Limping Along In Robot Land

Once it was hailed as the ultimate manufacturing industry, an enterprise that would cut American labor costs, boost productivity and rack up as much as $4 billion in sales by 1990. Blue-chip giants stampeded to buy into the action; bankers panted to finance the heralded expansion. Optimism was seemingly unbounded for the U.S. robotics industry, which produced semi-intelligent machines that were expected to help American businesses compete with low-wage foreign rivals over the next two decades and to improve greatly the quality of American industrial production.

Well, that was five years ago. Rather than becoming the highly successful purveyor of tireless, reliable welders, assemblers and heavy lifters for the auto industry, aerospace and other industrial concerns, robotics today is an industrial accident victim, crippled by a two-year slump. Sales of U.S. robots are expected to decline from an anemic $580 million in 1986 to about $400 million this year, miles below those rosy billion-dollar projections. The number of manufacturers that make robots and related equipment dropped from 328 last year to 300 this year.

Analysts had predicted that 250,000 robots would be in American factories by 1990; today only 25,000 are installed, roughly twice as many as exist in West Germany, which has a much smaller industrial base. The U.S. lags far behind Japan, where 118,800 robots are in use. Along with sluggish domestic demand, U.S. manufacturers face a shrinking share of the roughly $1.9 billion global robotics market. Reason: Japanese competitors have gained a strong edge in the field and appear likely to continue their domination. Says Michael Cronin, president of Automatix, a Massachusetts-based robotmaker that lost $7.8 million last year: "It's a bloodbath out there."

A symbolically significant retrenchment took place last month, when Westinghouse Electric sold off part of its money-losing Unimation robotics division. The buyer: Prab Robots, a small Michigan-based manufacturer of industrial robots and conveyor machines. Westinghouse's 1983 purchase of Unimation for $107 million marked Big Business's arrival in robotics; IBM, Bendix and General Electric soon followed. Unimation, founded in 1959, was a robotics pioneer. Its first product was an $18,000 Unimate machine used by General Motors to load forged dies at a New Jersey auto-assembly plant. As recently as 1981, Unimation made 45% of all robots sold in the U.S. Another setback for robotics will take place next month, when GE plans to fold its $4 million robotmaking plant in Plymouth, Fla., idling 118 workers.

There is a measure of irony in the robot industry's plight. Although industrial robots account for only 2% of the $24 billion factory-automation business (such items as computers and other electronically controlled industrial machinery make up much of the rest), the mechanical menials have drastically altered many sectors of the American workplace. Robots perform more than 98% of the spot welding on Ford's highly successful Taurus and Sable cars. At Doehler-Jarvis, a major Ohio metal fabricator, robots load and unload die-casting machines, trim parts and ladle molten metal. At IBM factories across the country, robots insert disk drives into personal computers and snap keys onto electronic typewriter keyboards. At a General Dynamics plant in Fort Worth, one robot drills 550 holes in the vertical tail fins of an F-16 fighter in three hours. It used to take three workers eight hours to do the same job.

What led to the current U.S. debacle? One factor was a slowing in capital spending that began with last year's elimination of the investment tax credit, making it more expensive for companies to buy big-ticket items like robots. Beyond that, the technology was often overhyped. Robots also proved more expensive to operate than many manufacturers imagined. U.S. robotmakers depended heavily on the fortunes of a single industry: automaking. U.S. auto manufacturers have bought 50% of American robots in current use. By contrast, less than 10% of Japan's robots are operated by its auto firms.

As the domestic automakers' share of the American market has declined, so has the need for robots. GM alone accounts for 40% of U.S. demand for the devices; last year, when the No. 1 automaker killed a major program intended to build plastic-bodied cars, it canceled about $100 million in orders for robots and support equipment. That was bad news for GMF Robotics of Troy, Mich., the nation's biggest robotmaker (1986 sales: $186 million). GMF, a joint venture of GM and Fanuc, Japan's largest robotics firm, has cut its work force to 400 people, 60% of what it was two years ago.

