Term of the Moment
Ethernet switch
Definition: AI chip
An enhanced GPU chip used to train and execute AI systems. The training side (machine learning) requires the most processing, and large language models cause quadrillions of calculations per second to be executed for days, weeks and even months.
New to the AI Chip Race
In 2023, AMD introduced its MI300 AI chips, which are expected to generate considerable revenue for the company (see Instinct MI300). In 2024, Amazon developed the Trainium and Inferentia chips (see Amazon AI chips). There are also smaller companies racing to develop chips for the runtime (inference) side of AI. For example, when people ask ChatGPT or other chatbots a question or request that new content be generated, it is the runtime part of these AI systems that develops the results. See AI training vs. inference. See Tensor chip, neural processing unit, deep learning and Cerebras AI computer.
The Versal System-on-Chip
New to the AI Chip Race
In 2023, AMD introduced its MI300 AI chips, which are expected to generate considerable revenue for the company (see Instinct MI300). In 2024, Amazon developed the Trainium and Inferentia chips (see Amazon AI chips). There are also smaller companies racing to develop chips for the runtime (inference) side of AI. For example, when people ask ChatGPT or other chatbots a question or request that new content be generated, it is the runtime part of these AI systems that develops the results. See AI training vs. inference. See Tensor chip, neural processing unit, deep learning and Cerebras AI computer.
Today's chips often include AI processing. This Versal system-on-chip (SoC) contains more than 30 billion transistors and provides circuits for AI processing (green area). It also contains programmable hardware, which means the actual circuits are programmed at startup, a rarity on any SoC (red area). See SoC, FPGA and Versal. (Image courtesy of AMD.)