CryptocurrencyInvestingBanksReal Estate

How to invest in the Internet of Things

December 4, 2014, 1:05 PM UTC

With the “Internet of things,” it’s not a matter of “if.” It’s a matter of when, how big, and who will reap the princely profits. That’s the thinking, at least, among many investors, tech conglomerates, and investment banks. They see it as the biggest opportunity since smartphones and tablets swept the world.

You may or may not be attracted to the idea of using your phone to control your thermostat and home security system from miles away, or wearing a smartwatch or fitness tracker—the products that leap to mind when the Internet of things is mentioned. But the term also encompasses a much larger and less visible universe of uses—everything from cars to oil rigs and factory machinery that sends data to one another. Between the possible consumer and business applications, analysts have been tripping over each other to make the most grandiose predictions: 1.9 trillion from Gartner (IT), 7.1 trillion from IDC, 19 trillion from Cisco (CSCO). Are they referring to devices or dollars? What’s the difference? It’ll be huge! (For the record, they’re talking about dollars.)

For investors, the frenzy may cause a familiar anxiety: Call it the fear of missing out on the next big thing. In truth, there’s bound to be some disappointment. There are relatively few companies to invest in, and those with the biggest opportunities are either nascent and risky or buried inside enterprises so large that the effect of the connected products will be diluted. But, as we’ll see, there are some opportunities in unexpected places.

The very excitement among venture capitalists and large tech companies is an impediment for retail investors. For example, when Josh Elman, a partner at venture capital firm Greylock Partners, invested in “smart home” startup SmartThings, he expected to wait two to three years before the sector took off. Then Google (GOOG) bought Nest, maker of Internet-connected thermostats, for $3.2 billion. Shortly after, Nest itself scooped up Dropcam, a maker of Internet-connected video­cameras, for $555 million. “The moment that happened, all hell broke loose,” Elman says. With Google moving aggressively, large tech companies scrambled to come up with their own strategies. In practice that means acquisitions. Sure enough, in August, less than a year into Elman’s investment, Samsung snapped up SmartThings.

As a result there are few pure plays for investors. SmartThings won’t contribute meaningful income to Samsung (SSNLF) for years. Nest and Dropcam’s revenues will cause nary a ripple on Google’s $56 billion top line. Other successful consumer products—Pebble, Fitbit, and Jawbone—are still privately held. The closest thing to a good bet in this realm is Garmin (GRMN), a car navigation company that remade itself as a seller of smartwatches, fitness trackers, and pet-tracking gadgets. Last quarter, half of Garmin’s revenues came from sales of nonautomotive products—its stock is up 33% this year—but its moment in the sun may be brief: It and other smartwatch makers could soon see their sales crushed by Apple’s offering.

Fortunately for investors, consumer products are only a small piece of the overall market. The uses for sensors, a component in virtually every connected device, are endless. They are expected to penetrate just about every category of product, from stoplights and parking spaces to jet engines and tires. “It’s not just going to be a technology investment thing,” says Deborah Koch, co-manager of Northern Trust’s technology fund. “It’ll be a productivity movement that will drive the entire economy.” Koch contends that it’s too early to pick winners, as older sectors aren’t likely to adopt this technology for another three to five years.

Much of the progress on Internet-connected devices is occurring inside old-economy stalwarts such as GE. The company attributes more than $1 billion in revenues to some 43 “industrial Internet” offerings, and the biggest buyers have been aviation and locomotive customers, according to CMO Beth Comstock. Those ­companies already have IT departments that know what to do with data from sensors, like improving fuel efficiency and making trains run faster, Comstock says, stressing that the business is in its earliest days.

Rather than focus on end products, many investors are betting on the suppliers. After all, someone has to make the billions of chips that enable the connections. “Relative to the more obvious ways to play the Internet of things, we view the demand for chips as being healthy and being underpriced for the market,” says Paul ­Ebner, senior portfolio manager at BlackRock (BLK). His firm has backed chipmakers that sell into the auto industry, since carmakers are buying 8% of semiconductors. (Indeed, these days cars are as likely to tout their 4G LTE capabilities as they are the horsepower of their engines.)

