Jump to content

AMD

From Alfawiki
Advanced Micro Devices, Inc.
Company typePublic
IndustrySemiconductors, Technology
FoundedMay 1, 1969; 57 years ago (1969-05-01) in Sunnyvale, California, United States
FounderJerry Sanders and associates
Headquarters
Santa Clara, California
,
United States
Area served
Worldwide
Key people
Lisa Su (CEO)
ProductsCPUs, GPUs, FPGAs, SoCs, high-performance computing technologies
Number of employees
Not specified
ParentNone
Websiteamd.com

Advanced Micro Devices, Inc. (AMD) is a United States-based multinational corporation and technology company with its headquarters located in Santa Clara, California and significant operations situated in Austin, Texas. Operating as a fabless enterprise, AMD specializes in the design and development of central processing units (CPUs), graphics processing units (GPUs), field-programmable gate arrays (FPGAs), system-on-chip (SoC) solutions, and high-performance computing technologies. Its product range caters to various sectors such as gaming, data centers, artificial intelligence (AI), and embedded systems.

The company’s primary portfolio encompasses microprocessors, motherboard chipsets, embedded processors, and graphics processors tailored for servers, workstations, personal computers, and embedded devices. AMD has strategically expanded into markets including data centers, gaming, and high-performance computing. Its processors are integral to numerous devices like PCs, servers, laptops, and gaming consoles. Following the divestiture of its manufacturing division as GlobalFoundries in 2009, AMD concentrated on chip design while outsourcing production. The acquisition of Xilinx in 2022 broadened AMD’s FPGA offerings.

Established in 1969 by entrepreneur Jerry Sanders alongside a group of technology specialists, AMD initially focused on manufacturing memory chips before venturing into the microprocessor arena in 1975 to challenge Intel. The early 2000s marked growth fueled by the success of Athlon and Opteron processors. Nevertheless, AMD encountered difficulties competing with Intel’s faster chips during the late 2000s and early 2010s.

In the latter part of the 2010s, AMD regained market presence through aggressive pricing strategies and the introduction of Ryzen processors that delivered competitive performance at appealing price points compared to Intel’s products. By 2022, AMD outpaced Intel in market capitalization for the first time, underscoring its strategic vision and expansion within the semiconductor sector.

Foundational years

Advanced Micro Devices was founded on May 1, 1969, by Jerry Sanders, who had previously served as marketing director at Fairchild Semiconductor, along with seven associates. Dissatisfied with the limited backing he received at Fairchild, Sanders sought to establish a flexible semiconductor firm inspired by Intel founders Robert Noyce and Gordon Moore.

In September 1969, AMD moved its operations to Sunnyvale, California. At its inception, it functioned as a secondary source supplier for microchips originally produced by Fairchild and National Semiconductor to rapidly build its customer base. The company concentrated on manufacturing logic chips adhering to U.S. Military Standard MIL-STD-883 quality controls, which attracted early computer manufacturers and telecommunications companies.

AMD’s initial product launch was the Am9300 4-bit MSI shift register in 1970, succeeded by the well-received Am2501 logic counter and Am2505 multiplier. In 1971, AMD entered the RAM segment with its Am3101 bipolar RAM chip and broadened sales of linear integrated circuits, achieving $4.6 million in annual revenue.

The firm went public in September 1972. By 1973, it had become a second-source supplier for Intel MOS/LSI circuits. By 1975, AMD’s catalog included 212 products featuring proprietary RAM and low-power Schottky MSI circuits.

AMD entered the microprocessor industry in 1975 with the release of the Am9080—a reverse-engineered clone of Intel's 8080—and introduced the Am2900 bit-slice microprocessor family. Starting in 1976, a cross-licensing agreement with Intel granted AMD access to Intel’s microcode.

In 1977, AMD partnered with Siemens through a joint venture aimed at boosting technological expertise and strengthening presence in the U.S. market; Siemens acquired a 20% stake that provided capital for product development. Together they launched Advanced Micro Computers (AMC), focusing on microcomputer development utilizing AMD’s second-source Zilog Z8000 processors. Due to differing strategic objectives, AMD repurchased Siemens’ shares by 1979 and shuttered AMC by 1981 to concentrate on second-source Intel x86 microprocessors.

By fiscal year 1978, AMD’s revenues exceeded $100 million. The company made its debut on the New York Stock Exchange in 1979 and commenced production at its semiconductor fabrication facility in Austin, Texas—complementing overseas assembly operations in Penang and Manila. Construction began on a San Antonio fabrication plant in 1981. In 1980, AMD expanded its semiconductor supply efforts into the burgeoning telecommunications industry.

Intel collaboration

Intel launched the initial x86 microprocessors in 1978. When IBM developed its PC in 1981, it selected Intel's x86 processors but required an alternate manufacturer due to patent restrictions. As a result, Intel and AMD forged a decade-long technology exchange agreement between 1981 and 1982. This pact permitted both companies to produce and market semiconductor products created by the other under licensing and royalty arrangements. Through this agreement, AMD secured second-source manufacturing rights for Intel's 8086, 80186, and 80286 processors, gaining access to essential technical documentation.

Beginning in 1982, AMD commenced mass production of Intel-licensed x86 chips and introduced its Am286 clone by 1984, aiming at the expanding IBM PC and compatible markets. Simultaneously, AMD concentrated on proprietary bipolar integrated circuits and made substantial investments in research and development, achieving the world's first 512K EPROM in 1984. That same year, AMD earned recognition as one of the "100 Best Companies to Work for in America" and entered the Fortune 500 ranking in 1985.

