Andy Rubin, father of the Android OS, revealed on Wednesday to the N.Y. Times his next "moon shot" project for Google (GOOG): androids. Pursuing an old avocation, Rubin wants to build robots with human-like capabilities that the current generation of industrial robots can't match. The payoff for Google could be lower cost domestic manufacturing, packaging and shipping of devices such as smartphones.
Rubin's Next Moon Shot
The reason for the lower cost is that final assembly of mobile devices and smartphones still requires a human being with manual dexterity and intelligence to put together the component parts. Likewise, packaging for shipment, warehousing and pulling items out of inventory usually requires a human being. Rubin thinks robotics technology has matured to the point that these jobs could be filled by next generation robots.
However, it won't be easy to match the cost effectiveness of the contract manufacturers such as Foxconn that have armies of highly skilled assemblers who get paid about $2.00 an hour, according to analysis by the N.Y. Times. The challenge for manufacturers in the United States is also about skills. A large labor force with the skills required for fine assembly work just isn't available in the U.S.
Developing robots that can compete on cost and quality with the likes of Foxconn is a tall order. Why does Google want to go there? There are a number of reasons:
1) Clearly Google wants to become a vertically integrated devices/software/services company in the Apple (AAPL), and now Microsoft (MSFT), vein. This was the justification for the Motorola acquisition. And while Google's Android has overwhelmed iOS in market and user share, Apple is the acknowledged leader in mobile device profitability. As described in a recent Forbes article, while iOS devices made up only 12.9% of mobile devices shipped in Q3, Apple captured 56% of the profits.
With Chromebooks, Nexus phones and tablets, as well as Motorola devices, Google has been broadening its portfolio of hardware offerings. I no longer think of Google as just a software company.
2) Although most of Google's products are still made abroad, Google is thinking hard about the benefits of repatriating some of that work, as has Apple. Foxconn recently announced a pay increase for workers, so labor rates may not be favorable indefinitely. Combined with higher transportation costs, driven by ever higher energy costs, domestic manufacturing becomes more attractive.
The Moto X experience has probably been sobering for Google. Touted as a premium, assembled in the U.S. smartphone intended to restore Motorola's preeminence, the phone has been a disappointment. It was widely reported to have only sold about 500,000 units in Q3, and is now on sale on the AT&T site for a bargain basement $50 with contract. IHS estimated the cost to manufacture the phone (components plus assembly) to be $226, so it's unlikely that Google is making money on the phone.
Although the assembly cost is not a large part of the overall cost of the phone, it was significantly higher than the IHS estimate for the Apple iPhone 5s at $12 vs. $8, pointing out the existing labor cost disadvantage of assembly in the U.S.
Even more sobering, early reviews (such as Engadget's) of the phone pointed to cheap construction and product defects, so Google's domestic manufacturing suffered a quality disadvantage as well, relative to foreign manufacturing.
Google appears to be serious about remedying both disadvantages through its robotics initiatives. They've put arguably their best technologist (Rubin) on the project and shelled out for a number of robotics technology start ups. The crown jewels among the acquisitions were Meka and Redwood Robotics, both startups co-founded by Dr. Aaron Edsinger, who studied under Rodney Brooks at MIT, who also is CEO of his own robotics company, Rethink Robotics. With Edsinger, Google gets someone with direct research experience on the problem of robots with human-like motor skills.
Apple's More Cautious Approach
While Google seeks a leap-ahead breakthrough in robotics, Apple has taken a characteristically more cautious and conservative approach to repatriating manufacturing through its new, assembled in the U.S. Mac Pro. Apple has taken on an easier assembly problem in its new desktop Mac, and still uses people for the minimal final assembly it requires, but just about everything else in the new Mac is manufactured by automation. Clearly, when Apple execs wandered around the Foxconn plants, it wasn't just for photo-ops. They were checking it out, and decided that for some things, they could do better.
But Apple's motivations are basically the same as Google's: concern about rising labor and energy costs that will eventually remove the advantage of manufacturing in the Far East, as well as ethical concerns about labor exploitation and environmental impacts.
Will Rubin's initiative pan out? He has a very impressive track record of accomplishment and impeccable timing. I think his timing in this case is right on. Robots such as the Baxter of Rethink Robotics already have manual dexterity approaching that of people, but the real problem is integrating sensors, mechanisms and intelligence into a machine that can assemble a Moto X or Nexus 5.
Google is already gaining much of that systems integration experience through its autonomous car project, which poses a similar set of sensor/mechanism/intelligence integration problems. With its recent acquisitions, Google probably has all the expertise required to solve those problems, if they can be solved.
As Rubin related to the N.Y. Times, "I feel with robotics it's a green field," he said. "We're building hardware, we're building software. We're building systems, so one team will be able to understand the whole stack."