Qualcomm (QCOM), the largest chipmaker in the world, powers almost every new smartphone in some capacity. That dominance will make the company an important player. Indeed, its recent $2.5 billion acquisition of British semiconductor company CSR indicates Qualcomm’s intention to go big in the Internet of things: CSR sells chips for cars, printers, and wireless audio, new markets that Qualcomm CEO Steven Mollenkopf has said will drive the company’s projected 8% to 10% annual growth in the coming years. The company’s share price was recently bruised by news of disputes with Chinese licensees, leaving its price/earnings ratio at 14, a bargain for a growing tech company.

Then there’s ARM Holdings (ARMH), which licenses the processor architecture that powers 95% of the world’s smartphones. The small company ($1.8 billion in 2013 revenues) has trounced giant Intel in mobile chips, and ARM is already on its way to becoming a top licenser of connected devices: Half of the 10 billion chips its licensees sold last year were for nonmobile items like appliances. In October the company introduced a chip designed specifically for use in factory machines, cars, and smart homes.

The closest one can get to a pure investment among the enterprise-focused companies is Sierra Wireless, a wireless-device maker with a $1.2 billion market cap. The company earns all its income from the category, serving auto­motive customers like Chrysler, Renault, and Dezo, which supplies Toyota, as well as industrial companies. Sierra has a high forward P/E of 33, but it’s justified by the company’s expected annual earnings growth of 81% for the next five years. “You’re able to move the needle faster on a smaller company at lower penetration levels than you are to move the tanker ship such as Cisco, Qualcomm, or GE,” says John Bright, a director and senior research analyst with Avondale Partners.

“It’s not just going to be a technology investment thing,” says Northern Trust’s Koch. “It’ll be a productivity movement that will drive the entire economy.”

Cisco, the tanker ship, is moving as fast as it can. CEO John Chambers has staked the company’s future on what it calls the “Internet of everything,” introducing 800 products, like wireless networks for mining sites and manufacturing floors. “Within the networking industry, they have been way out ahead of the competition,” says Goldman Sachs analyst Simona Jankowski. Cisco’s 10,000 connected-equipment customers make up $2.4 billion in revenues out of the company’s $49 billion total. That may not be enough to revive its shares, which succumbed to rigor mortis after the dotcom crash of 2000 (the company has attempted multiple turnarounds as demand for its networking hardware shrinks) and have shown no movement ever since.

Still, all those toasters, garage doors, and air conditioners need to connect to the web. In the short term, Cisco and its peers will be in high demand. Ted Scalise, who manages the TIAA-CREF Mid-Cap Growth Fund, which has averaged 16.9% returns per year over the past five years (vs. 15.9% for the S&P 500) believes that web-connected hardware will eventually become commoditized. But he thinks wireless networking companies like Aruba, Ruckus Wireless, and Netgear could enjoy some good years before pricing pressure crimps their returns.

Once the hardware is in place a few years from now, experts say, software players will rise to help companies make sense of all the data the hardware is collecting. “It’s still evolving, and it’s the linchpin to how this works in the future,” Koch says. Big-data companies such as Splunk (whose shares she owns) and Hortonworks (which has filed to go public later next year) are out in front of the trend.

For now the potential of the Internet of things radically outpaces the reality. Research firm Gartner predicts the hype will soon collapse into a “trough of disillusionment,” followed by a “slope of enlightenment,” and then, eventually, a “plateau of productivity.” Notes Tim Herbert, vice president of research and market intelligence at CompTIA: “We often overestimate a technology’s impact in the short term and underestimate it in the long term.”

The good news is that the media hype has not translated to stratospheric stock valuations. “This is more like 1995 in the Internet phenomenon more than, say, 1999 or 2000,” says Jankowski of Goldman Sachs. “It’s a real trend, and there is a lot of interest, but I wouldn’t say anything has overheated yet.” The keyword, of course, is “yet.”

This story is from the December 22, 2014 issue of Fortune.