Over the last decade, the rise of smartphones and tablets has fundamentally transformed the semiconductor business. As the market has matured, however, an increasingly large number of companies have struggled to remain profitable. Firms like Samsung and Apple earn the vast majority of the profits on smartphone sales. A handful of SoC providers dominate the component business, with Qualcomm at the relative top of the market and companies like Rockchip, Allwinner, and Mediatek holding large swathes of the budget and midrange segments.
The above means less profit for other firms. If Computex was any indication, these companies are looking towards emerging markets in self-driving cars, the Internet of Things, next-generation connectivity solutions, artificial intelligence, and a range of other topics.
“We are going from hype phase to more a reality phase with real products. You can see them, you can feel them,” Hugo Swart, head of Qualcomm’s consumer electronics and IoT businesses, told Reuters. “I see last year was a year of a lot of promises and this year is a material realization.”
ExtremeTech spoke to Anshel Sag, an analyst with Moor Insights & Strategy, who talked about how the focus and nature of the Computex trade show is changing. “Computex was always more of a traditionally PC hardware for both consumers and business show. Now it’s turned into more of a high-end PC hardware show with IoT to boot,” Sag said. “Many of the Taiwanese vendors want to get involved in IoT and that’s why you’re seeing so much support for them from the chip vendors.” Chip vendors, in this case, refers to companies like ARM, Intel, MediaTek, Qualcomm, and other SoC designers.
These reports jive with what we’ve seen from companies like AMD, Nvidia, Tesla, Uber, and Waymo over the past few years. After largely sitting out the AI market, AMD is positioning its upcoming Vega products for AI and HPC workloads, Nvidia is making a huge push into self-driving vehicles, Tesla already has its Autopilot feature on the market, and Alphabet’s Waymo project could team up with Honda to bring self-driving cars out by 2020. We tend to think about the markets for these products as being strictly tied to the sale of just one vehicle, but the number of discrete components that go into creating one self-driving car is enormous.
According to Sag, devices like smart toasters and fridges draw a fair amount of derision, but they aren’t where most companies are actually focusing their R&D efforts. At the same time, however, nobody really knows what killer capabilities, analytics, or technologies will be the first to drive mass consumer adoption and support. Wearables, to date, have only had a modest impact. But that could change as technology, battery life, and software all improve. Nobody wants to be caught napping when and if that happens, so we’re seeing a lot of mud get flung at the wall in the name of finding out what might stick.
Let’s block ads! (Why?)