Cover Story: Beyond the Smartphone: The Next Big Thing
The semiconductor industry has been on a roll, pushing well past $300 billion in annual revenues. But with smartphones and tablets showing signs of slowing down, the industry’s perennial question returns: What is the next big thing?
Lacking a clear consensus on specifics, most market watchers nonetheless agree that it will have something to do with the Internet of Things (IoT). A June 2014 report by International Data Corporation (IDC) estimates that sales of billions of smart, connected gadgets like Google Glass will generate a stunning $7.1 trillion in annual sales by 2020, with most of the total in services and supporting software.
According to Rob Lineback, an analyst at IC Insights, some people in the industry are looking for one or two huge growth drivers; others believe the next big thing will be made up of many products to serve many markets.
Automotive electronics is a large segment that is showing healthy growth, he said. Wireless medical systems and fitness trackers are generating a lot of interest, and the category is projected to grow by a CAGR of about 37%, reaching $4.7 billion in 2017. The semiconductor content, however, “is just not big enough to move the needle much for a $300 billion industry,” said Lineback, who co-authored the IC Insights IC Market Drivers Report.
He noted that medical systems are set to grow to about $65 billion by 2017, while medical semiconductors are roughly a $5 billion market this year and forecast to grow to $6.8 billion by 2017—a relatively small market compared with PCs or smartphones.
“When we look ahead, there are plenty of things to watch. Some will be a fad, a flash in the pan, while others may take hold. It’s the old saying: ‘You don’t know what the next big thing is until you get there,’” Lineback said.
A Medical Breakthrough?
Eileen Tanghal, director of Applied Ventures LLC, a venture capital fund, sees medical as a prime IoT opportunity. A variety of medical testing, personal diagnostic, gene sequencing, and drug discovery procedures are still being done on glass slides in labs, using labor-intensive methods.
“Replacing those with silicon-based testing would significantly reduce the cost-per-test and speed up the ability to develop new drugs,” she said, estimating that if the medical testing and remote diagnostics industries in the developed nations move to silicon-based solutions, it could add about $1.2 billion to the worldwide semiconductor equipment market. If these tools are adopted worldwide, the impact on the silicon equipment industry could reach the $5 billion level.
Food production and clean energy are two other areas ripe for silicon-based solutions, according to Tanghal.
Charting a Course for IoT Data
Mike Rosa, a MEMS expert and senior strategic and technical marketing manager for 200mm Emerging Technology Products at Applied Materials, believes that the semiconductor industry’s next wave lies in the thousands of IoT device applications for sensors, microcontrollers and communications that will combine to help people make decisions. “But generating value out of the data, gathering the right data to make an informed decision: therein lies the value of the Internet of Things,” he said.
Indeed, thinking primarily about the connected gadgets at the “edge” of a network is selling the IoT short. Alongside the low-cost edge devices are gateways, cloud-based computing and all of the services and software required to analyze the data streaming in from edge devices.
While much attention is focused on the low-cost “intelligent edge” devices, David Formisano, Intel’s director of strategy of integrated Internet of Things, said Intel will profit from the need for gateway products and cloud-based computing services. As data comes streaming in from sensors, gateway systems will filter data so as not to overwhelm the network infrastructure. “A gateway device would decide: ‘Is this important? Should I send it or not?’”
Speaking at SEMI’s recent IoT conference, held in Austin, Texas, Formisano said the first IoT waves are being seen in industrial applications, such as sensor-driven systems for manufacturing semiconductors, chemicals, and other high-value products. “Intel is using the IoT to make our chips. Gone are the run cards affixed to the run box. Today’s leading-node factories rely on high levels of advanced process control and large amounts of data coming through the sensor networks,” he said.
Figure 1. The Internet of Things will create a dramatic transformation in industrial economies, as connected devices, cloud-based computing, and Big Data analytics converge. The biggest impact, already underway, is in manufacturing, with an estimated $3.9 trillion in added value over the next ten years.
“Also interesting is this notion of ‘brown field’ things, which refers to the 85% of the things out there which are not connected. Street lights are still not intelligent, semi-trailer trucks are still not part of a fleet management system. To get them connected you don’t have to build a brand new bus, or truck, or train. So there is a lot of excitement today in getting today’s things on the IoT,” Formisano said.
