Courtesy of FANUC

As President Donald Trump prevents manufacturers from leaving the U.S., expect them to use robots to keep labor costs down. While this trend is likely to be greeted with alarm by union leaders, the case can be made that using robots actually helps keep whole industries from exiting American shores. Among U.S. car makers, which have been enthusiastic robot buyers in recent years, domestic employment has been not only steady, but rising. A far greater threat to U.S. workers than mechanized colleagues turning up at hometown plants is the warm welcome robots are receiving in China.

Already the world’s largest buyer of robots, China plans to close the gap with developed nations on robot density, or the number of robots in service per human worker. The idea isn’t just to drive down production costs. It’s to improve quality and one day compete more effectively in high-value goods like cars. It’s also to offset the million workers per year that China is expected to lose as its population ages. In the U.S., the demographic challenge is less dire, but present. The working population is growing slowly, at about 0.5% a year.

For long-term investors, robots could be one key to securing healthy corporate profit growth, and stock returns, even as wages rise. There are specific opportunities, too. Japan’s Fanuc (ticker: 6954.Japan) is far and away the U.S. market leader in industrial robots, and it’s quickly ramping up production. Its shares have been outperforming, and they could offer 20% more upside over the next year. Germany’s Kuka (KU2.Germany), which sold a majority stake last year to China’s appliance giant, Midea Group (000333.China), has similar return potential. Other stocks with high exposure to industrial robotics and factory automation include Rockwell Automation (ROK), Switzerland’s ABB(ABB), and Yaskawa Electric (6506.Japan). And for one-stop shoppers, there’s the Robo Global Robotics & Automation Index exchange-traded fund (ROBO), which tracks 85 stocks, charges annual expenses of 0.95%, and has returned 37% over the past year.

FACTORY ROBOTS LOOK nothing like Rosie from The Jetsons, the nameless B9 model from Lost in Space, or the one on the cover of this magazine. Many are hulking arms with rotating joints and interchangeable tools that can weld, stack, paint, assemble, and more. Among the robot definitions offered by Merriam Webster is “a mechanism guided by automatic controls.” The word itself is a play on a Czech word for forced labor, introduced in the 1921 play R.U.R., which stands for Rossumovi univerzální roboti (Rossum’s universal robots). Science-fiction writer Isaac Asimov introduced the term “robotics” and proposed a set of laws for them in his 1942 short story Runaround.

The first law, in brief, is that a robot may not injure a human being. This is more fiction than science. If there is a defining characteristic of robots that can lift cars or pack foods at blurring speeds, it’s that they’re quite capable of injuring humans. Until recently, factory robots have been isolated in cages. That is changing with the introduction of collaborative robots, which don’t lift as much, or move as fast, but can work side-by-side with people.

A Kuka robotic arm on a Mercedes-Benz assembly line Thomas Niedermueller/Getty Images

Roughly three-quarters of all robots are sold in five countries: China, South Korea, Japan, the U.S., and Germany, according to the International Federation of Robotics, a trade group. The auto industry has the most robots in use, but electronics and metals companies—which are in second and third place, respectively—have faster growth. Unit sales of industrial robots are expected to rise 13% a year through 2019, pushing the number of robots in operation to 2.6 million, up from 1.8 million currently.

Orders in China will gallop higher by 20% a year, predicts the IFR, compared with 5% to 10% in North America. The U.S. operates 176 robots per 10,000 workers, ranking eighth, between Denmark and Belgium. Japan, Germany, and Singapore are over 300, and Korea, over 500. China employs 49 robots per 10,000 workers but aspires to reach the top 10 by 2020, which would require fourfold growth.

ONE THING THAT COULD accelerate U.S. robot deployments is a corporate tax cut, which would reduce the overall cost of manufacturing in the U.S., but not the labor cost. Another is a border adjustment tax, which would reward exporters while penalizing importers. Accelerated depreciation on capital investments would give companies an immediate tax break on money spent to automate factories.

If the carrots don’t work, there’s always the stick. President Donald Trump used both—aTwitter shakedown and a state tax perk—to lure United Technologies (UTX) to keep a Carrier air-conditioning factory in Indiana, rather than moving it to Mexico. The Mexico move would have been cheaper, United CEO Greg Hayes told a television interviewer, so he will invest in automation to close the gap. “You Are Paying to Replace American Workers With Robots,” read a subsequent headline at OurFuture.org, one of many responses from left-leaning Websites. The robot makers, naturally, disagree.

