Artificial intelligence is developing rapidly, and deep learning chip sales are expected to reach 2.9 billion in 2025

GPUs and CPUs currently lead in market share, but ASICs will take the lead by 2022, with opportunities growing for SoC accelerators and FPGAs.

In today’s AI market, hardware is key to solving many of the industry’s challenges, and chipsets are at the heart of that hardware solution. Given the widespread use of AI, it’s almost certain that every future application will require some kind of acceleration using an AI chipset, whether in the data center or at the edge. Acceleration can take many forms, from simple AI libraries running on the CPU to more complex custom hardware.

Tractica predicts that this growth and evolution of the AI ​​market will drive deep learning chipset shipments from 164.9 million in 2018 to more than 2.9 billion units in 2025. By then, the global market for deep learning chipsets will reach $72.6 billion. By 2025, application-specific integrated circuits (ASICs) will account for the largest share of total revenue, followed by graphics processing units (GPUs), central processing units (CPUs), system-on-chip (SoC) accelerators, and field-programmable gate arrays (FPGAs).

“The deep learning chipset market has experienced a period of rapid development over the past two years, led by NVIDIA and Intel. However, emerging ASIC chip companies are somewhat behind in delivery timelines. Validation work for the edge market has begun, and validation work for the enterprise market will begin in 2019. The rapid growth of deep learning chipsets will start from 2020, during which time the winners will begin to emerge,” said Anand Joshi, principal analyst at Insights.