Qualcomm Sees IoT Driving Growth, AI at the Edge

SAN JOSE, Calif. -- Qualcomm executives this week sketched a broad view of the growth opportunities in the Internet of Things (IoT) and how it will affect the chip market, including shifting more computing power and security needs to the "edge" of networks, including actual IoT devices themselves.

It's a compelling view of a big shift in computing architecture, as IoT drives the need for more techniques such as deep learning and artificial intelligence (AI) to be embedded directly into IoT devices or IoT gateways to fuel applications such as connected cars, augmented reality/virtual reality (AR/VR), and voice-command interfaces for all kinds of devices.

Put it all together, and the result will be a steady growing market for powerful chips, according to Qualcomm. To back this claim up, Qualcomm announced on Tuesday that it is selling one million chips per day into IoT markets, which includes wearables, voice and music, connected cameras, robotics and drones, home control and automation, home entertainment, and commercial and industrial IoT.

Raj Talluri, SVP and GM of Qualcomm's IoT business, told analysts at the IoT workshop here on Tuesday that IoT is a solid, profitable, and growing business for Qualcomm -- but that hype about the market may have elevated expectations.

"The hype [on IoT] has been a lot more than what it is," said Talluri. "But we are happy with the numbers. Our number is growing. It's actually a sizable business and it's profitable. The hype has been so big that some people have expected something much bigger."

Outlining the market development, the theme of edge computing was popular. Qualcomm executives said that voice user interfaces are becoming more prevalent, and that will drive the need for computing power and deep learning, whether that be in an IoT gateway, kitchen appliance, automobile, or device itself, such as a drone. While some AI and deep learning will always be executed in the cloud, more of this function can be offloaded to edge devices, said Qualcomm executives.

"At the edge you have connectivity, but it very quickly grows into sophisticated edge computing," said Seshu Madhavapeddy, VP, Product Management, in an introductory IoT session. "Deep learning is a gold mine in IoT edge computing."

This is a broad trend expected to drive growth for more sophisticated IoT chips at the edge that have wide-ranging image-processing, connectivity, compute, and security features.

Madhavapeddy said that Qualcomm is working with ecosystems of specialized companies to develop convolutional neural networks (CNN) solutions -- a form of deep learning -- for specific use cases for connected cameras, smart assistance, robotics, drones, and AR/VR. The new generation of chips will be optimized for handling problems such as neural networking processing at the edge.

Another area that Qualcomm will be focusing on is security, by building security into the chip level for IoT applications. Qualcomm said chips targeting IoT will come with security features such as Secure Boot, Secure Storage, Secure Debug, and hardware-based key provisioning.

Jeffrey Torrance, VP Business Development, gave an overview of the Industrial IoT space, which he described as growing reliably.

"It's boring, it grows nicely, and it makes money. That's what we like about this Industrial IoT space."

Torrance estimated the serviceable addressable market (SAM) in IIoT at about $1 billion and growing 20 percent annually. Some of the key markets include:

• Industrial Security
• Energy & Metering
• Industrial IoT Gateways
• Industrial handhelds
• Asset Tracking
• Retail payments

Of course, all this doesn't mean that IoT is an easy gravy train. Some of the challenges discussed during the IoT workshop include fragmented IoT standards (see: Qualcomm Sees Demise of Proprietary IoT Nets) and a wide variety of needs for distinct vertical markets.

It's a big project and opportunity for Qualcomm, which soon will be reorganizing its entire business to pave the way for growth in IoT as it merges with NXP Semiconductor, which leads development of the connected automobile market. After the merger, Qualcomm's semiconductor business is expected to be organized around three areas: 1) Mobile (smartphones, tablets, mobile computing), 2) Automotive (the bulk of the NXP business), and 3) IoT.

It's quite possible that after the NXP merger, Qualcomm will become the largest provider of chips into IoT, which makes it a good proxy for the industry.

For a more detailed analysis of the Industrial IoT market, purchase our 50-page Ultimate Industrial Internet of Things (IIoT) Report, which covers a wide range of communications and cloud technologies that are being applied to businesses around the world to provide connectivity, analysis, automation, and optimization of a range of industrial applications.