Another area of vulnerability for U.S. manufacturers was the hydraulic- robot technology pioneered by Unimation. The company's robots, which became the American industry standard, were large (up to 4,000 lbs.), powerful, multipurpose and expensive, ranging in price from $30,000 to $200,000 apiece. But these bulky hydraulic machines, originally programmed to perform tasks by means of magnetic tape similar to that used in tape recorders, were often inaccurate and susceptible to breakdowns. Says Raj Reddy, director of the Robotics Institute at Pittsburgh's Carnegie-Mellon University: "U.S. companies dragged their feet on innovation because they wanted to squeeze every last penny out of their existing equipment." Despite those drawbacks, in the early 1980s hydraulic robots appeared to be the best workhorses available for such automated tasks as parts assembly, materials handling and heavy-duty lifting.

Even as U.S. robotmakers wallowed in their early success, the Japanese, who imported their first hydraulic robot in 1967, were coming up with a new product. Fitted with high-speed computer chips and sophisticated circuitry, the new electric machines received instructions via computer-software programs. The machines tended to be smaller, less expensive ($5,000 to $40,000 each) and not as prone to breakdowns as their U.S. hydraulic counterparts. - Though electric robots were less powerful, and thus less capable of heavy industrial tasks, their greater accuracy in tasks such as delicate manipulation and precision welding made them more attractive for the automotive, aerospace and electronics industries.

American manufacturers were soon suffering as foreign-built electronic robots stormed the U.S. market. At Unimation, sales dropped from $70 million in 1983 -- the year Westinghouse bought the firm -- to $45 million in 1985, then to $38 million last year. Overall, Japanese robotmakers account for 28% of all sales in the U.S. Worldwide, Japanese sales exceed those of U.S. competitors by 30%.

Exaggerated claims about what U.S. robots could do for businesses proved to be just as debilitating. Says Laura Conigliaro, analyst with the Prudential Bache investment firm: "The robot industry promised more than it could deliver. The technology was not as advanced or sophisticated as promised." Many companies discovered that buying an industrial robot was only the first, and least expensive, step in automating their factories. Says Carnegie- Mellon's Reddy: "Suddenly they needed experts in computer science, communications and database technology. The number of people in factories with this expertise is probably zero." Adds Warren Seering, professor of mechanical engineering at M.I.T.: "In general, robots are much more expensive than people."

Some of those who rushed to buy an expensive robotic system got less than they bargained for. At a Ford Motor plant in St. Louis, snags in 200 production-line robots delayed the 1986 introduction of the Aerostar minivan. Then the discovery that the same robots had been skipping many key welds led to the recall three months later of some 30,000 of the vehicles. In another disastrous episode, a Campbell Soup plant in Napoleon, Ohio, was outfitted with a $215,000 system designed to lift 50-lb. cases of soup. But anytime it encountered defective cases, the machine would drop them, causing thousands of dollars in damage. Eventually the robot was donated to a local university and replaced by three humans. Says Warren Helmer, the company's manager of engineering research and development: "Campbell's was ready for robots, but robots weren't ready for Campbell's."

The current slump does not necessarily signal the demise of American robotics. The industry is expected to perk up again by the end of 1988, partly because of increases in U.S. competitiveness caused by the falling dollar. Struggling American manufacturers have begun to adopt the electronic robot technologies of the Japanese and, like U.S. automakers, are moving their own assembly plants overseas to help cut costs. Above all, U.S. robotmakers have adjusted their own expectations of how the industry will perform in the future. "We're in a solid business with solid growth," says Bruce Haupt, a marketing manager in the division that oversees robotmaking at IBM. "Our early expectations were out of line, but they have been altered."

Even so, most analysts expect the number of U.S. robotmakers to keep shrinking through the mid-1990s. By that time robotics technology may have taken another impressive leap forward, with the U.S. once again expected to be the technological trailblazer. Advances now being explored in American universities and research laboratories could lead to the creation of machines capable of walking, improvising tasks and seeing (some robots can already do this crudely, through computerized video cameras). By then, the robots' masters may have learned how to exploit their wondrous inventions without falling into the kind of painful doldrums that now afflict their once glamorous industry.