSoC), which combine FPGA technology with a heterogeneous processing system, as well as the industry’s first RFSoC architecture with integrated radio frequency (RF) data converters (Zynq UltraScale+ RFSoC and Zynq RFSoC DFE). The AMD Versal™ portfolio, composed of software-programmable adaptive SoCs, is a heterogeneous compute platform that combines a processing system, programmable logic, AI Engines, and digital signal processing (DSP) Engines. The Versal devices achieve dramatic system-level performance improvements over today's fastest FPGA competitors’ solutions and accelerate applications in a wide variety of markets, including aerospace and defense, automotive, industrial, vision and healthcare, communications infrastructure, test and measurement, emulation and prototyping, audio, video and broadcasting, and data center. Our Versal RF Series delivers compute performance for aerospace and defense applications. We introduced our Versal Premium Series Gen 2, the first adaptive compute devices supporting next-generation CXL® and PCIe® connectivity.

Development Boards, Kits and Configuration Products.

We offer development kits for all our FPGA and adaptive SoCs. These kits include hardware, development tools, IP, and reference designs that streamline and accelerate the development of domain-specific and market-specific applications. We continue to enhance developer productivity with expanded support for AMD Kria™ SOMs and starter kits, enabling rapid prototyping and deployment for motor control and edge AI applications. Our Kria K24 SOM provides high determinism and low latency for powering electric drives and motor controllers used in compute-intensive DSP applications at the edge. Coupled with our Kria KD240 Drives Starter Kit, an out-of-the-box-ready motor control-based development platform, the products offer a seamless path to production deployment.

Design Platforms and Services

Adaptable Platforms.

We offer two types of platforms that support our customers' designs and reduce their development efforts: FPGAs and adaptive SoCs. FPGAs feature reconfigurable hardware as well as integrated memory, DSP, analog mixed signal, high-speed serial transceivers and networking cores coupled with advanced software for a broad range of applications in all of our end markets. Our adaptive SoCs feature a heterogeneous processing system with integrated programmable hardware fabric targeting embedded systems that need real-time control, analytics, sensor fusion and adaptable hardware for differentiation and acceleration. Our Zynq UltraScale+ RFSoCs feature integrated high-performance RF data converters targeting wireless, radar and cable access applications. Enabled by both hardware and software design tools and an extensive operating system, middleware, software stack and IP ecosystem, adaptive SoCs target software developers as well as traditional hardware designers. Our Versal portfolio combines a processing system, programmable logic, AI Engines, and DSP Engines with leading-edge memory and interfacing technologies to deliver powerful heterogeneous acceleration for any application. This product portfolio is ideally suited to accelerate a broad set of applications in the emerging era of big data and AI.

Software Development Platform.

Our AMD Vivado™ Design Suite provides hardware design teams with the tools and methodology needed to program FPGAs and adaptive SoCs. Our Vitis™ unified software platform enables the development and deployment of embedded software and accelerated applications on our FPGAs and adaptive SoCs. Our Vitis AI unified software platform enables the development and deployment of AI software on our FPGAs and adaptive SoCs.

Sales and Marketing

We sell our products via a direct sales force, our OEMs (as described in detail below) and through independent distributors and sales representatives in both domestic and international markets. Our sales arrangements generally operate on the basis of product forecasts provided by the particular customer, but do not typically include any commitment or requirement for minimum product purchases. We use purchase orders and related documents, and contractual agreements as evidence of our sales arrangements. Our agreements typically contain standard terms and conditions covering matters such as payment terms, warranties and indemnities for issues specific to our products.

We generally warrant that our products sold to our customers will conform to our approved specifications and be free from defects in material and workmanship under normal use and conditions for one year. We offer up to three-year limited warranties for certain product types and sometimes provide other warranty periods based on negotiated terms with certain customers.

We market and sell our latest products under the AMD trademark.

Our product brand for server microprocessors is AMD EPYC™ processors.

Our product brands for data center graphics are AMD Instinct™ accelerators and AMD Radeon™ PRO V Series.

Our product brands for data center networking are AMD Pensando™ Pollara 400 AI NIC and AMD Pensando Salina DPU.

Our product brand for low-latency adapters for FinTech is AMD Solarflare™ adapters.

Our client product brands for desktop and notebook PCs include: AMD Ryzen™, AMD Ryzen™ AI, AMD Ryzen™ PRO, AMD Ryzen™ Threadripper™, AMD Ryzen™ Threadripper™ PRO, AMD Athlon™, and AMD PRO A-Series processors.

Our product brand for the consumer graphics market is AMD Radeon™ graphics, and AMD Embedded Radeon graphics is our product brand for the embedded graphics market.

Our product brand for professional graphics products is AMD Radeon™ PRO graphics.

We also market and sell our chipsets under AMD trademarks.

Our FPGA products are Virtex™ 6, Virtex 7, Virtex UltraScale+, Kintex™ 7, Kintex UltraScale™, Kintex UltraScale+, Artix™ 7, Artix UltraScale+, Spartan™ 6, and Spartan 7 brands.

Our product brands for adaptive SoCs are Zynq™ 7000, Zynq UltraScale+™ MPSoC, Zynq UltraScale+ RFSoCs, Versal™ HBM, Versal Premium, Versal Prime, Versal™ AI Core and Versal AI Edge.

The software tools for our programmable devices are Vitis™, Vitis AI™, and Vivado™ tools.

Our System-on-Module (SOM) product is the Kria™ brand.

We also sell low-power versions of our AMD Athlon™, as well as AMD Geode™, AMD Ryzen™, AMD EPYC™, AMD R Series and G Series processors as embedded processor solutions.

Our compute and network acceleration board products are sold under the Alveo™ and Pensando™ brands.

We market our products through direct marketing and co-marketing programs. In addition, we have cooperative advertising and marketing programs with customers and third parties, including market development programs, pursuant to which we may provide product information, training, marketing materials and funds. Under our co-marketing development programs, eligible customers can use market development funds as reimbursement for advertisements and marketing programs related to our products and third-party systems integrating our products, subject to meeting defined criteria.

Customers

Our Data Center Segment customers consist primarily of hyperscale data centers, original equipment manufacturers (OEMs), original design manufacturers (ODMs), system integrators and independent distributors in both domestic and international markets. ODMs provide design and/or manufacturing services to branded and unbranded private label resellers, OEMs and system builders.

Our Client and Gaming Segment customers consist primarily of PC OEMs, a network of independent distributors and, for chipset products, ODMs that manufacture motherboards. PC OEMs and independent distributors as well as add-in-board manufacturers (AIBs), independent game console and portable gaming devices manufacturers and contract manufacturers for AMD branded graphics cards.

Our Embedded Segment products are sold to customers in a very wide range of markets such as aerospace and defense, automotive, industrial, vision and healthcare, communications infrastructure, test and measurement, emulation and prototyping, audio, video and broadcasting and data center. For these products, we either sell directly to our customers or through a network of distributors and OEM partners. We are also developing a network of Value Added Resellers (VARs) and Integrated Solution Vendors (ISVs) for our Alveo products.

We work closely with our customers to define product features, performance and timing of new products so that the products we are developing meet our customers’ needs. We also employ application engineers to assist our customers in designing, testing and qualifying system designs that incorporate our products. We believe that our commitment to customer service and design support improves our customers’ time to market and fosters relationships that encourage customers to use the next generation of our products.

We also work with our customers to create differentiated products that leverage our CPU, GPU, APU, DPU, AI NIC, and FPGA technology. Certain customers pay us non-recurring engineering fees for design and development services and a purchase price for the resulting products.

Original Equipment Manufacturers

We focus on three types of OEM partners: multi-nationals, selected regional accounts and selected global and local system integrators, who target commercial and consumer end customers of all sizes. Large multi-nationals and regional accounts are the core of our OEM partners’ business; however, we are increasingly focused on the VAR channel, which resells OEM systems to the mid-market and the small and medium business (SMB) segments. Additionally, we have increased our focus on global system integrators, which resell OEM systems, coupled with their software and services solutions into Enterprise, High Performance Computing (HPC) and Cloud Service Provider customers. Our OEM customers include numerous foreign and domestic manufacturers of servers and workstations, desktops, notebooks, PC motherboards and game consoles.

Hyperscale Data Centers

Large multi-national public cloud service providers and hyperscale private data centers directly and indirectly purchase a substantial portion of our data center-focused products, including server CPUs, GPU accelerators, DPUs, AI NICs, FPGAs and adaptive SoCs. These products are incorporated into servers and other data center appliances sold by OEMs to the hyperscale customers or into custom servers or hardware designed by or for these customers and manufactured by ODMs or contract manufacturers. Hyperscale data centers use these products to operate web-based applications or to support public cloud computing and storage service offerings, including but not limited to AI workloads such as generative AI models.

Third-Party Distributors

Our authorized channel distributors resell to sub-distributors, OEMs, ODMs and other customers. Typically, distributors handle a wide variety of products, and may include products from other manufacturers that compete with our products. Distributors typically maintain an inventory of our products. In most instances, our agreements with distributors protect their inventory of our products against price reductions and provide certain return rights with respect to any product that we have removed from our price book or are otherwise subject to discontinuation. In addition, some agreements with our distributors may contain standard stock rotation provisions permitting limited product returns.

Add-in-Board (AIB) Manufacturers and System Integrators

We offer component-level graphics and chipset products to AIB manufacturers who in turn build and sell board-level products using our technology to system integrators (SIs), retailers and sub distributors. In most instances, our agreements with AIBs protect their inventory of our products against price reductions. We also sell directly to some SI customers. SIs typically sell from positions of regional or product-based strength in the market. They usually operate on short design cycles and can respond quickly with new technologies. SIs often use discrete graphics solutions as a means to differentiate their products and add value to their customers.

Competition

The markets in which our products are sold are highly competitive and rapidly evolving and delivering the latest and best products to market on time is critical to revenue growth. The competitiveness of our products depends on a number of factors including, performance, total cost of ownership, timely product introductions, product quality and reliability, product features and capabilities, energy efficiency (including power consumption and battery life, given their impact on total cost of ownership), size (or form factor), selling price, cost, adherence to industry standards (and the creation of open industry standards), level of integration, software and hardware compatibility, ease of use and functionality of software design tools, completeness of applicable software solutions, security and stability, brand recognition and availability. Competition is expected to remain intense, driven by rapid technological change, evolving standards, shifting customer preferences, product obsolescence, and frequent product launches from both established and new competitors.

Competition in Data Center Segment

In the Data Center segment, we compete primarily against Intel Corporation (Intel) and Nvidia Corporation (Nvidia) with our CPU, GPU DPU and AI NIC server products. In addition, we compete against Altera with our FPGA and adaptive SoC server products. A variety of smaller fabless silicon companies offer proprietary accelerator solutions and Arm®-based CPUs targeting data center use cases. In addition, some of our customers are internally developing their own data center microprocessor products and accelerator products which could impact the available market for our products.

Competition in Client and Gaming Segment

Our primary competitor in the supply of CPUs and APUs is Intel. A variety of companies provide or have developed Arm-based microprocessors and platforms which could lead to further adoption of Arm-based PC solutions.

Our principal competitor in the supply of discrete graphics is Nvidia, who is the discrete GPU market share leader, and Intel, who manufactures and sells integrated graphics processors and gaming-focused discrete GPUs. With respect to integrated graphics, higher unit shipments of our APUs and Intel’s integrated graphics may drive computer manufacturers to reduce the number of systems they build paired with discrete graphics components, particularly for notebooks, because they may offer satisfactory graphics performance for most mainstream PC users at a lower cost. We are the market share leader in semi-custom game console products, where graphics performance is critical.

Competition in Embedded Segment

We expect continued competition from our primary FPGA competitors such as Altera, Lattice Semiconductor Corporation and Microsemi Corporation, and from ASSP vendors such as Broadcom Corporation, Marvell Technology Group, Ltd., Analog Devices, Texas Instruments Incorporated, NXP Semiconductors N.V., Qualcomm Incorporated and NVIDIA. In addition, we expect continued competition from the ASIC market, which has been ongoing since the inception of FPGAs. Intel is our main competitor for embedded CPUs. Other competitors include manufacturers of high-density programmable logic products characterized by FPGA-type architectures; high-volume and low-cost FPGAs as programmable replacements for ASICs and ASSPs; ASICs and ASSPs with incremental amounts of embedded programmable logic; high-speed, low-density CPLDs; high-performance DSP devices; products with embedded processors; products with embedded multi-gigabit transceivers; discrete general-purpose GPUs targeting data center and automotive applications; and other new or emerging programmable logic products.

Research and Development

We focus our research and development (R&D) activities on designing and developing products. Our main area of focus is on delivering the next generation of processors (CPU and GPU), FPGAs and adaptive SoCs, accelerators (adaptive, graphics and DPU), SOMs and AI NICs, and the associated software for these products. We focus on designing, with security in mind, new integrated circuits (ICs) with improved performance and performance-per-watt in advanced semiconductor manufacturing processes, the design of logic and interface IP, advanced packaging technologies, and heterogeneous integration technologies. We also focus on software as part of the development of our products, including design automation tools for hardware, embedded software, AI stack, and optimized software tools and libraries that extend the reach of our platforms to software and AI developers. Through our R&D efforts, we were able to introduce a number of new products and enhance our IP core offerings and software.

We also work with industry leaders on process technology, design tools, intellectual property, software and other industry consortia to conduct early-stage R&D. We are also actively contributing to numerous industry open-source software initiatives and industry consortia across a broad range of technologies. We conduct product and system R&D activities for our products in the United States with additional design and development engineering teams located in various countries who undertake specific activities at the direction of our U.S. headquarters.

Manufacturing Arrangements and Assembly and Test Facilities

Third-Party Wafer Foundry Facilities

We utilize Taiwan Semiconductor Manufacturing Company Limited (TSMC) for the production of wafers for our HPC, FPGA and adaptive SoC products and GLOBALFOUNDRIES Inc. (GF), with respect to wafer purchases for our HPC products at the 12 nm and 14 nm technology nodes. Additionally, we utilize TSMC, United Microelectronics Corporation (UMC) and Samsung Electronics Co., Ltd. for the production of our ICs in the form of programmable logic devices.

Other Third-Party Manufacturers

We outsource board-level graphics product manufacturing to third-party manufacturers.

Assembly, Test, Mark and Packaging Facilities

Wafers for our products are either sorted by the foundry or delivered by the foundry to our assembly, test, mark and packaging (ATMP) partners or subcontractors located in the Asia-Pacific region, who package and test our final semiconductor products. We are party to two ATMP joint ventures (collectively, the ATMP JVs) with Tongfu Microelectronics Co., Ltd. The ATMP JVs, Siliconware Precision Industries Ltd. (SPIL) and King Yuan Electronics Company (KYEC) provide ATMP services for our products.

Intellectual Property and Licensing

We rely on contracts and intellectual property rights to protect our products and technologies from unauthorized third-party copying and use. Intellectual property rights include copyrights, patents, patent applications, trademarks, trade secrets and mask work rights. As of December 27, 2025, we had approximately 7,200 patents in the United States and approximately 2,200 patent applications pending in the United States. In certain cases, we have filed corresponding applications in foreign jurisdictions. Including United States and foreign matters, we have approximately 18,900 patent matters worldwide consisting of approximately 12,600 issued patents and 6,300 patent applications pending. We expect to file future patent applications in both the United States and abroad on significant inventions, as we deem appropriate. We do not believe that any individual patent, or the expiration of any patent, is or would be material to our business. As is typical in the semiconductor industry, we have numerous cross-licensing and technology exchange agreements with other companies under which we both transfer and receive intellectual property and technology rights. We have acquired various licenses from external parties to certain technologies that are implemented in our products, including our IP cores and devices. These licenses support our continuing ability to make and sell our products. We have also acquired licenses to certain proprietary software, open-source software, and related technologies, such as compilers, for our design tools. Continued use of such software and technology is important to the operation of the design tools upon which our customers depend.