A family of Fanuc robots Courtesy of FANUC

“Robots save jobs,” says John Roemisch, vice president of Fanuc America. “You can’t do it the old way just because that employs the most people. If you don’t adapt, you’re not going to survive.” In Roemisch’s view, robots replace unpleasant or dangerous jobs, but also spawn new ones. “Instead of a worker picking up a screwdriver, the robot does that and the worker operates the robot,” he says. “His skill level has to be a little bit higher, but it’s not like you need engineers on the shop floor.” Indeed, newer robots come with features like intuitive touchscreen controls that decrease the amount of training workers need to use them.

That’s one thing driving down the cost of installing robots. Another is the spreading use of camera systems and pressure sensitivity—sight and touch senses—that allow robots to learn how to handle various objects on the go. Chips from companies with videogame expertise, like Nvidia (NVDA), make quick work of the heavy thinking. Networks allow groups of robots to learn from one another.

Collaborative robots, which require less investment in safety systems, promise to expand robotics to smaller factories and more varied applications.

Fanuc uses a simple color system to help customers tell its models apart at a glance. Yellow robots, the best sellers by far, are the ones that aren’t safe for human workers to be around. Green ones, the newest line, have soft outer shells, stop gently on contact, and can be pushed out of the way—yet some can still lift 70-pound objects with ease. White robots are for health-care and pharmaceutical use. Aluminum robots, used for painting, don’t get a coat of their own.

The cost savings for a well-placed robot can be substantial. According to an example cited by the Robotic Industries Association, a trade group, a typical $250,000 installation, including training and parts, can pay for itself in two years in reduced payroll costs and increased productivity. Seven or eight years in, the cumulative cash flow gained can reach $1.5 million. Once the upfront costs are paid, medium-size robots can cost just 50 cents an hour to operate, and large robots, $1.

IN A NOVEMBER newspaper interview, Trump was asked whether he was worried about losing jobs to robots. His answer was that America should make the robots. That’s a worthy long-term goal, and the U.S. is a leader in robot research and a player in the tiny market for service robots that vacuum floors, clean gutters, and the like, as well as surgical robots. For factory robots, however, the reality is that the U.S. long ago ceded the business to overseas competitors, which have mostly consolidated in Asia. If the U.S. is to keep up with China in factory automation, it will have to use outside suppliers for now. And it should.

“Historically, the reason China hasn’t been able to export its cars is that the cars are crap, and that’s because they’ve been made using humans, rather than robots,” says Frank Tobe, editor of the Robot Report, an industry newsletter. “Now, China has a strategic long-term plan to deploy more robots, and the U.S. is only giving lip service.”

The automotive market is one where the U.S. has regained competitiveness in the face of fierce overseas competition. It’s also one where the U.S. ranks near the top in robot density: over 1,200 robots per 10,000 workers. Policy makers should think in terms of how to support duplicating that success across other manufacturing industries. More robots means more market share and, ultimately, more and better jobs.

FOR INVESTORS, ROBOT STOCKS have rarely traded cheaply relative to measures of fundamental value, such as earnings. Barron’s looked for bargains in the space four years ago and highlighted three (“Cheaper Robots, Pricier Stocks,” Jan. 19, 2013). Since then, Kuka has soared in value almost 250% to 97 euros ($103) a share, and Milwaukee-based Rockwell Automation has climbed 74%, to $153. ABB, a conglomerate with lower exposure to robotics than the others, has gained only a smidgen. We would have been better off with Fanuc, which is up 50%, to more than 22,000 Japanese yen ($193) per share. On average, our trio is up 110%, versus 60% for the Standard & Poor’s 500 index. The robotics ETF, which launched in October 2013, is up 27% since then.

For best bets now, favor companies with meaningful U.S. market share. China’s rapid growth is well understood, but an acceleration in U.S. robot demand might not yet be baked into earnings estimates. That list includes Fanuc, with a 55% share, and Yaskawa, ABB, and Kuka, with about 10% apiece. Rockwell, a partner to companies like Fanuc and Kuka, is the only sizable pure play on factory automation, which makes it a frequent subject of takeover speculation.

