# Xilinx

Xilinx was an American semiconductor company that invented the first commercially viable FPGA and pioneered the fabless manufacturing model. Acquired by AMD in 2022, its adaptive computing products now support AI acceleration in edge and data center applications.

Xilinx, Inc. (pronounced ZY-links) was an American technology and semiconductor company that primarily supplied programmable logic devices. The company is renowned for inventing the first commercially viable field-programmable gate array (FPGA) and for pioneering the fabless manufacturing model, where chip design is done in-house but fabrication is outsourced. Founded in 1984 in Silicon Valley, Xilinx became a leading supplier of adaptive computing solutions, with products ranging from FPGAs to system-on-chips (SoCs) and adaptive compute acceleration platforms (ACAPs). In February 2022, [AMD](https://www.wikiprompt.org/wiki/amd) completed an all-stock acquisition of Xilinx valued at approximately $60 billion, and the brand was phased out in June 2023, with product lines rebranded under AMD.

Xilinx's technology has been foundational in enabling flexible, reconfigurable hardware for a wide range of applications, from telecommunications and automotive to data centers and [artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence) (AI). The company's FPGAs allow designers to program logic blocks and interconnects after manufacturing, providing a middle ground between fixed-function ASICs and general-purpose processors. This flexibility has made Xilinx a key player in the evolution of computing, particularly in accelerating AI workloads at the edge and in the cloud.

## Company overview

Xilinx was headquartered in San Jose, California, with additional offices in Longmont (United States), Dublin, Singapore, Hyderabad (India), Beijing, Shanghai, Brisbane, Tokyo, and Yerevan. The company sold a broad range of FPGAs, complex programmable logic devices (CPLDs), design tools, intellectual property (IP) cores, and reference designs. According to Bill Carter, former CTO and fellow at Xilinx, the name "Xilinx" derives from the chemical symbol for silicon (Si), with "linx" representing programmable links that connect programmable logic blocks, and the 'X's at each end symbolizing the programmable logic blocks themselves.

At its peak, Xilinx held approximately 51% of the programmable logic market, with its strongest competitor Altera (later acquired by Intel) holding 34%. Other players included Actel (now part of Microsemi) and Lattice Semiconductor. Xilinx's products served diverse sectors, including aerospace, automotive, broadcast, consumer electronics, data centers, industrial, medical, and wired/wireless communications.

## History

### Early history

Ross Freeman, Bernard Vonderschmitt, and James V Barnett II, all former employees of Zilog, co-founded Xilinx in 1984. While at Zilog, Freeman envisioned chips that acted like a "blank tape," allowing users to program the technology themselves. This concept required many transistors, which were considered precious at the time, and many thought the idea was impractical. However, Freeman and Barnett left Zilog and teamed up with Vonderschmitt to raise $4.5 million in venture funding to design the first commercially viable FPGA. They incorporated the company in 1984 and began selling its first product by 1985.

By late 1987, Xilinx had raised more than $18 million in venture capital (equivalent to about $51 million in 2025) and was generating nearly $14 million in annual revenue. The company's early success demonstrated that FPGAs could be profitable despite the prevailing economics favoring high-volume generic circuits.

### Expansion

From 1988 to 1990, Xilinx's revenue grew from $30 million to $100 million. During this period, Monolithic Memories Inc. (MMI), which had been providing funding to Xilinx, was purchased by AMD. As a result, Xilinx dissolved its deal with MMI and went public on the Nasdaq in 1989 (though some sources say 1990). The company also moved to a 144,000-square-foot facility in San Jose to handle orders from HP, Apple, IBM, and Sun Microsystems.

Other FPGA makers emerged in the mid-1990s, but Xilinx continued to grow, reaching $550 million in revenue by 1995. The company expanded operations to India, Asia, and Europe. By fiscal year 2018, sales reached $2.53 billion. Under CEO Moshe Gavrielov (appointed in 2008), Xilinx introduced targeted design platforms that combined FPGAs with software, IP cores, boards, and kits to address specific applications, offering an alternative to costly ASICs and ASSPs. In January 2018, Victor Peng, the company's COO, replaced Gavrielov as CEO.

### Recent history and AMD acquisition

In 2011, Xilinx introduced the Virtex-7 2000T, the first product based on 2.5D stacked silicon (using silicon interposer technology) to deliver larger FPGAs than possible with monolithic silicon. The company later adapted this technology to combine FPGA fabric with transceivers on heterogeneous process technology, boosting bandwidth while reducing power. The Zynq-7000 line of 28 nm SoCs combined an ARM core with an FPGA, shifting Xilinx's position from a programmable logic supplier to delivering "all things programmable."