As demand for IoT devices skyrockets, Jeremy Read, head of service marketing at Applied Global Services, said semiconductor manufacturers increasingly will upgrade their existing 200mm tools to run the modern automation software and advanced process control (APC) applications that are key parts of what he calls an industrial IoT (IIoT) approach to chipmaking. Data gathered from APC technologies and by sensors attached to thousands of machines can be used to detect, classify, diagnose, control, predict and prevent various failure modes. “It is the new data that makes predictive maintenance a reality,” Read said.
Standards and Security
Bill Curtis, director of new business development at ARM, Ltd., argued that much work remains to be done to develop the standards required for high-volume IoT markets. “The things around us are increasingly smart and connected, creating an opportunity for shipping billions of devices. However, today we live in a siloed world. There are many different standards, and almost none of the stuff out there can communicate.”
Formisano said the corporate world needs interoperability standards so that IT and operations organizations can share data. And ever since the public became aware of how much data is being secretly gathered, security and privacy concerns are increasing. “Security and privacy is a big challenge. The government is interested in the IoT, but almost every week is asking, ‘How do we deal with the privacy and security issues?’”
Costs are another major concern. To put a sensor, control circuitry, Internet communications capabilities, and some form of energy harvesting on an edge device selling for less than a dollar or two is a formidable challenge. Curtis points to ARM cores aimed at IoT applications: 32-bit processors consisting of just 12,000 gates. “The processors are pretty doggone small. Going forward we will want a world of very small, very cheap devices,” he said. And a whole new world of communications protocols is evolving to ensure that data can be sent over the Internet at very low levels of power consumption.
Rajiv Kumar, marketing manager at Freescale Semiconductor’s microcontroller unit, said tiny 32-bitters are powerful enough to recognize hand-motion interfaces, such as those used on NEST home controllers. Kumar sees the IoT putting new vigor into the electronics industry, as startups work on “very tangible problems” across all walks of life. “Now, almost everyone has a mobile phone, and tablets compete with PCs for market share. I see the next wave as a lot of connected waves, becoming more and more relevant.”
Freescale offers a multitude of IoT-class microcontrollers, using a Lego-like building-block approach to quickly configure ARM-based solutions. And design teams can download tools and use credit-card sized development boards to quickly come up with IoT-type solutions for hundreds of markets, Kumar said.
Printed Electronics Needed?
Rosa said he believes silicon-based solutions will not be inexpensive enough to meet the cost targets mandated by many IoT applications. Printed electronics on cheap plastic substrates will finally come to the fore, undercutting the cost to process silicon wafers. While some market research firms predict that a trillion sensors will be in use by 2020 or 2022, Rosa said that is unlikely unless production costs are driven down. “There are about 12–14 billion MEMS in use today, so a trillion sensors by 2020 is not going to happen unless we can cut the manufacturing costs by one to two orders of magnitude.”
Dean Freeman, semiconductor manufacturing analyst at Gartner, said printed electronics are appealing, but thus far they haven’t found a commercial high-volume application. “Many of the IoT-class MCUs and communications devices will be made in 300mm fabs, using legacy process technology.”
Freeman calculates that 10 billion IoT things, each with an average of three MEMs devices, may be made in 2020. That would create demand for about 22 new or converted 200mm fabs, each running 50k wafers per month. “That is assuming an 80% yield and a 4mm2 device. That is about 1.1 to 1.2 million wafer starts per month,” he said. Viewed another way, a moderately successful IoT market would add about 10% to the annual wafer fab equipment (WFE) expenditures, about equal to today’s market for refurbished equipment, he added.
The Next Big Thing: IoT or I Don’t Know?
So all of this debate leaves us with a few questions: Will the next big thing rely more on chip technology, new materials technology, or software? Or all of the above? And at the end of the day, what is it that will have us lining up at some big-box store at midnight to buy one?
Applied Ventures’ Tanghal, who considers herself an optimist, firmly believes that something will emerge as a high-volume, high-value product akin to what the smartphone has been since 2007. The next big thing is out there, she said, probably being developed by young people in ambitious startups with dreams of a buy-out by companies like Google, Amazon, or others. Maybe personal robots, drones or other forms of human assistance systems will be winners. Or something else?
“Innovation doesn’t stop,” said Tanghal.
What do you think will be the next big thing? Send your ideas to firstname.lastname@example.org