Click graphic for larger version

Fanuc’s high U.S. market share dates back to the launch of a joint venture with General Motors (GM) in 1982. It has a 10% share in China, too. In both countries, market share is constrained by Fanuc’s ability to produce robots, according to UBS analyst Hikaru Mizuno. The company is working to expand manufacturing capacity by about 60% through 2018—and yes, it leans heavily on its own robots to make new robots. Shares trade at a lofty 38 times projected earnings for the current fiscal year, which runs through this month. But profits are well below peak levels, due to currency effects, soft demand for machines used to make smartphones in China, and high costs to add capacity for robots that are selling well. Mizuno sees Fanuc’s earnings per share rebounding 55% over the next two years, to JPY924. The shares trade at 24 times that figure.

Kuka, like Fanuc, is a key supplier of automotive robots. Its 2014 acquisition of Swisslog, which now accounts for 20% of revenue, added exposure to automation in warehouses and distribution centers, as well as hospitals. Kuka shares sell for 30 times last year’s estimated earnings. (Fourth-quarter numbers are slated for release later this month.) But Kuka, too, is giving up some margin in the near term to invest in longer-term growth. Management aims to boost annual revenue to €4 billion to €4.5 billion by 2020, up from an estimated €3 billion last year, and to drive operating margin above 7.5% from under 6%. That looks achievable; bulls see operating margin topping 9% by 2020. If they’re right, EPS by then could approach €6, double recent levels.

ABB IS LESS EXPENSIVE than the other names on this list, at 19 times last year’s earnings, versus 20 times for the S&P 500 index and 18 times for the Stoxx Europe 600. The trade-off is that, in addition to its automation divisions, source of about 45% of last year’s revenue, ABB has a big power-grid business, which competes with the likes ofGeneral Electric (GE) and Siemens (SIE.Germany), and an electrification unit that goes up against Schneider Electric (SU.France) and Eaton (ETN).

Overall EPS is below 2013 levels, but ABB has been divesting underperforming business lines like cables and making small acquisitions in automation and robotics. Some investors have called for a sale of the power-grid unit. For now, ABB is instead working toward cutting corporate bloat to save $1.3 billion a year. Morgan Stanley analyst Ben Uglow predicts a return to growth this year that will take EPS 57% higher by 2019. ABB is adding hundreds of jobs at a Michigan plant as part of a plan to begin making robots in the U.S., a first among major competitors. The company’s American depositary receipts recently sold for $22 and change.

Yaskawa collects two-thirds of its sales from Asian markets, especially Japan and China. China is working to foster robot development at home, which could one day pose a threat to Yaskawa. For now, analysts say, the company’s expertise in software applications and reliability set it apart from Chinese upstarts in manufacturing key goods like cars and semiconductors. Yaskawa, too, is reportedly considering making robots in the U.S., although it hasn’t announced any plans. At JPY2137, its shares trade at 28 times projected earnings for the fiscal year ending this month. Like Fanuc, it has been in a profit slump, but Wall Street predicts a return to growth, with earnings rising a cumulative 40% in the two years ahead.

Rockwell Automation trades at 24 times projected earnings for its fiscal year through September, well above its five-year average of 18. Put differently, it fetches a 33% premium to the S&P 500, up from an average of 9%. Growth potential is healthy, but unremarkable, projected at close to 10% a year, compounded, over the next three years. That’s reason for caution. But Rockwell, which says that 70% of its sales now include software, could one day make a strategic asset for a larger player. Rumors swirled last fall about a Schneider takeover, but were eventually dismissed. As a stand-alone, Rockwell’s returns from here could depend on its ability to beat earnings expectations. Last quarter, it did so by more than 20%. The earnings consensus for fiscal 2018 is up 4% since the end of December.

Among other U.S. names in robotics are privately held Rethink Robotics in Boston, maker of user-friendly collaborative robots called Baxter and Sawyer, and publicly traded companies with in-house robotics operations, like Amazon.com (AMZN). It bought warehouse automation outfit Kiva Systems in 2012 and now uses tens of thousands of Kiva robots to whisk packages around. Now, Amazon is exploring residential package delivery using flying robots, also known as drones. And two years ago, privately held Uber poached 40 researchers and scientists from Carnegie Mellon University’s Robotics Institute to jump-start the ride-sharing company’s efforts in driverless cars.

Alphabet ’s (GOOGL) Google unit, too, has a robot division—Boston Dynamics—which serves up the occasional viral video on YouTube. One last summer featured a robo-dog called SpotMini that can load the dishwasher. One from this past week, which has already been viewed five million times, features Handle, which, balanced upright on two wheels, stands 6½ feet and can jump four feet vertically. The commercial applications there are unclear, but perhaps the NBA’s small forwards should be nervous.