ASIC asserts rights to ML chip market

The Machine Learning Chip Market report forecasts the global ML chip market will be worth $37.8 billion in 2025, demonstrating a compound annual growth rate (CAGR) of 40.8% between 2018 and 2025.

The report considers chips by type (GPU, neuromorphic, flash-based, ASICs and FPGAs) across consumer electronics, automotive, healthcare and the robotics industry, where the rise in the use of artificial intelligence (AI), ML and quantum computing are expected to increase demand for ML chips.

Although the GPU is the largest sector, accounting for nearly one third of total revenue in 2017, it is expected to be overtaken by the ASIC market in 2025, which is predicted to grow at 21.33% CAGR from 2018 to 2025.

Other factors driving demand are the proliferation of smart homes and cities, the quest for more human-like robots, and the spread of the IoT, although this will be counter-balanced by “AI phobia” which may restrain the market.

A machine learning chip can be used across media and advertising, manufacturing, finance, healthcare, automotive and transportation, for time-saving monitoring and prevention.

AI encompasses ML, robotics, and collaborative systems. Here, an ML chip can accelerate the system without human intervention. Examples of use range from chatbots to autonomous cars.

A warning is sounded in that ML chips require complex algorithms and software engineering skills which are in short supply. The report warns that a “dearth of skilled workforce is a prominent restraining force of the market”.

The report discuses technologies such as system-on-chip (SoC), system-in-package (SIP), multi-chip module, and others. The system-on-chip segment is likely to account for nearly half of the total market revenue by 2025.

Regionally, North American currently dominates the market, while Asia-Pacific is expected to grow fastest at a CAGR of 21% through 2025 (Figure 1).