Backlog

Sales are made primarily pursuant to purchase orders for current delivery or agreements covering purchases over a period of time. Although such orders or agreements may provide visibility into future quarters, they may not necessarily be indicative of actual sales for any succeeding period as some of these orders or agreements may be revised or canceled without penalty.

Additional Information

AMD was incorporated under the laws of Delaware on May 1, 1969 and became a publicly held company in 1972. Our common stock is currently listed on The NASDAQ Global Select Market (NASDAQ) under the symbol “AMD”. Our mailing address and executive offices are located at 2485 Augustine Drive, Santa Clara, California 95054, and our telephone number is (408) 749-4000. For financial information about geographic areas and for segment information with respect to revenues and operating results, refer to the information set forth in Note 4 of our consolidated financial statements. We use a 52- or 53-week fiscal year ending on the last Saturday in December. References in this report to 2025, 2024 and 2023 refer to the fiscal year unless explicitly stated otherwise.

AMD, the AMD Arrow logo, 3D V-Cache, Alveo, Athlon, Atrix, CDNA, CoolRunner, EPYC, FidelityFX, FirePro, FreeSync, Geode, Infinity Fabric, AMD Instinct, Kintex, Kria, Opteron, Pensando, Radeon, RDNA, ROCm, Ryzen, Spartan, Threadripper, UltraScale, UltraScale+, Versal, Virtex, Vitis, Vivado, XDNA, Xilinx, Zynq and combinations thereof are trademarks of Advanced Micro Devices, Inc.

Microsoft, Windows, and Xbox One are registered trademarks of Microsoft Corporation in the United States and/or other countries. Linux is the registered trademark of Linus Torvalds in the United States and/or other countries. PlayStation is a registered trademark of Sony Interactive Entertainment, Inc. Arm is a registered trademark of Arm Limited (or its subsidiaries) in the United States and/or other countries.

Other names are for informational purposes only and are used to identify companies and products and may be trademarks of their respective owners.

Website Access to Our SEC Filings and Corporate Governance Documents

On the Investor Relations pages of our website, http://ir.amd.com, we post links to our filings with the SEC, our Principles of Corporate Governance, our Code of Ethics for our executive officers, all other senior finance executives and certain representatives from legal and internal audit, including our Chief Executive Officer, Chief Financial Officer, Chief Accounting Officer and persons performing similar functions, our Global Code of Conduct, which applies to our Board of Directors and all of our employees, and the charters of the committees of our Board of Directors.

Our filings with the SEC are posted on our website as soon as reasonably practical after they are electronically filed with, or furnished to, the SEC. The SEC website, www.sec.gov, contains reports, proxy and information statements, and other information regarding issuers that file electronically with the SEC. You can also obtain copies of these documents by writing to us at: Corporate Secretary, AMD, 2485 Augustine Drive, Santa Clara, California 95054, or emailing us at: Corporate.Secretary@amd.com. All of these documents and filings are available free of charge.

If we make substantive amendments to our Code of Ethics or grant any waiver, including any implicit waiver, to our principal executive officer, principal financial officer, principal accounting officer, controller or persons performing similar functions, we intend to disclose the nature of such amendment or waiver on our website.

The information contained on our website is not incorporated by reference in, or considered to be a part of, this report.

Section II

Risk Factors

ITEM 1A. RISK FACTORS

The risks and uncertainties described below are not the only ones we face. If any of the following risks actually occurs, our business, financial condition or results of operations could be materially adversely affected. In addition, you should consider the interrelationship and compounding effects of two or more risks occurring simultaneously.

Risk Factors Summary

The following is a summary of the principal risks that could adversely affect our business, financial condition and results of operations.

Economic and Strategic Risks

Operational and Technology Risks

Section III

Management Discussion & Analysis

ITEM 7. MANAGEMENT’S DISCUSSION AND ANALYSIS OF FINANCIAL CONDITION AND RESULTS OF OPERATIONS

The following discussion should be read in conjunction with the Consolidated Financial Statements as of December 27, 2025 and December 28, 2024 and for each of the three years in the period ended December 27, 2025 and related notes, which are included in this Annual Report on Form 10-K as well as with the other sections of this Annual Report on Form 10-K, “Part II, Item 8: Financial Statements and Supplementary Data.”

Introduction

In this section, we will describe the general financial condition and the results of operations of Advanced Micro Devices, Inc. and its wholly-owned subsidiaries (collectively, “us,” “our” or “AMD”), including a discussion of our results of operations for 2025 compared to 2024, an analysis of changes in our financial condition and a discussion of our off-balance sheet arrangements. Discussions of 2023 items and year-to-year comparisons between 2024 and 2023 that are not included in this Form 10-K can be found in “Management’s Discussion and Analysis of Financial Condition and Results of Operations” in Part II, Item 7 of our Annual Report on Form 10-K for the fiscal year ended December 28, 2024.

Beginning in the first quarter of fiscal year 2025, we combined the Client and Gaming segments into one reportable segment to align with how we manage our business. All prior period segment data were retrospectively adjusted.

Overview

In 2025, we delivered strong annual revenue growth with net revenue increasing 34% to $34.6 billion, compared to $25.8 billion in 2024. This growth was driven by the performance of our Data Center and Client and Gaming segments. Data Center net revenue of $16.6 billion increased by 32% compared to $12.6 billion in 2024, primarily driven by strong demand for our 5th generation AMD EPYC™ processors and AMD Instinct™ MI350 Series GPUs. Client and Gaming segment net revenue of $14.6 billion in 2025 increased by 51% compared to $9.6 billion in 2024, primarily driven by strong demand for our AMD Ryzen™ processors, semi-custom game consoles SoCs and Radeon™ gaming GPUs. The increase in annual net revenue was partially offset by a decrease in net revenue in our Embedded segment. Embedded net revenue of $3.5 billion decreased by 3% compared to net revenue of $3.6 billion in 2024, as certain end market demand remained mixed.

Gross margin of 50% increased by 1% compared to 49% in 2024, primarily due to product mix partially offset by approximately $440 million of net inventory and related charges associated with the U.S. government export control on AMD Instinct™ MI308 Data Center GPU products.

Cash, cash equivalents and short-term investments as of December 27, 2025 were $10.6 billion, compared to $5.1 billion at the end of 2024. Our aggregate principal amount of total debt as of December 27, 2025 was $3.3 billion, compared to $1.8 billion as of December 28, 2024.

In 2025, we returned a total of $1.3 billion to shareholders through the repurchase of 12.4 million shares of common stock under our stock repurchase program. As of December 27, 2025, $9.4 billion remained available for future stock repurchases under this program. The stock repurchase program does not obligate us to acquire any common stock, has no termination date and may be suspended or discontinued at any time.

During 2025, we launched multiple leadership products and made significant progress executing our AI strategy. A priority in 2025 was accelerating growth in the Data Center segment. Demand for our data center AI GPU products was strong as large hyperscale customers, OEMs and ODMs deployed our AMD Instinct MI350X Series GPUs. We advanced our AMD AI GPU roadmap to deliver an annual cadence of leadership for AMD Instinct solutions, beginning with the AMD Instinct MI350 Series GPUs in 2025. Beyond GPUs, we launched the 5th Gen AMD EPYC family of server processors in 2025, which deliver leadership performance and capabilities for a wide range of data center workloads, including AI. We also expanded the data center portfolio with new networking solutions, including the AMD Pensando™ “Pollara” 400 AI NICs and “Vulcano” AI NICs, which deliver high-speed connectivity across GPU clusters providing high-performance, AI-ready, flexible solutions for scale-out networking. In addition, we previewed our Helios AI rack-scale platform solution that incorporates all of our data center products (CPUs, GPUs and Networking) to address the growing AI compute requirements.

