For many organizations, the biggest reason for not replacing manual labor with robots is purely economic. We believe that most organizations begin to ramp up their investment in automation when the cost of employing human labor rises high enough above the cost of owning and operating robotics systems to make human labor less cost-effective. We have assumed conservatively that this point is reached when the cost of human labor becomes 15 percent higher than the cost of robotics labor. Yet even with the relentless rise in wages around the world, the gap remains wide enough to prevent mass deployment.
The economics of advanced robotics are improving rapidly, however. For example, the total cost of purchasing and deploying a robotics system for spot welding in the U.S. automotive industry plunged from an average of $182,000 in 2005 to $133,000 in 2014 (not adjusted for inflation). By 2025, the total cost is projected to drop by approximately another 22 percent, to around $103,000. The prices of robotics hardware and software, which account for only one-quarter of that total cost, are around 40 percent lower than they were a decade ago. The cost of systems engineering—which includes installing, programming, and integrating a robotics system into a factory—has declined even more. Ten years ago, the average systems-engineering costs of a spot-welding robot amounted to $81,000. Those costs are now down to around $46,000, on average, and are likely to keep dropping for the rest of the decade. The costs of peripheral equipment—such as sensors, displays, and expensive safety structures that protect workers and that together typically cost more than the robots themselves—are plunging as well. (See Exhibit 2.) In fact, safety barriers may not be required at all for many next-generation robots.
At the same time that costs have been declining, the performance of robotics systems has been improving by around 5 percent per year. Taken together, the changes in price and performance for spot welding, for example, have been translating into an annual 8 percent improvement in the cost of robotics. To put this into perspective: an investment of $100,000 today buys a robotics system that is capable of performing more than twice as much work as a robotics system costing the same amount a decade ago. This pace of improvement in price and performance is expected to be sustained for the foreseeable future.
Robotics systems are thus becoming an economically viable alternative to human labor in more and more industries. A human welder today earns around $25 per hour (including benefits), while the equivalent operating cost per hour for a robot is around $8 when installation, maintenance, and the operating costs of all hardware, software, and peripherals are amortized over a five-year depreciation period. In 15 years, that gap will widen even more dramatically. The operating cost per hour for a robot doing similar welding tasks could plunge to as little as $2 when improvements in its performance are factored in. (See Exhibit 3.)
Other industries are quickly approaching inflection points. The U.S. electronics and electrical-equipment manufacturing industries currently deploy about 3,300 industrial robots, many of them relatively basic and designed to perform simple tasks. But manufacturers are in the process of adding more versatile—and more expensive—robots to take on increasingly complex tasks. Today, the cost of a “generic” robotics system—which has a high degree of flexibility and thus can take on many different types of work—is around $28 per hour. By 2020, this cost is projected to fall to less than $20 per hour, which would be below the average human worker’s wage. This will enable a significant increase in the number of tasks that can be automated. We estimate that, as a result, the percentage of tasks handled by advanced robots will rise from 8 percent today to 26 percent by the end of the decade.
Some industries will be slower to adopt. In furniture manufacturing, where tasks remain more difficult to automate, the economic payoff of using robots is still a number of years away. Based on our estimates, the adoption of robots in that industry won’t begin in earnest until 2020; and it will be near the end of the next decade before even 10 percent of tasks are automated.