The mid-1980s downturn in the microchip sector, spurred by Japanese dumping practices and saturation of the U.S. market, posed challenges for AMD. In response, the company pursued vigorous innovation efforts, exemplified by the Liberty Chip initiative—designing a new chip or chipset every week during fiscal year 1986—and advocated for U.S. trade restrictions against unfair pricing tactics. During this timeframe, AMD withdrew from the DRAM segment and expanded its capabilities in CMOS technology.

AMD found success with the AMD7910/7911 "World Chip" FSK modem and adopted RISC architecture by developing the Am29000 processor series, which gained traction as embedded processors. The company also grew its presence in the EPROM market. Throughout the 1980s, AMD continued as a second-source supplier for Intel's x86 processors but began creating its own designs with the launch of the Am386 in 1991, entering direct rivalry with Intel.

Even amid the dotcom crash, AMD sustained a robust flash memory division. To bolster cash flow amid fierce competition from Intel in microprocessors, AMD spun off its flash memory manufacturing into Spansion in 2003—a joint venture initially established with Fujitsu in 1993. In December 2005, AMD divested Spansion via an initial public offering to concentrate resources on its primary microprocessor business.

2006–2019

On July 24, 2006, AMD revealed plans to acquire Canadian 3D graphics firm ATI Technologies for $5.4 billion; this acquisition concluded on October 25, 2006. By August 2010, AMD phased out the ATI brand name, integrating all graphics chipsets under the AMD identity.

In October 2008, AMD separated its manufacturing division into GlobalFoundries, a joint venture with Abu Dhabi's Advanced Technology Investment Co., thereby securing capital to emphasize chip design activities. CEO Hector Ruiz resigned in July 2008 to lead GlobalFoundries as chairman, while President and COO Dirk Meyer assumed the role of AMD CEO. The economic downturn in 2009 compelled AMD to reduce its workforce by approximately 1,100 employees.

In August 2011, former Lenovo executive Rory Read succeeded Meyer as CEO. Confronted with declining revenues, AMD implemented layoffs exceeding 10% of its staff by late 2011 and an additional 15% reduction in 2012. The integration of AMD processors within the PlayStation 4 and Xbox One gaming consoles played a critical role in preventing bankruptcy.

At the start of 2012, AMD acquired low-power server manufacturer SeaMicro to advance development of an Arm64 server processor. In October 2014, Rory Read stepped down after three years at the helm; COO Lisa Su was appointed CEO and initiated a restructuring that segmented AMD into two divisions: Computing and Graphics (covering desktop/notebook CPUs and discrete GPUs) and Enterprise, Embedded, and Semi-Custom (encompassing server CPUs and semi-custom SoCs including gaming console platforms). This reorganization included a workforce reduction of about 7% by year-end.

Following the spin-off of GlobalFoundries and subsequent layoffs, AMD vacated its longstanding Sunnyvale headquarters. In August 2016, it moved into a new facility spanning 220,000 square feet within Santa Clara Square—strategically located near competitor Intel’s headquarters. The previous Sunnyvale property was sold to the Irvine Company, which obtained approval in April 2019 to redevelop it into residential housing.

2020–present

In October 2020, AMD revealed its plan to acquire Xilinx, a prominent company specializing in field programmable gate arrays and complex programmable logic devices (FPGAs and CPLDs), through an all-stock transaction valued near $50 billion. This deal was completed in February 2022, considerably broadening AMD's product range and market presence.

AMD expanded its AI software capabilities by acquiring Nod.ai, an open-source AI software firm, in October 2023, reinforcing its foothold within the artificial intelligence sector.

By January 2024, AMD ceased manufacturing all CPLD products obtained from Xilinx, thereby optimizing its product lineup.

A notable rise in semiconductor stock prices during March 2024 elevated AMD’s market capitalization beyond $300 billion for the first time, signaling robust investor confidence and business expansion.

In July 2024, AMD announced the acquisition of Finnish AI startup Silo AI for $665 million, aiming to enhance its AI technology portfolio and better compete with industry leader Nvidia in the AI chip sector.

Further strengthening its AI hardware capabilities, AMD signed a $4.9 billion agreement in August 2024 to acquire ZT Systems, a company focused on custom computing infrastructure for AI workloads.

Since early 2024, AMD has made considerable progress in the server CPU market, closing the gap with Intel. As of July 2025, Intel maintains a 63.3% market share while AMD has grown to 36.5%, demonstrating AMD’s notable upward momentum despite Intel’s ongoing dominance.

List of CEOs

Name Years Position, education
Jerry Sanders 1969–2002 Founder, electrical engineer
Hector Ruiz 2002–2008 Electrical engineer
Dirk Meyer 2008–2011 Computer engineer
Rory Read 2011–2014 Information Systems
Lisa Su 2014–present Electrical engineer

IBM PC and the x86 architecture

In February 1982, AMD secured a license to act as a second-source manufacturer for Intel’s 8086 and 8088 processors. This arrangement allowed IBM to source these chips from multiple suppliers for its PC line. Under this agreement, AMD also produced the Am286 processor. However, when Intel stopped sharing technical documentation for the Intel 80386 in 1984, AMD initiated arbitration proceedings which it won in 1992. Intel resisted this ruling until the California Supreme Court affirmed AMD's rights in 1994.

Amid continuing legal confrontations, AMD engineered clean-room versions of Intel’s microcode for its x386 and x486 processors. The Am386 debuted in March 1991 and achieved sales of one million units by October that year. In 1993, AMD launched the Am486 series, which gained substantial OEM adoption including an exclusive contract with Compaq. The Am5x86 followed in November 1995 as a high-performance yet cost-efficient processor option.