Across Client and Gaming, we continued to strengthen our leadership with expanding enterprise adoption and a growing portfolio of AMD Ryzen processors. For gamers, creators and developers, we brought to market AMD Radeon 9000 Series GPUs and Radeon AI PRO 9700 GPUs based on the AMD RDNA 4 graphics architecture to market along new high-performance Ryzen Threadripper™ 9000 Series processors. Our x86 Embedded portfolio continued to expand in 2025 with the introduction of three new AMD EPYC embedded processor series: AMD EPYC Embedded 9005 Series, EPYC Embedded 4005 Series and EPYC Embedded 2005 Series.

We also made strategic investments through acquisitions to further advance our software capabilities including compiler and AI expertise in machine learning, inference and performance optimization, and enable highly optimized solutions across the stack; to scale our ability to support and develop a variety of photonics and co-packaged optics solutions across next-gen AI systems; and to bring deep expertise in high-speed inference and reasoning-based AI technologies for large-scale deployments, reinforcing our enterprise AI software stack.

To execute our AI strategy, we brought in multiple AI teams across AMD to drive development of a comprehensive software ecosystem spanning our full product portfolio. We delivered key optimizations and expanded framework and library support in the latest version of AMD ROCm™ software, improving performance for generative AI workloads and simplifying the developer experience across training and inference.

In March 2025, we completed the acquisition of ZT Systems for $3.2 billion in cash and 8.3 million shares of our common stock. We retained select intellectual property and employees associated with the design operations (ZT Design Business), and in October 2025, we sold the ZT data center infrastructure manufacturing business (ZT Manufacturing Business) to Sanmina Corporation (Sanmina) for $2.4 billion in cash, subject to certain purchase price adjustments, and 1.2 million shares of Sanmina common stock. We are eligible to receive additional contingent cash consideration of up to $450 million from Sanmina to the extent certain conditions are met. Sanmina will also be our preferred partner for manufacturing capabilities in building complex AI solutions. Following the close of the sale of the ZT Manufacturing Business to Sanmina, we retained certain intellectual property and former employees of ZT Systems (ZT Design Business) and settled the contingent consideration liability with the former ZT shareholders and warrant holders.

In October 2025, we entered into a product purchase agreement with OpenAI OpCo, LLC, (OpenAI) to deploy 6 gigawatts of AMD GPUs, with the deployment of the first gigawatt of capacity powered by our AMD Instinct MI450 series products. Concurrent with the agreement, we issued to OpenAI a warrant to purchase up to an aggregate of 160 million shares of AMD’s common stock at an exercise price of $0.01 per share. The warrant shares will vest in tranches based on certain AMD Instinct GPU purchase milestones by OpenAI, or its affiliates, or indirectly through third parties, and achievement of specified AMD stock price targets and stock performance. Each vested tranche is further subject to the fulfillment of certain other technical and commercial conditions prior to exercise. Subject to certain conditions, the warrant is exercisable through October 5, 2030. None of the warrant shares met the vesting or exercise conditions and the warrant had no impact to our financial statements for the year ended December 27, 2025.

During the second quarter of fiscal year 2025, the Company recorded approximately $800 million of inventory and related charges on AMD Instinct MI308 Data Center GPU products due to new U.S. export restrictions on certain semiconductors to China. We applied for and were granted some licenses by the U.S. government that allow us to ship MI308 products to certain China-based customers. During the fourth quarter of fiscal year 2025, we began shipping products and reversed approximately $360 million of the inventory and related charges recorded earlier in the year. U.S. government officials have expressed an expectation that the U.S. government will receive 15% of the revenue generated from licensed MI308 sales to China; however, to date, the U.S. government has not published a regulation establishing such requirement.

We intend the discussion of our financial condition and results of operations that follows to provide information that will assist in understanding our financial statements, the changes in certain key items in those financial statements from period to period, the primary factors that resulted in those changes, and how certain accounting principles, policies and estimates affect our financial statements.

Critical Accounting Estimates

Our discussion and analysis of our financial condition and results of operations are based upon our Consolidated Financial Statements, which have been prepared in accordance with U.S. generally accepted accounting principles (U.S. GAAP).