The future belongs to the machines. More than one billion smartphones will be shipped this year, and several million smartwatches and fitness monitors of various sorts will be worn. But if predictions about the so-called Internet of Things are right, the number of unmanned computing devices that populate the world, connected over wired and wireless networks, eventually will dwarf the number of devices that people carry around with them.
Market research from Gartner predicts that the tally of network-connected devices, most of them not manipulated by a person, will soar from three billion to 25 billion in just seven years.
Illustration: Dan Picasso for Barron’s
These connected things span industries. They include automobiles with built-in Internet connections, packaged goods on store shelves that can have wireless transceivers in their labels, and fleets of trucks on the road connected to networks that track their progress.
In addition, heating, ventilation, and air-conditioning systems will be better able to control humidity and temperature, airlines will make smarter use of fuel, and hospitals will better coordinate the availability of operating-room equipment with surgical appointments.
TWO CLASSIC TECH TRENDS are at work here.
First, technology, as it often does, is democratizing the process of connecting things. Industry titans such as General Electric (ticker: GE) have been automating industrial-control systems on oil rigs and such for decades. And the kinds of networked infrastructure mentioned here can be built in soup-to-nuts fashion by IBM ’s (IBM) Smarter Planet business, which generates $5.5 billion in service revenue for the tech giant each year, analysts estimate.
But typically, “those kinds of approaches could only be afforded by the biggest companies in the world,” says Ed Maguire, a software analyst with boutique research firm CLSA Americas, who has written numerous reports this year digging into every nook of IoT. “It was not something that was accessible to the average company.”
With the decline in the cost of sensors, the proliferation of inexpensive wired and wireless connectivity systems for computing, and the arrival of affordable data centers from Amazon.com (AMZN) and other providers to gather and analyze information from machines around the world, it is becoming increasingly possible for companies large and small to be connected in ways that only the biggest once dreamed of.
Second, capabilities are ramping up. Now, for example, a simple motion, light, or temperature sensor might be attached to a thing and used to control it. But soon, the software behind a network of sensors will monitor an entire city’s street lamps, making them more efficient and saving taxpayers money. Or airlines might use a service that knows when parts for a jet engine are about to fail and can have a replacement delivered to minimize downtime.
It’s similar to how the PC initially replaced pen and paper for number-crunching, but later led to the proliferation of far more valuable inventions, such as databases and the World Wide Web.
The explosion of connected things offers opportunities for technology companies across the spectrum, including chip makers such as Qualcomm (QCOM) and Thin Film Electronics (THIN.Norway), traditional networking firms such as Cisco Systems (CSCO), software vendors such as PTC (PTC) and Splunk (SPLK), and cloud-computing service providers such as Amazon and Salesforce.com (CRM).
Although Amazon, Cisco, Qualcomm, and the like are too big to get any near-term kick from the Internet of things, Splunk, Thin Film, and PTC are highly leveraged to its growth (see table).
Another big payoff will accrue to old-economy companies such as Monsanto (MON) and GE. They may produce a renaissance from things as humble as industrial equipment and crop seeds.
THERE ARE THREE STAGES to the IoT. First, lots of things get connected, including homes, cars, freight trains, medical equipment, and packaged goods. Next, more and more analyzable information is extracted from those connected devices. And in the final stage, manufacturers run their businesses based on that data.
The first stage is all about cheap sensors and cheap chips. There are plenty of flashy examples, such as electric-car maker Tesla (TSLA). Last week, it announced a new version of its Model S, with sensors that over time will lead to autopiloting on the highways.
But there are humbler examples proliferating out of site. Chip start-up Gimbal of San Diego was spun out of Qualcomm in May. It makes what’s called an iBeacon, a name actually trademarked by Apple (AAPL), which developed the standard.
Gimbal’s iBeacons are pieces of plastic costing $5 to $20 that can be as small as an inch square and contain sensors, wireless transceivers, and some memory. They can be stuck in various places and transmit information to devices that pass by them as far as 50 meters away.
Apple uses Gimbal’s devices in its retail stores: When a shopper browses, his iPhone automatically connects to the iBeacons hidden under display tables. The iBeacon then sends alerts to his screen, such as a reminder about an upcoming Genius Bar appointment.
That’s today, but in the future this technology will find its way into uses that involve no direct human participation, predicts Gimbal’s chief operating officer, Kevin Hunter. For example, some partners, which Hunter declined to name, use temperature sensors built into the iBeacon to send back information about the climate at points throughout a facility, so that an HVAC system can be adjusted optimally.
Price is key: In the past, the approach would have involved Wi-Fi wireless access points costing 10 or 20 times as much as an iBeacon network does.
EVEN CHEAPER are electronics printed on flimsy pieces of plastic. Thin Film of Oslo has the technology to imprint sensors and circuitry on such transparent materials. Thin Film CEO Davor Sutija says that the actual sensor, which is just a single black line traced onto the plastic, costs a half-cent to produce. A full sensor hub, which has a central processing unit, radio-frequency circuits, antenna, and even a battery with a few months’ life, costs a buck. In a few years, that could drop to just a quarter. “Think about it, four times less costly in a few years,” says Sutija.