A conclusive settlement in 1996 granted AMD access to Intel’s x386 and x486 microcode but excluded subsequent processor generations. This resolution enabled AMD to establish a stronger position within the x86-compatible processor market and enhance its competitiveness within the semiconductor industry.

K5, K6, Athlon, Duron, and Sempron

In 1996, AMD introduced its inaugural in-house x86 processor, the K5, marking a milestone in its processor design journey. The letter "K" stood for Kryptonite, representing AMD’s intent to challenge Intel’s dominance. The numeral "5" indicated the fifth generation of x86 CPUs, aligning with Intel’s Pentium naming convention.

That same year, AMD acquired NexGen, known for its Nx series of x86-compatible processors. This acquisition paved the way for the K6 processor, launched in 1997, which delivered superior performance compared to Intel’s Pentium II models such as the K6-III/450, while retaining compatibility with Socket 7 motherboards.

The seventh generation of x86 processors arrived in 1999 with the Athlon line, which introduced a new Slot A interface due to licensing restrictions on Intel’s Slot 1. AMD later expanded its offerings with the Duron, a more affordable version of Athlon featuring a smaller L2 cache. Subsequently, the Sempron series was unveiled as a cost-effective successor to Athlon XP, evolving to support newer sockets including AM3.

On October 9, 2001, AMD launched the Athlon XP, followed by an upgraded model in February 2003 that incorporated a larger 512 KB L2 cache, bolstering its competitive edge in the processor market.

Athlon 64, Opteron, and Phenom

The K8 microarchitecture represented a major leap forward by introducing 64-bit computing (x86-64/AMD64), an integrated memory controller, and the high-speed HyperTransport interconnect within AMD’s Direct Connect Architecture. Initially released in April 2003 as the server-focused Opteron processor, this design was soon adapted for desktop use under the Athlon 64 brand.

In April 2005, AMD debuted its first dual-core Opteron processors for servers, followed by dual-core Athlon 64 X2 processors for desktops in May. By May 2007, AMD streamlined its desktop dual-core branding by dropping the "64" suffix to emphasize value-oriented systems. Additionally, dual-core Sempron processors were introduced in China to complete AMD’s dual-core lineup across various market segments.

September 2007 brought forth K10-based Opteron processors available in dual-, triple-, and quad-core configurations, followed by the November launch of the Phenom desktop processor. These products were part of the Spider platform, which combined Phenom CPUs with ATI R770 GPUs and AMD’s 700 series chipsets. However, their manufacturing at 65nm lagged behind Intel’s more advanced 45nm process technology.

In January 2009, AMD refreshed its product range with the Phenom II series fabricated using a 45nm process. The Dragon platform paired Phenom II CPUs with R770 GPUs and 790 GX/FX chipsets. Phenom II featured dual-, triple-, and quad-core variants on a single die and resolved previous limitations such as low clock speeds and restricted cache sizes. It also supported DDR3 memory through the AM3 socket while maintaining backward compatibility with AM2+ motherboards and DDR2 memory.

April 2010 saw the release of the six-core Phenom II X6 "Thuban," based on the Istanbul Opteron die and equipped with "turbo core" technology that dynamically increased core frequencies under heavy workloads.

That year also introduced server processors including the Opteron 6100-series "Magny-Cours" (offering 8 to 12 cores) and the 4100-series "Lisbon" (4 to 6 cores). Magny-Cours employed a multi-chip module design with two six-core dies on socket G34 for dual- or quad-socket configurations targeting high performance. Lisbon focused on energy efficiency for single- or dual-socket systems using socket C32. Both families utilized a 45nm SOI manufacturing process.

This progression highlights AMD’s strategic emphasis on broadening its presence within server and desktop markets through innovations in multi-core designs, platform integration, and pricing strategies tailored to evolving customer demands.

Fusion evolves into the AMD APU

Following AMD's acquisition of the Canadian graphics company ATI Technologies in 2006, the company initiated the Fusion project aimed at merging CPU and GPU functionalities onto a single chip, thereby removing the need for a separate northbridge on the motherboard. This architecture delegated certain CPU operations to the GPU, which was better suited for those tasks, and integrated a native PCI Express interface. The Fusion concept was eventually renamed the AMD APU (Accelerated Processing Unit).

Llano represented AMD’s inaugural APU designed primarily for laptops and was the second APU released overall,[1] targeting mainstream consumers.[2] It unified CPU, GPU, and northbridge capabilities onto one die, utilizing "Socket FM1" and supporting DDR3 memory. The CPU core architecture was derived from the Phenom II "Deneb" processor. Manufacturing challenges with Llano unexpectedly impacted AMD’s revenue negatively.Cite error: The opening <ref> tag is malformed or has a bad name The APU product range later grew to include budget-friendly E1 and E2 models and the Vision A-series (A4 through A10) aimed at mainstream users. These processors featured next-generation Radeon graphics cores, with premium variants equipped with Radeon R4 or R6 GPUs.

Bulldozer: High-performance, high-power cores

Debuting on October 12, 2011, Bulldozer introduced a new microarchitecture for AMD’s server and desktop AMD FX processors, succeeding the family 10h (K10) architecture. Designed from scratch for thermal design power (TDP) ranges between 10 and 125 watts, Bulldozer targeted improvements in performance-per-watt for demanding computing workloads. Despite these goals, Bulldozer frequently fell short of expectations and often underperformed compared to its predecessors in benchmark tests.

