A glimpse into the AI chip market

By Warren Miller, contributing writer

Artificial intelligence (AI) is making its way into almost every area of consumer electronics, as evidenced by its ubiquity at the 2018 Consumer Electronics Show (CES). Already a staple of smartphones and digital assistants for your home, AI is expanding into everything from cars to virtual reality. AI is even being incorporated into medical devices and technology — can you imagine surgeries in the future being performed by autonomous robots? You might not have to wait too much longer. This new wave of innovation in the AI field means more opportunities for semiconductor companies to create chipsets that can function in a variety of different systems.

One such chipset is Samsung’s new Exynos 9 Series 9810 , premium application processors designed for AI applications in smartphones and other multimedia devices. Its deep-learning capabilities and custom eight-core CPU allow it to recognize people and objects in photos and even scan people’s faces in three dimensions. The Exynos 9810 also boasts a 6CA LTE modem, meaning that it can stream or broadcast videos in UHD resolution and 360-degree video. “The Exynos 9810 will be a key catalyst for innovation in smart platforms such as smartphones, personal computing, and automotive for the coming AI era,” according to Ben Hur , vice president of System LSI marketing at Samsung.

Samsung’s Exynos 9 Series 9810 application processor for AI applications. Image source: Samsung.

NXP Semiconductor NV also has the multimedia sector in mind with its i.MX 8M application processors , specifically designed for home automation and IoT applications. The processors are engineered to provide a single platform from which to operate a multitude of home electronics and utilities via voice command, a la digital assistants like Amazon’s Alexa. The i.MX 8M series processes commands more quickly and can improve voice command functionality for things like home security systems, lighting, heating systems, and home irrigation systems, as well as multimedia platforms like smart TVs, home stereo systems, and digital streaming devices.

“Interacting with machines will be as natural as using your human senses,” said Martyn Humphries , NXP’s vice president of consumer and industrial i.MX applications. “For instance, you can give a voice command to stream a specific TV episode and then ask a contextual question about the actor, which initiates a search and displays results on the screen — all while your show is still streaming.”

Another key factor in the proliferation of AI-enhanced technologies is energy efficiency. To that end, startup Gyrfalcon has created the Lightspeeur 2801S family of high-performance AI processors, capable of operating at just 9.3 TOPS/watt. By using memory as the processing unit, the processor eschews the necessity for external memory architecture. The Lightspeeur series offers a plethora of options for platforms ranging from consumer electronic devices, IoT applications, and AR/VR systems to face recognition, language processing, and deep-learning-enabled technologies.

AI-based chipsets could be the future of the semiconductor industry. The AI chipset market is forecast to exceed $16 billion worldwide by 2022, according to a new report from MarketsandMarkets . This segment of semiconductors is driving growth opportunities for companies looking to develop and evolve.

“AI influences the development of the semiconductor sector in two ways: creating demand for new types of technologies and improving the product fabrication process,” said Jian-Hong Lin , research manager for TrendForce, a market intelligence provider.

There are a host of tech companies ramping up their AI developments, as exhibited at CES 2018 earlier this year, including speech-recognition-AI-based technologies and reference designs in addition to processors.