In April 2012, Xilinx introduced the Vivado Design Suite, a next-generation design environment for advanced electronic systems. In May 2014, it shipped the first 20nm UltraScale FPGAs. In September 2017, Xilinx partnered with [Amazon Web Services](https://www.wikiprompt.org/wiki/amazon-web-services) (AWS) to promote FPGA adoption, enabling AWS Marketplace AMIs and FPGA instances. In July 2018, Xilinx acquired DeepPhi Technology, a Chinese machine-learning startup. In October 2018, Xilinx FPGAs were used in a cloud-based video coding service with NGCodec's H.265 encoder, achieving 35-45% lower bitrate than GPU-based solutions.

In November 2018, the Zynq UltraScale+ MPSoC family achieved SIL 3 HFT1 certification under IEC 61508, enabling use in AI-based safety applications in automotive, aerospace, and industrial 4.0. In January 2019, ZF Friedrichshafen AG used Zynq in its ProAI automotive control unit for automated driving, with Xilinx handling data aggregation, pre-processing, and AI acceleration.

In October 2020, AMD announced its intention to acquire Xilinx. The all-stock transaction, valued at approximately $60 billion, was completed on February 14, 2022. Xilinx remained a wholly owned subsidiary of AMD until the brand was phased out in June 2023, with products now sold under AMD's Adaptive and Embedded Computing Group.

## Technology and products

Xilinx's core technology is the FPGA, a semiconductor device containing programmable logic blocks and interconnects that can be configured after manufacturing. This reconfigurability allows designers to implement custom digital circuits without the cost and time of ASIC development. Xilinx also developed CPLDs, which are simpler programmable devices, and later introduced SoCs that combine FPGA fabric with processor cores, such as the Zynq family.

Key product families included:
- Virtex: high-end FPGAs for performance-intensive applications.
- Kintex: mid-range FPGAs balancing performance and cost.
- Artix: low-cost, low-power FPGAs for cost-sensitive applications.
- Zynq: SoCs integrating ARM processors with FPGA fabric.
- UltraScale and UltraScale+: advanced 20nm and 16nm/7nm architectures.
- Versal: ACAPs combining FPGA fabric, CPU cores, and AI engines.

Xilinx also provided design software, including the ISE Design Suite and later Vivado, along with IP cores for common functions like memory controllers, DSP, and communication protocols.

## AI acceleration and adaptive computing

Xilinx positioned its products as "adaptive computing" solutions, particularly for AI acceleration. FPGAs offer low latency and high throughput for specific workloads, making them suitable for edge AI inference and data center acceleration. Xilinx's AI engines, introduced in the Versal ACAP, are designed for [deep learning](https://www.wikiprompt.org/wiki/deep-learning) inference and signal processing.

In data centers, Xilinx FPGAs were deployed on cloud platforms such as [AWS](https://www.wikiprompt.org/wiki/aws-trainium) (through F1 instances) and [Microsoft Azure](https://www.wikiprompt.org/wiki/azure) to accelerate [machine learning](https://www.wikiprompt.org/wiki/machine-learning), video processing, and networking. The acquisition by AMD aimed to combine AMD's CPU and GPU expertise with Xilinx's adaptive computing to compete in the growing AI infrastructure market.

## Impact and legacy

Xilinx's invention of the FPGA revolutionized digital design, enabling rapid prototyping and flexible hardware. The fabless model it pioneered became the standard for many semiconductor companies, including [NVIDIA](https://www.wikiprompt.org/wiki/nvidia) (though not listed) and [Qualcomm](https://www.wikiprompt.org/wiki/qualcomm). Xilinx's technology has been used in countless applications, from [Tesla's autopilot](https://www.wikiprompt.org/wiki/tesla-autopilot) (as of early models) to [Waymo's](https://www.wikiprompt.org/wiki/waymo) self-driving systems, and in telecom infrastructure.

The company's contributions to AI include enabling [generative AI](https://www.wikiprompt.org/wiki/generative-ai) inference at the edge and in the cloud, with products supporting [large language models](https://www.wikiprompt.org/wiki/large-language-model) and [transformer](https://www.wikiprompt.org/wiki/transformer) architectures. Xilinx's adaptive platforms are also used in [robotics](https://www.wikiprompt.org/wiki/robotics), medical imaging, and broadcast equipment.

## See also

- [AMD](https://www.wikiprompt.org/wiki/amd)
- [Intel](https://www.wikiprompt.org/wiki/intel) (Altera)
- [Artificial intelligence](https://www.wikiprompt.org/wiki/artificial-intelligence)
- Field-programmable gate array (not in list, but relevant)

## References

This article is based on information from public sources, including Wikipedia (CC BY-SA). Specific facts and figures are drawn from the provided source material.


---
Source: https://www.wikiprompt.org/wiki/xilinx
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-09-13T03:56:15.371434+00:00