In 2012, Bulldozer was succeeded by the Piledriver microarchitecture, which delivered higher clock frequencies and enhanced performance across FX processors, APUs, and Opteron products.[3] The Steamroller architecture arrived in 2013 with a focus on APUs, improving instruction-level parallelism.[4][5]

Excavator followed in 2015 as the last iteration of the Bulldozer family,[6] placing emphasis on greater energy efficiency.Cite error: The opening <ref> tag is malformed or has a bad name This microarchitecture served as a bridge before AMD transitioned to its high-performance Zen architecture.

Low-power Cat cores

The Bobcat microarchitecture, unveiled by AMD EVP Henri Richard at Computex 2007 and launched in early 2011, targeted low-power applications with cores optimized for power envelopes between 1 and 10 watts. It addressed competitive challenges within the 10–100 watt x86 market segment while also being scalable for sub-1 watt devices suitable for handheld gadgets.

Released in 2013 as Bobcat’s successor, Jaguar powers AMD’s affordable low-power APUs and significantly boosted AMD’s presence in embedded systems and gaming markets. Custom APUs based on Jaguar architecture form critical components of major gaming consoles such as the PlayStation 4, Xbox One, PlayStation 4 Pro, Xbox One S, and Xbox One X—highlighting AMD’s strategic role in global partnerships and operations.

The Puma microarchitecture, introduced in 2014 following Jaguar, further enhanced AMD’s portfolio of low-power computing solutions.

ARM architecture-based designs

In 2012, AMD revealed plans to develop ARM-based solutions aimed at the semi-custom and server sectors. Their first server offering, the 8-core Cortex-A57 64-bit Opteron A1100 SoC, was introduced in 2014 but only became available in 2016 due to delays related to software ecosystem readiness. The product's launch was hindered by limited endorsement from key vendors, which restricted its market penetration. AMD also intended to create an APU featuring a Graphics Core Next GPU and unveiled the custom K12 ARM core in 2014, targeting server, embedded, and semi-custom applications. Nonetheless, the K12 project experienced setbacks with no release date announced as AMD shifted focus toward advancing its x86-based Zen microarchitecture.

Zen-based CPUs and APUs

The Zen microarchitecture represents AMD’s x86-64 design foundation for the Ryzen CPU and APU lineup, debuting in 2017. Initiated under Jim Keller’s guidance starting in 2012, Zen achieved a substantial leap with a 52% boost in instructions per cycle (IPC), exceeding the original goal of a 40% increase. Manufactured using a 14 nm FinFET process, Zen delivered enhanced single-thread performance and improved power efficiency compared to AMD’s earlier 28 nm and 32 nm generations.

Zen established a unified platform supporting both CPUs and APUs on the Socket AM4 interface and incorporated simultaneous multithreading (SMT), bringing AMD’s technology closer to Intel’s hyper-threading capabilities. It also introduced DDR4 memory support, improving overall system performance and compatibility.

AMD launched the high-performance Ryzen 7 "Summit Ridge" CPUs based on Zen in March 2017, followed by Ryzen 5 mid-tier and Ryzen 3 entry-level processors later that year. The company expanded into the server market with Epyc processors built on Zen for single- and dual-processor configurations. In October 2017, AMD released Ryzen Mobile APUs equipped with Vega graphics cores. The Zen+ refresh using a 12 nm process arrived in April 2018, succeeded by the 7 nm Zen 2 architecture in June 2019, which also refreshed the Epyc server lineup. The Zen 3 generation was announced for release in the third quarter of 2020.

By 2019, AMD’s Ryzen CPUs reportedly surpassed Intel’s consumer desktop sales, indicating significant market traction. At CES 2020, AMD introduced Ryzen Mobile 4000 series—the first x86 mobile processors fabricated on a 7 nm node offering up to eight cores and sixteen threads designed for ultrathin laptops—still utilizing Zen 2 technology. Later that year saw the launch of Zen 3-based processors, with models like the Ryzen 5 5600X achieving leading single-thread performance scores.

In April 2020, AMD targeted commercial high-performance computing (HPC) and hyperconverged infrastructure markets by releasing new high-frequency Epyc SKUs on its second-generation Rome platform, supported by partners such as Dell EMC, Hewlett Packard Enterprise, Lenovo, Supermicro, Nutanix, and IBM Cloud as its first public cloud collaborator. August 2022 marked the unveiling of AMD’s initial family of Zen 4-based CPUs.

Prominent gaming devices including the Steam Deck, PlayStation 5, and Xbox Series X and Series S employ customized Zen 2-derived chips specifically optimized for their performance demands.

In March 2025, AMD announced Instella, an open-source large language model project highlighting the company’s expanding focus beyond traditional semiconductor manufacturing into AI-driven technologies.

Radeon under AMD management

In 2007, AMD’s ATI division unveiled the TeraScale microarchitecture, which introduced a unified shader framework that replaced traditional fixed-function units with programmable shaders. This architecture first appeared in the Xbox 360’s GPU and was subsequently used in the Radeon HD 2000 series. Between 2007 and 2015, three iterations of TeraScale were developed.

Integration of GPU and CPU units

AMD consolidated its CPU and GPU departments in 2009 to propel its APU strategy, combining graphics processing with general-purpose computing. The follow-up to TeraScale, named Graphics Core Next (GCN), debuted in 2011. GCN improved general-purpose GPU computing capabilities and enabled heterogeneous computing on APUs by adopting a reduced instruction set computing (RISC) approach, markedly enhancing computational efficiency compared to TeraScale’s VLIW architecture. The Radeon HD 7000 series introduced GCN, which has evolved through five generations from 2011 to at least 2018, underpinning AMD’s strengths in both graphics rendering and compute workloads.

Radeon Technologies Group

September 2015 saw the formation of the Radeon Technologies Group (RTG), an autonomous internal division dedicated to graphics innovation, headed by Raja Koduri. This reorganization empowered RTG with independent authority over product development and marketing efforts, bolstering AMD’s competitiveness in the graphics sector. RTG was responsible for the creation of the Polaris and Vega microarchitectures, released in 2016 and 2017 respectively. The Vega design, representing AMD’s fifth generation of GCN, brought substantial improvements in performance and computational capabilities.

Following Raja Koduri’s departure from RTG in November 2017, CEO Lisa Su took over leadership responsibilities. Early in 2018, industry experts Mike Rayfield and David Wang joined as senior vice president/general manager and senior vice president of engineering respectively, strengthening RTG’s executive team. By January 2020, AMD announced its work on the second generation of the RDNA architecture aimed at directly competing with Nvidia’s RTX lineup. The launch of the RX 6000 GPUs based on RDNA2 in October 2020 introduced native ray-tracing capabilities designed to rival Nvidia’s premium RTX 3000 series.

This concentrated focus within RTG exemplifies AMD’s strategic intent to grow its share in the graphics processing market through continuous innovation, key leadership appointments, and product development aligned with evolving industry demands.

Semi-custom Solutions and Console Partnerships

In 2012, then-CEO Rory Read initiated AMD's semi-custom design program, allowing clients to co-develop specialized chips derived from AMD’s proprietary technology. This approach involves a one-time engineering fee combined with purchase costs for the customized components. Leveraging its unique capability to offer both x86 CPU cores and graphics intellectual property, AMD secured critical design wins as the APU provider for major gaming consoles such as the PlayStation 4, Xbox One, PlayStation 4 Pro, Xbox One S, Xbox One X, Xbox Series X/S, and PlayStation 5. These semi-custom offerings became a vital revenue source, accounting for most of AMD’s earnings in 2016. In November 2017, AMD partnered with Intel to co-market a product integrating an Intel Core processor with a semi-custom AMD Radeon GPU and HBM2 memory within a single package, underscoring AMD’s strategic alliances aimed at expanding its market reach.

AMD motherboard chipsets

Prior to the 2003 debut of Athlon 64 processors, AMD engineered chipsets for its K6 and K7 series, such as the AMD-640, AMD-751, and AMD-761 models. With the introduction of Athlon 64, AMD altered its approach by embracing third-party chipset manufacturers through the "Open Platform Management Architecture" program, allowing companies like ATI, VIA, and SiS to produce chipsets compatible with Athlon 64 CPUs. Nvidia contributed the chipset for the Quad FX platform.

In 2004, AMD withdrew from server chipset development following the release of the AMD-8111 chipset for Opteron processors, ceding leadership in server chipset innovation to Nvidia and Broadcom. The acquisition of ATI Technologies in 2006 brought valuable chipset expertise, including the Radeon Xpress 200 and 3200 series, which AMD subsequently rebranded—for example, the CrossFire Xpress 3200 became known as AMD 580X CrossFire.

February 2007 marked AMD's return to branded chipsets with the launch of the AMD 690G, aimed at mainstream integrated graphics users and pioneering motherboard support for HDMI 1.2, achieving shipments exceeding one million units. Plans to develop Intel IGP chipsets were abandoned due to licensing complications.

On November 15, 2007, AMD unveiled the 7-Series chipsets, covering a range from enthusiast-level multi-GPU setups to entry-level integrated graphics solutions. This new lineup replaced previous families such as the 480/570/580 and 690. The release coincided with the "Spider" desktop platform designed for discrete graphics cards and was followed by "Cartwheel" in 2008, targeting integrated graphics processors.

AMD reentered the server chipset market with the 800S, offering features including SATA 6.0 Gbit/s ports, C6 power state support for Fusion processors, and AHCI 1.2 with FIS-based switching capabilities. This series supports AMD Phenom CPUs and enthusiast platforms like Quad FX (890FX) and integrated graphics solutions (890GX).

With the arrival of APUs in 2011, AMD consolidated northbridge functionalities directly into the APU die while pairing it with a single-chip Fusion Controller Hub (FCH) that provides traditional southbridge features.

In 2017, AMD introduced new chipsets tailored for its Ryzen processors built on the Zen microarchitecture, which incorporated much of the northbridge functions internally. These AM4 socket-based chipsets differed primarily in PCI Express lane counts, USB ports, and SATA connections and were developed in partnership with ASMedia.

Embedded CPUs

During the early 1990s, AMD launched its Élan family of embedded system-on-a-chip (SoC) solutions beginning with models SC300 and SC310. These integrated a low-voltage 32-bit Am386SX CPU alongside memory controllers, peripheral interfaces, a real-time clock, and ISA bus compatibility. The SC300 model added PC card slots and a CGA-compatible LCD controller. In 1996, the SC4xx series introduced VESA Local Bus support combined with Am486 CPUs running up to 100 MHz; devices like the SC450 were utilized in products such as the Nokia 9000 Communicator. By 1999, the SC520 featured an Am586 CPU operating at either 100 or 133 MHz with SDRAM and PCI support.

In February 2002, AMD expanded into MIPS-based embedded processors for handheld and portable media applications by acquiring Alchemy Semiconductor; however, this product line was transferred to Raza Microelectronics by June 2006.

