Companies & labs

AMD

US semiconductor company designing CPUs, GPUs and Instinct AI accelerators

6recordsFirst record 2025Latest Sep 10, 2026

Profile

Country
United States
Headquarters
Santa Clara, California
Founded
1 May 1969
Founders
Jerry Sanders and seven colleagues
Type
Public company (NASDAQ: AMD), incorporated in Delaware

AMD (Advanced Micro Devices) is an American semiconductor company. It was founded in Silicon Valley on 1 May 1969 by Jerry Sanders, who had worked at Fairchild Semiconductor, together with seven colleagues. Its headquarters is in Santa Clara, California. It is incorporated under the law of the US state of Delaware, became a public company in 1972 and is now listed on the US NASDAQ market under the symbol 'AMD'. Lisa Su, who became chief executive (CEO) in October 2014, is chair and CEO as of October 2026.

AMD designs and sells chips such as CPUs (central processing units, which act as a computer's 'brain') and GPUs (graphics processing units, chips that do many calculations at once). For the 2025 financial year it had three business segments. 'Data Center' covers products such as EPYC server CPUs and Instinct AI accelerators (chips that speed up AI calculations); 'Client and Gaming' covers products such as Ryzen CPUs for PCs, Radeon graphics chips and custom chips for the PlayStation 5 and the Xbox Series X and S games consoles. 'Embedded' covers chips built into other machines, such as cars and industrial equipment. AMD spun off its manufacturing into GlobalFoundries in 2009 and now relies on outside foundries (companies that make chips for others), such as TSMC, to produce its chips. It bought the graphics chip company ATI in 2006 and Xilinx in 2022.

AMD's main AI products are its Instinct GPUs for data centres. The company says the MI300 and MI350 series, among others, are designed for AI training and inference (the process in which a trained AI produces answers). Starting with the MI350 series in 2025, it set out a plan to release new Instinct products every year, and in 2026 it launched the MI400 series, which includes the MI455X, and Helios, which puts 72 MI455X chips into a single rack (a large frame that holds server equipment). It also develops ROCm, software that lets AI programs run on these chips. For PCs, it makes Ryzen AI processors that include an NPU (neural processing unit, a circuit just for AI calculations).

Made by AIThis profile was written by AI (Claude) from the 8 sources above (official pages, filings, news reports).How it's made

Records 6

Newest first. Open one to see it with its source.

vLLM, AMD, Embedded LLM
vLLM, AMD and Embedded LLM optimize MiniMax M3 on AMD Instinct MI355X

A record, in concrete numbers, of AMD GPUs being optimized in a major open inference engine into a real alternative to NVIDIA.

AMD
AMD launches Helios rack-scale system and Instinct MI455X

72 GPUs per rack, a single scale-up domain and 31TB of HBM4 — alternative silicon entering real-world deployment

AMD
AMD announces MLPerf Training v6.0 results

Material showing on a standard benchmark that ROCm narrowed the gap to single-digit percentages in 1 year. A real crack in CUDA's monopoly

OpenAI, AMD
6 GW supply deal gives OpenAI warrants for up to 160 million AMD shares

Institutionalizes a structure in which the customer receives equity in the supplier

Nvidia, AMD, US government
Nvidia and AMD reportedly agree to pay US 15% of China AI chip revenue

Export controls in effect turned into a negotiation over tariffs and revenue sharing

AMD
AMD makes its first MLPerf Training submission

The first primary source making ROCm's maturity checkable on a standard benchmark. The debate over CUDA alternatives moved from claims to measurements