Google's purchase of Boston Dynamics signals move from hobby to industry.
Has robotics’ time finally arrived? With global sales up a mere 2% last year, according to the International Federation of Robotics (IFR), that may be a tricky conclusion to draw. However, with internet giant Google’s acquisition of Boston Dynamics, robotics time in the sun is certainly getting closer.
Spun out from US-based Massachusetts Institute of Technology (MIT) in 1992 based on research by Marc Raibert, Boston Dynamics has grabbed headlines and attracted millions of YouTube views for its animal-inspired robotics capable of moving at speeds and crossing tough terrain. The firm has also received nine awards from the US government’s Small Business Innovation Research (SBIR) programme to develop its growing catalogue of robots, such as a robot which can paint radio towers and a robotic mule to carry equipment.
Bolstered by funding from the US’ Defence Advanced Research Projects Agency (DARPA), Dynamics also draws on its association with MIT, being one of thousands of spin-outs to have passed through the institute. MIT produces more patent applications than any other single university in the world, with 179 in 2011, and its active companies created by its alumni bring in aggregate revenue of about $2 trillion – the equivalent of the 11th-largest economy in the world.
However, despite these strong connections, robotics as a whole sector has received relatively little venture capital. Instead, most of the research has been driven by established corporations, such as Honda, as well as government-funded research at universities and laboratories.
One company overlooking the current sales forecast and dearth of venture investment is Google, which has been following and buying some of the best robotics firms on the market (see box below). Earlier, Google, according to research by news provider New York Times, reportedly bought seven other robotics companies this year, including Japan-based Schaft. Yuto Nakanish, chief executive of Schaft, had been working on similar problems since the 1980s after spinning out from Tokyo University. Google’s other acquisitions in robotics are notable for having no public venture capital backers, instead seeming to rely on public grants and commercial contracts.
Blog Hizook’s list of VC Funding for Robotics in 2012 showed robotics companies raised about $220m (emitting deals under $1m), such as Anki’s $12m series A round and Finland-based ZenRobotics’ €13m ($17m). It was a similar amount to the year before, but in total it is still less than one single deal can raise in consumer or enterprise internet companies, such as Google-backed Surveymonkey.
Instead, Hizook noted “innumerable crowd-funded robotics campaigns launched new companies; and robotics-specific grants to academia seemed to be on the up, for example, US government’s National Science Foundation’s (NSF) National Robotics Initiative (NRI) and Defense Advanced Research Projects Agency’s (Darpa) Maximum Mobility and Manipulation (M3)] and ARM Humanoid programs”.
The slow rise of robotics is reminiscent of the early years of personal computing. As the power of computing hardware increased, Moore’s Law, named after Intel co-founder Gordon Moore, stated in 1965 that the number of transistors on integrated circuits doubles approximately every two years. Then, discussion at the Stanford Research Institute (subsequently spun off in 1970 from the eponymous university and renamed SRI International) under Doug Engelbart and Stanford Artificial Intelligence Laboratory (Sail) under John McCarthy turned on whether computing would augment humans or create an artificial intelligence that could go beyond Moore’s forecast.
With developments such as quantum computing power, augmentations and prosthetics, and machine-learning algorithms on the horizon, this debate has been re-opened, as has a conversation about which groups will reap the rewards on the investment into robotics.
But just as taxpayers, via government agencies, were replaced as the primary source of money funding for personal computing and sparking what venture capitalist and former Intel executive John Doerr called the “largest legal creation of wealth on the planet”, so robotics is moving outside of the labs as corporate venturing units, angel investors and VC firms become more active. This move to alternate funding was a topic of discussion led by Robert Bosch Venture Capital, the corporate venturing unit of the eponymous Germany-based industrial company, at this year’s Global Corporate Venturing Symposium, which sponsored the robotics and industrials sector roundtable.
An example of robotics’ rising backing comes from Dmitry Grishin, one of the co-founders of Russia-based technology company Mail.Ru, who set up Grishin Robotics, which it said was the first global investment company focused on the robotics industry. More information on Grishin can be found over on our sister publication.
Google Ventures followed its parent’s lead on robotics by beefing up its staff in the area (although admittedly it is adding everywhere judging by its 2013 report). Last month, Google’s eponymous corporate venturing unit hired Shanna Tellerman and David Munichiello as investment partners as it backed Building Robotics, a US-based creator of software that allows people to control office temperature, with $1.1m seed funding.