In August 2003, AMD enhanced its embedded x86 processor offerings by acquiring the Geode business (originally Cyrix MediaGX from National Semiconductor). In 2004, it introduced low-power Geode NX processors based on the K7 Thoroughbred core architecture, ranging from fanless models at 667 MHz to fan-cooled units at up to 1.4 GHz with a thermal design power (TDP) of approximately 25 W. These processors found use in embedded systems such as casino slot machines, kiosks, ultra-mobile PCs (UMPCs) popular in Asian markets, and notably in the One Laptop per Child XO-1 device. The Geode LX processor debuted in 2005 and remained available through at least 2015.

AMD broadened its embedded portfolio by incorporating 64-bit processors starting with Opteron chips aimed at high-end telecommunications and storage markets utilizing HyperTransport and Direct Connect Architecture technologies. In 2007, AMD supplemented this lineup with Athlon, Turion, and Mobile Sempron processors optimized for embedded applications at lower power consumption levels. By 2008, a variety of embedded x86-64 options were available ranging from low-power single-core to multi-core processors designed for extended availability.

The ATI acquisition in 2006 included Imageon and Xilleon product lines; however, AMD divested its handheld division to Qualcomm in late 2008—leading to Qualcomm’s Adreno GPU series—and sold Xilleon assets to Broadcom.

April 2007 saw AMD introduce the M690T integrated graphics chipset targeting embedded designs requiring robust 3D rendering and video performance—ideal for digital signage, kiosks, and point-of-sale systems. This was followed by the M690E variant featuring dual DVI outputs specifically tailored for embedded applications.

In January 2011, AMD launched its first Embedded G-Series Accelerated Processing Unit (APU), marking its entry into embedded APUs. Subsequent revisions appeared in both 2013 and 2016.

May 2012 brought forth the Embedded R-Series APU family incorporating Bulldozer CPU cores alongside Radeon HD 7000G graphics capabilities. A system-on-chip variant released in 2015 enhanced CPU and GPU performance further while adding DDR4 memory support.

These embedded product lines illustrate AMD’s strategic expansion into diverse markets demanding scalable x86-based solutions across a wide range of power requirements and application scenarios.

Embedded Graphics

AMD designs graphics processors tailored for embedded applications, catering to industries such as healthcare, casinos, and industrial equipment. Their embedded GPU solutions combine memory and graphics processing units within a compact multi-chip module. The embedded GPU line was initially launched by ATI with the E2400 model in 2008. Since then, AMD has progressively enhanced this product family through updates released in 2009, 2011, 2015, and 2016, continuously advancing GPU technology to solidify its position in the embedded systems market.

CPU and APU Offerings

As of 2020, AMD’s lineup of CPUs and APUs encompasses:

  • Athlon – entry-level CPUs utilizing Excavator cores and APUs based on Ryzen architecture
  • A-series – consumer desktop and laptop APUs built on Excavator-class technology
  • G-series – energy-efficient embedded APUs leveraging Excavator and Jaguar designs
  • Ryzen – mainstream consumer CPUs and APUs targeting a wide range of users
  • Ryzen Threadripper – premium CPUs designed for prosumer and professional workloads
  • R-series – high-performance embedded APUs employing Excavator-class cores
  • Epyc – enterprise-grade server processors intended for data center deployment
  • Opteron – microserver APUs designed for scalable server infrastructures

This comprehensive portfolio supports AMD’s strategic objective to expand its presence across consumer electronics, embedded systems, and enterprise markets.

Graphics Products

AMD’s dedicated graphics processor lineup consists of:

  • Radeon – consumer-focused graphics cards originating from ATI’s product heritage.
  • Radeon Pro – workstation-class graphics cards succeeding the FirePro series.
  • Radeon Instinct – specialized GPUs for servers and workstations aimed at machine learning and GPGPU workloads.

RAM

In 2011, AMD introduced Radeon-branded DDR3 SDRAM modules designed to enhance memory bandwidth for its APUs, produced by partners such as Patriot Memory and VisionTek. The range expanded in 2013 with higher-frequency DDR3 gaming memory, followed by the release of DDR4 modules in 2015 despite limited compatibility with AMD CPUs at that time. By 2017, AMD noted that these memory products were primarily distributed within Eastern European markets but continued to be part of their offerings.

Solid-State Drives

AMD launched Radeon-branded solid-state drives in 2014, manufactured by OCZ, providing SATA-based storage solutions with capacities reaching up to 480 GB.

CPU Hardware

Key technologies integrated into AMD CPUs and APUs include:

Graphics Hardware

Technological innovations in AMD GPUs encompass:

  • AMD Eyefinity – enabling multi-display configurations supporting up to six monitors per graphics card.
  • AMD FreeSync – a display synchronization technology based on the VESA Adaptive Sync standard.
  • AMD TrueAudio – hardware acceleration for enhanced audio processing capabilities.
  • AMD XConnect – facilitating external GPU connectivity via Thunderbolt 3.
  • AMD CrossFire – technology allowing simultaneous operation of multiple GPUs.
  • Unified Video Decoder (UVD) – hardware acceleration focused on video decoding tasks.
  • Video Coding Engine (VCE) – dedicated hardware for accelerating video encoding processes.

Software

Over the past ten years, AMD has prioritized open software tools beyond just firmware, strengthening its business approach by encouraging developer participation and expanding its ecosystem.

Distribution

AMD Radeon Software functions as AMD’s main official software distribution channel, integrating both proprietary and free elements for Microsoft Windows and Linux operating systems.

CPU

  • AOCC is AMD’s proprietary C/C++ compiler for Linux, built on the LLVM framework.
  • AMDuProf offers CPU performance and power profiling capabilities on both Linux and Windows platforms.
  • AMD has contributed to the open-source coreboot initiative, which aims to replace closed BIOS firmware; although collaboration paused in 2013, AMD has recently suggested potential renewed support for Ryzen processors with coreboot.

