Tractica reckons the global market for deep learning chipsets was worth $1.6 billion in 2017, jump to about $4 billion in 2018 and then jump again to $10 billion in 2019. By 2021 the market could be worth $26 billion roughly equivalent to some forecasts for the microcontroller market.
GPUs and conventional processors are being used for machine learning today but there is an expanding role for FPGAs, ASICs, SoC accelerators and other emerging chipsets. And Tractica said that in the year to May 2018 more than 60 companies, of all sizes, announced deep learning chips, chipsets or intellectual property (IP) designs.
Deep learning chipset market revenue by type 2016 to 2025. Source: Tractica LLC
Over the period 2017 to 2025 GPUs and ASICs are set to dominate. GPUs can do a reasonable job of accelerating a wide range of AI algorithms and have the advantage of being produced in high volume at leading-edge nodes. ASICs can be architected to reach higher per performance and energy efficiency but this is often at the expense of being application specific and therefore with smaller volumes and higher cost.
The annual market is set to hit $66.3 billion in 2025 representing a compound annual growth rate (CAGR) of about 60 percent.
Next: By type
ASIC solutions will lead the market in 2025 followed by GPUs and CPUs. Edge computation, where AI is provided on the electronic equipment, is expected to be more than 75 percent of the market with the balance being in cloud/data center environments. Mobile phones will be a major driver of the edge market, and other prominent edge categories include automotive, smart cameras, robots, and drones.
“Tractica expects that 2019 and 2020 will be the years when a ramp-up in deep learning chipset volumes will take place and winners will begin to emerge,” said Tractica analyst Anand Joshi, in a statement.
Related links and articles: