(1) For NVIDIA AI chips, see
GPU and
CUDA.
(2) For Google AI chips, see
Tensor Processing Unit and
Tensor chip.
(3) For AMD AI chips, see
AMD Instinct.
(4) For Amazon AI chips, see
Amazon AI chips.
(5) For Microsoft AI chips, see
Maia chip.
(6) For the Groq AI language processing unit (LPU), see
Groq. (This is not xAI's Grok.)
(7) For a chip dedicated to AI image generation, see
Normal Computing.
(8) For the photonic AI chip, see
Native Processing Unit.
(9) For wafer-scale AI chips, see
Cerebras AI computer.
(10) For low-power AI chips, see
Normal Computing and
Native Processing Unit.
(11) For the microcontroller that uses an AI convolutional neural network (CNN), see
Pulsar.
(12) For companies that make AI chips, see
semiconductor companies.
A Decade Makes a Difference
The ARM-based chips Apple designs and uses in its phones and tablets have evolved significantly over the years. There was no AI processor in 2015, but there was in 2025. See
neural processing unit.
Apple A9 Apple A19 Pro
(2015) (2025)
14 nm chip 2 nm chip
2-core CPU 6-core CPU
6-core GPU 6-core GPU
6x 0.65GHz 6x 1.62GHz
--- 16-core neural engine
The Versal System-on-Chip
Debuting in 2018, this system-on-chip (SoC) from AMD has an unusual feature. The "adaptable engines" (red) are actual logic circuits that are programmed by the customer. The "intelligent engines" (green) include AI and digital signal processors (DSPs). (Image courtesy of AMD.)