GPU

AMD’s most prominent publicly available software solutions center around GPUs.

The company has made both its graphics and compute frameworks open:

  • GPUOpen represents AMD’s graphics stack, incorporating technologies such as FidelityFX Super Resolution.
  • ROCm (Radeon Open Compute platform) facilitates machine learning and high-performance computing by utilizing LLVM compiler infrastructure. Within ROCm, AMDgpu serves as AMD’s open-source Linux device driver supporting GCN and later architectures, employed by both graphics and compute components.

Other

  • AMD actively explores heterogeneous computing advancements.
  • Its software suite includes the AMD Core Math Library alongside the open-source AMD Performance Library.
  • The company contributes to various open-source projects, collaborating with organizations like Sun Microsystems to enhance OpenSolaris and Sun xVM for AMD hardware. AMD also maintains its own distribution of the Open64 compiler, regularly contributing improvements back to the community.
  • Since 2008, AMD has published low-level GPU programming specifications and works closely with the X.Org Foundation to develop drivers supporting its graphics cards.
  • AMD introduced extensions for software parallelism (xSP) aimed at accelerating multi-threaded and multi-core workloads, including Light Weight Profiling (LWP) for runtime analysis and SSE5 instruction set enhancements.
  • The SIMFIRE utility assists with interoperability testing within the Desktop and mobile Architecture for System Hardware (DASH) open framework.

Production and fabrication

Historically, AMD produced its semiconductor chips in company-owned foundries and partnered with IBM and Motorola to co-develop manufacturing technologies. Founder Jerry Sanders referred to this approach as the "Virtual Gorilla" strategy to compete against Intel's larger fabrication investments.

In 2008, rising costs associated with advancing process nodes led AMD to spin off its foundry operations into a separate entity named GlobalFoundries. The Emirate of Abu Dhabi acquired GlobalFoundries through its subsidiary Advanced Technology Investment Company (ATIC), purchasing AMD’s remaining shares in 2009.

After divesting its manufacturing facilities, AMD transitioned into a fabless semiconductor company focused on designing products manufactured by third-party foundries. Although the agreement with GlobalFoundries included wafer supply commitments, AMD also established partnerships with other fabricators such as TSMC and Samsung to diversify production sources and mitigate risks.

In 2018, following GlobalFoundries’ decision to discontinue development of its 7 nm process technology, AMD shifted CPU and GPU manufacturing primarily to TSMC. By 2019, AMD revised its wafer purchase agreement to allow selection of foundries for 7 nm and smaller nodes freely while maintaining supply commitments for 12 nm processes and above through 2021.

Business trends

The principal financial and operational indicators for AMD, measured by fiscal years ending in December, are as follows:

Revenue (US$ bn) Net profit (US$ m) Total assets (US$ bn) Employees
2017 5.3 43 3.5 8,900
2018 6.4 337 4.5 10,100
2019 6.7 341 6.0 11,400
2020 9.7 2,490 8.9 12,600
2021 16.4 3,162 12.4 15,500
2022 23.6 1,320 67.5 25,000
2023 22.6 854 67.8 26,000

Partnerships

To strengthen its market position and challenge Intel's dominance, AMD has cultivated numerous strategic alliances:

  • The joint effort with Alpha Processor Inc. resulted in the creation of HyperTransport, which has become a standard technology for AMD-compatible motherboard platforms.
  • Through collaboration with IBM, AMD gained access to advanced silicon on insulator (SOI) manufacturing techniques ranging from 90 nm down to 22 nm process nodes.
  • AMD sustains robust relationships with leading original equipment manufacturers including HP, Dell, Asus, Acer, and Microsoft to ensure widespread processor availability.
  • In 1993, AMD and Fujitsu established a 50–50 joint venture named FASL LLC; this entity went public as Spansion in 2005 before AMD exited the Flash memory business under a non-compete agreement.
  • Dell ended its exclusive reliance on Intel in 2006 by integrating AMD Opteron and Athlon CPUs into their server and desktop product lines.
  • Starting in 2011, HP incorporated AMD accelerated processing units (APUs) into both business and consumer notebook models.
  • By 2013, AMD secured a prominent role by supplying processors for all three major next-generation gaming consoles: Xbox One, PlayStation 4, and Nintendo Wii U, facilitating enhanced cross-platform development and support.
  • The company partnered with Hindustan Semiconductor Manufacturing Corporation (HSMC) to manufacture chips tailored for the Indian market.
  • As an original member of the HSA Foundation, AMD advocates for heterogeneous system architecture that integrates CPUs and GPUs to boost computing performance.
  • In 2016, AMD entered a joint venture aimed at producing x86 server processors specifically for China’s market demands.
  • Collaborations with U.S. Department of Energy laboratories and corporations such as Cray Inc. and Hewlett Packard Enterprise have positioned AMD’s CPUs and GPUs at the heart of exascale supercomputers like Frontier (2021) and El Capitan (2023).
  • AMD supplies the ninth-generation gaming consoles from Microsoft and Sony unveiled in 2020.
  • In 2021, the company formed a partnership with Meta Platforms to develop semiconductor solutions oriented toward metaverse applications.
  • A collaboration established in 2022 with Samsung led to the creation of the Exynos 2022 mobile processor featuring AMD’s RDNA 2 graphics architecture.

These partnerships underpin AMD’s business approach by expanding its global footprint, spurring innovation across product lines, and reinforcing its competitive stance in critical markets.