Sutija is working with partners and customers to put the flexible printed smart labels on the exterior of packaged goods such as produce and diary. A supermarket worker or a consumer could, say, check the temperature status and history of these items on a digital display. An estimated five trillion to 10 trillion disposable items were sold globally in 2012. If each contained a printed electronic label costing, say, 1% of the item’s retail price, Sutija envisions a $100 billion market.
The next stage is for data to be harvested from all of the connected devices in the field.
PTC, which has for many years sold software used to design products, has acquired a new business called ThingWorx. It lets companies throw together feeds from sensors in the field and combine them with network resources of various kinds. A customer could take sensor data from machinery and pair it with the names of customers in a database run by Salesforce, the most prominent outfit offering customer-relationship management software, as a cloud-computing service.
CLSA’s Maguire is enthusiastic about software vendor Splunk, which got its start helping information-technology teams sort reams of data about the health of their computer servers. An IoT business unit could do something similar with the flood of data coming from sensors.
One customer is New York Air Brake, a maker of locomotive parts such as valves, hoses, and cylinders. Using Splunk, the company can analyze tons of data about vibrations, track conditions, and motion coming from sensors on freight trains.
“Like Microsoft had with Windows, Splunk is a platform,” says Maguire. “They have more than 7,000 customers and many thousands of developers, and they are building this enormous concentration of expertise by solving business problems.”
Chip giant Intel (INTC) will be among the companies bringing that data to the cloud. It already sells chips to 20 different computer assemblers building gateways—boxes that can be put in the field, either attached to smart things or near them. Gateways can perform some analysis themselves, or can send data to the cloud for heavier lifting. Intel’s chips have been selected for GE’s analytics efforts.
Illustration: Dan Picasso for Barron’s
Today’s venture-backed start-ups probably will quickly command investors’ attention when they go public. Jasper Technologies of Mountain View, Calif., provides some of the functions behind the OnStar connected-vehicle system. Jasper also monitors smart devices in the field for numerous clients. Another start-up, Aeris Communications of Santa Clara, provides telecoms with technology to sell remote monitoring of parking meters, vending machines, and other connected devices.
IN THE THIRD STAGE of IoT, companies change how they earn revenue and learn how to cut costs. This is perhaps the hardest area to quantify because it involves teasing out all of the inefficiencies in existing business models.
Cisco, which has realigned its business to sell packages of switches, software, and services for the Internet of things, points to the example of Barcelona, Spain, which is using sensors and networks to gain greater insight into traffic patterns and reduce congestion. “There is no sector of the economy where there aren’t billions of dollars trapped in inefficiencies,” observes Harbor Research analyst Glen Allmendinger. “Look at the way PCs in the 1990s improved office productivity. The impact of IoT will be many times greater.”
Consider the maintenance of GE jet engines. When an airline such as United needs a part for a plane already at a gate, it usually barters with another carrier with a nearby plane, causing delays. However, with the IoT, says Allmendinger, GE could remotely monitor the engine, anticipate the part’s failure, and quickly deliver a replacement, so the jet wouldn’t be stranded on the tarmac.
“Our genesis is always about the machines, but I think industrial companies really have to go long in analytics and software,” GE CEO Jeffrey Immelt tells Barron’s. “We’ve always been a services company by the nature of the way our assets operate, but [the Internet of Things] is going to make us even a bigger services company as time goes on.”
Maguire agrees. “The investing opportunity is for nontech investors, as nontech companies pull away from the pack” by making their businesses smarter, and hence faster-growing and more profitable, he says.
For example, Jones Lang LaSalle (JLL), the Chicago-based property manager, has devised a computer system called IntelliCommand to more intelligently oversee a building’s electrical usage. Jones Lang claims that, in a building it manages for Procter & Gamble, the system’s cost was offset in just three months by the amount saved on energy.
IN A SIMILAR VEIN, MONSANTO spent $930 million to buy Climate Corp., a weather-prediction start-up. Using ground sensors that measure pH levels and soil composition, the big seed producer hopes to offer a service that can tell farmers which crops will work best in their fields under a given weather condition.
The biggest issue looming over IoT is security. Obviously, very bad things could happen if someone hacks into networks connected to the nation’s energy grid or rail system.
“Lots of people have put things on networks far faster than they have taken time to secure them. That’s the biggest risk, categorically,” says Allmendinger. Here, too, there may be future tech stars, including start-up Mocana of San Francisco, whose business is application security. Another could be Symantec (SYMC), which probably doesn’t get enough credit for the work it is doing to devise safeguards for the IoT era.
In the meantime, while strolling through a thicket of iBeacons, we humans might contemplate whether we like living in an age in which buildings are becoming smarter than us, shop-floor robots are taking more of our jobs, and self-driving cars can be remotely repossessed. Of course, our concerns might be irrelevant. People aren’t the centerpiece of this computing revolution; machines are.