Tellerman was the founder of Sim Ops Studios (Wild Pockets), a spin-off from Carnegie Mellon University that focused on democratizing 3D game development, and which was acquired by Autodesk in 2010. Carnegie Mellon’s Community Robotics, Education, and Technology Empowerment Lab has partnered with Google on a number of mapping projects that use the university’s GigaPan technology.
Munichiello joined Google Ventures from being senior director of solutions design and professional services at Kiva Systems, a logistics robotics company sold to online retailer Amazon for $775m. The Kiva sale last year was followed by Japan-based conglomerate Softbank buying 80% of Aldebaran, a France-based robot maker backed by Intel Capital, the corporate venturing unit of the eponymous chipmaker, for more than $100m.
Both exits sparked renewed interest in robotics as signs of rapid growth are starting to emerge. For the period 2013 to 2016, IFR forecasted a near-75% increase to about 94,800 units (primarily for the military and milking) with a value of $17.1bn. This would be an increase from the 126,000-plus service robots for professional use counted since 1998, even if IFR said the total number of professional service robots sold in 2012 “rose by a relatively low 2%” to 16,067 units (worth $3.4bn), up from 15,776 in 2011.
The market will increase substantially within the next 20 years, according to IFR. Sales of all types of robots for domestic tasks (vacuum cleaning, lawn-mowing, window cleaning and other types) could reach almost 15.5 million units in the period 2013-2016, with an estimated value of $5.6bn. Toy robots are due to add a further 3.5 million units, with an additional 3 million robots for education and research and about 6,400 robots for elderly and handicap assistance.
The bigger picture than headline unit sales, however, is the data and information. As Andy Rubin said in his New York Times interview: “I feel with robotics it’s a green field. We’re building hardware, we’re building software. We’re building systems, so one team will be able to understand the whole stack.”
Rubin, founder of what is now Google’s Android mobile operating system, has been looking for access into robotics technologies that could potentially help it develop the equivalent of its search engine, maps and Android systems – free to use and build upon but where Google can understand the data and uses of the system. The move into robotics also offers insights into the so-called internet of things, which currently usually relies on Bluetooth connections between a developer and an application but is reportedly struggling to find its standards to allow a wider communication systems – or roboweb.
Google itself has not yet clarified its overall strategy on robotics. However, it has ruled out further development of Dynamics’ product portfolio for the military use. Having said that, Google will still be obliged to fulfil existing contracts between Dynamics and DARPA, making the company with the motto “do no evil” an arms manufacturer for the time being.
Yet, a place for robotics can be envisaged within Google. For a company wanting to understand all information, getting into and shaping a world where there are potentially more robots than people and machine-to-machine communication is important.
Last year, news provider Forbes said there were about 8.7 billion devices connected to the internet. Next year, consultant Frost & Sullivan said machines would create more data and traffic than between humans. Given the rise of robotics on the horizon, perhaps it will be only a few years after that that Google’s shareholders choose to replace its human chief executive with a machine-learning robot the company has designed.
Box: Backgrounds to Google’s robotics purchases
A nice Upstart piece looks at what was known about the founders behind the eight robotics companies acquired by Google this year but there are some ties between most of them beyond their new owner, even if none of them have apparently filed funding documents with the Securities and Exchange Commission.
Marc Raibert founded Boston Dynamics after leading research laboratories at Carnegie Mellon University (CMU) and Massachusetts Institute of Technology (MIT) dedicated to building walking machines, including BigDog, Cheetah, Petman and Atlas.
Atlas competed in the US military’s DARPA-sponsored robotics competition along with Google-acquired Schaft, a team of Japanese roboticists who spun out of Tokyo University’s Jouhou System Kougaku (JSK) Robotics Lab.
Another, more recent, MIT spin-off acquired by Google is Meka Robotics, which was founded in 2006 by Aaron Edsinger and Jeff Weber. In 2012, the company teamed with the Human Centered Robotics Group at the University of Texas at Austin to develop HUME: a “bipedal robot for human-centred hyper-agility”.
Meka, along with research labs Willow Garage and SRI International, co-founded robotic arm maker Redwood Robotics, although the company is probably best known by its open source software suite ROS (Robot Operating System). Scott Hassan, an early Google employee, started Redwood in late 2006 and has now returned to the search engine provider after Google bought Redwood this year.
Other Willow spin-offs acquired by Google are Industrial Perception, which makes robotic three-dimension vision, and Holomni, a robot design firm apparently run by Bob Holmberg.
While the final reported Google acquisition is robotic camera system maker Bot & Dolly, which was co-founded by ex-Autofuss executives Jeff Linnell and Randall Stowell.