Against the giants: how AMD went from a 1969 start-up to Ryzen, EPYC and adaptive computing

AMD has spent much of its existence competing against companies with greater scale, deeper manufacturing resources or stronger market positions. Its survival has depended on repeatedly finding a technical opening.

Advanced Micro Devices was founded in Silicon Valley in 1969 by Jerry Sanders and a group of semiconductor veterans. The early company supplied logic and memory products before becoming deeply associated with processors compatible with the x86 PC ecosystem.

More than five decades later, AMD spans CPUs, GPUs, data-centre accelerators, gaming silicon, embedded products and adaptive computing.

Jerry Sanders built AMD around second-source semiconductors

The semiconductor industry of the late 1960s and 1970s often required customers to have more than one source for important components.

AMD built business by producing logic and memory devices and later entered processor markets through licensing and compatibility arrangements.

That path eventually made competition with Intel central to AMD’s identity.

x86 compatibility gave AMD access to the PC revolution

As IBM-compatible PCs spread, x86 became one of computing’s most important instruction-set architectures.

AMD produced compatible processors and later developed increasingly independent x86 designs.

The company’s Athlon line became particularly important around the turn of the century, proving AMD could compete at the high-performance end rather than only through lower prices.

AMD64 extended x86 instead of replacing it

One of AMD’s most enduring contributions arrived with the x86-64 instruction-set extension, widely known through AMD64.

The architecture expanded x86 computing into 64-bit addressing while retaining compatibility with existing software.

The approach became the industry standard for mainstream desktop and server x86 processors.

Buying ATI made graphics a permanent second pillar

AMD acquired graphics specialist ATI Technologies in 2006.

The deal brought Radeon GPUs into AMD and created the possibility of combining CPU and graphics technologies more closely.

That eventually supported accelerated processing units that integrated CPU and GPU functions, while Radeon remained a discrete graphics business in its own right.

The difficult years forced AMD to change its structure

Building leading-edge semiconductor factories became increasingly expensive.

AMD separated its manufacturing operations into GlobalFoundries, allowing the company to move towards a fabless model and use external foundries for advanced chips.

The transition reduced direct ownership of fabrication plants but made execution in chip design and supplier relationships even more important.

Lisa Su inherited a company that needed a technical reset

Dr Lisa Su joined AMD in 2012 and became chief executive in October 2014.

AMD was under intense competitive and financial pressure at the time. Its next major CPU architecture therefore needed to do more than deliver an incremental improvement.

Zen turned the comeback into products people could buy

The Zen CPU architecture became the foundation for AMD’s resurgence.

Ryzen desktop processors arrived in 2017, bringing a more competitive mix of core counts, performance and efficiency to consumer PCs.

EPYC applied the architecture to servers, giving AMD a route back into data centres.

Successive Zen generations improved performance while AMD adopted chiplet designs that allowed multiple pieces of silicon to be combined inside one processor package.

Chiplets changed how AMD scaled processors

Instead of building every function on one giant die, AMD increasingly separated CPU cores and I/O into specialised chiplets.

This can improve manufacturing economics because smaller dies generally yield better than extremely large ones, while different parts of a processor can use manufacturing processes suited to their needs.

Chiplets became a defining element of Ryzen and EPYC development.

Xilinx expanded AMD beyond conventional CPUs and GPUs

AMD completed its acquisition of Xilinx in February 2022.

Xilinx brought field-programmable gate arrays, adaptive SoCs and a large embedded-computing business.

The transaction changed how AMD described itself: not only as a CPU and graphics company, but as a high-performance and adaptive computing provider spanning cloud, edge and embedded workloads.

AI brought the data-centre GPU business into focus

The growth of generative AI increased demand for data-centre accelerators.

AMD expanded its Instinct portfolio and software stack to address training and inference workloads, adding another front to its long history of competing in markets dominated by much larger incumbents.

AMD’s history is a sequence of technical second chances

AMD has been written off more than once. Its continued relevance came from architectural decisions that opened new paths: x86 compatibility, AMD64, ATI graphics, the Zen redesign, chiplets and the Xilinx acquisition.

As of 18 September 2026, Lisa Su remains chair and chief executive. AMD now competes across PCs, servers, game consoles, graphics, embedded systems and AI accelerators.

This company-history page gives future TechnologyBlog.co.za Ryzen, Radeon, EPYC, Instinct and Xilinx-derived coverage a common brand authority page to link back to.

Sources reviewed include AMD corporate history material, AMD leadership pages and official releases for the Xilinx acquisition. Facts are current to 18 September 2026.