Litigation with Intel

AMD's competitive strategy has frequently involved significant legal confrontations with former collaborator and x86 originator Intel, influencing its market share and industry standing. Notable legal conflicts include:

  • In 1986, Intel violated an agreement permitting AMD to manufacture Intel microchips for IBM; after arbitration in 1987, a ruling in favor of AMD was issued in 1992. Intel challenged this verdict but ultimately the California Supreme Court upheld it in 1994, awarding damages due to contract breach.
  • Intel filed suit against AMD in 1990 alleging copyright infringement over the use of microcode for the 287 processor; however, a jury verdict in 1994 sided with AMD, confirming its right to utilize Intel’s microcode through the 486 generation.
  • The 1997 lawsuit initiated by Intel against AMD and Cyrix concerning unauthorized use of the MMX trademark concluded with a settlement recognizing MMX as Intel’s trademark while allowing AMD to market its K6 MMX processor under that brand.
  • In 2005, Japan’s Federal Trade Commission found Intel guilty of anti-competitive behavior. That same year, AMD prevailed in an antitrust case against Intel in Japan and filed an extensive U.S. antitrust complaint accusing Intel of employing secret rebates and discounts designed to exclude AMD processors worldwide. Major manufacturers such as Acer, Dell, Lenovo, HP, and Toshiba were subpoenaed during investigations.
  • The disputes culminated in November 2009 when Intel resolved all outstanding litigation by compensating AMD $1.25 billion along with renewing a five-year patent cross-licensing agreement.

These legal proceedings have played a pivotal role in defining AMD’s corporate evolution, enhancing its market influence, and solidifying its competitive positioning within the semiconductor sector.

Guinness World Record accomplishment

  • On August 31, 2011, AMD established a Guinness World Record for the "Highest computer processor frequency" by reaching 8.429 GHz with an 8-core FX-8150 CPU operating on a single active module and cooled using liquid helium. The former record was 8.308 GHz, held by an Intel Celeron 352.
  • On November 1, 2011, Taiwanese overclocker Andre Yang surpassed this achievement by pushing the FX-8150 to 8.461 GHz.
  • On November 19, 2012, Yang further advanced the record to 8.794 GHz utilizing an FX-8350 processor.

Acquisitions, mergers, and investments

Date Company Division or integration Transaction value
January 16, 1996 NexGen AMD K6 $857 million paid in AMD stock
February 6, 2002 Alchemy Semiconductor Embedded processor units Amount undisclosed
August 6, 2003 Coatue Non-volatile polymer memory technology Amount undisclosed
July 24, 2006 ATI Technologies Graphics and 3D software (Radeon GPUs) $5.4 billion
February 29, 2012 SeaMicro Data center platform solutions $334 million
June 29, 2016 HiAlgo Gaming optimization software (Radeon Chill, Boost, Swift) Amount undisclosed
April 10, 2017 Nitero 60 GHz wireless IP for augmented and virtual reality headsets Amount undisclosed
October 27, 2020 Xilinx Custom programmable chips (FPGA, adaptive SoCs) $49 billion
April 4, 2022 Pensando Data center and cloud infrastructure solutions (DPUs) $1.9 billion
August 29, 2023 Mipsology Artificial intelligence inference software Amount undisclosed
October 10, 2023 Nod.ai Open-source AI software platforms Amount undisclosed
July 10, 2024 Silo AI AI software development services $665 million
August 19, 2024 ZT Systems Hardware solutions for data centers $4.9 billion

Corporate responsibility

  • AMD’s sustainability report from 2022 highlights its dedication to embedding environmental stewardship across all operations, ensuring workplace safety throughout its global supply chain, and fostering community growth.
  • The company successfully reduced Scope 1 and Scope 2 greenhouse gas emissions by 19% relative to its 2020 levels, with these reductions validated by independent third-party evaluations.

Additional initiatives

  • AMD played a founding role in The Green Grid, collaborating with IBM, Sun Microsystems, and Microsoft to minimize energy consumption within data centers.

Sponsorships

AMD’s involvement in Formula 1 commenced in 2002 and currently features a partnership with the Mercedes-AMG Petronas team beginning in 2020. Previously, AMD collaborated with BMW Sauber and Scuderia Ferrari alongside prominent brands such as Intel and Vodafone. In April 2018, AMD secured a long-term agreement with Scuderia Ferrari. By February 2020, AMD expanded its presence by joining Mercedes’ sponsorship roster, boosting its visibility within high-performance technology markets.

In the esports arena, AMD has sponsored Victory Five (V5) competing in the League of Legends Pro League (LPL) since 2022 and has supported the Chinese Dota Pro Circuit through Perfect World.

During 2024, AMD acted as a Diamond sponsor for the World Artificial Intelligence Cannes Festival (WAICF) and held Platinum sponsorship status at HPE Discover 2024—an event in Las Vegas that highlights technology solutions for governmental and enterprise clients.

See also

Sources

External links

  1. ↑ Cite error: Invalid <ref> tag; no text was provided for refs named Hruska
  2. ↑ Cite error: Invalid <ref> tag; no text was provided for refs named APU
  3. ↑ Cite error: Invalid <ref> tag; no text was provided for refs named Piledriver
  4. ↑ Cite error: Invalid <ref> tag; no text was provided for refs named anandtech-kaveri-review
  5. ↑ Cite error: Invalid <ref> tag; no text was provided for refs named Su
  6. ↑ Cite error: Invalid <ref> tag; no text was provided for refs named PCWorld
Retrieved from "https://www.alfawiki.org/index.php?title=AMD&oldid=9264"