Over the past few years rapid technology advancements and related cost reductions have brought robots (in the broad sense of autonomous systems) to the forefront of news on disruptive industrial technologies, together with the Industrial Internet of Things, Big Data Analytics and 3D-Printing. As the impact of robots on industry, the economy and society at large will be profound, the debate on its implications will intensify. Social scientists are currently trying to understand how rapidly increasing robot deployment will affect employment and social governance, while robot manufacturers, users and even governments are at pains to portray the technology as something that will augment or support human activity rather than replace it. And even if it does replace some tasks, even jobs, they argue, more net jobs in other fields will be created –as has happened throughout history following the industrial revolution.
Of course it is not difficult to see that this argument is not watertight. Robots are different than other forms of technology in that they can “think” (whatever that may mean). Advances in sensing technology, dexterity, algorithms and machine learning mean that they can now take on tasks previously thought to be exclusively in the human domain. We will not engage in this debate here, interesting as it may be, other than by supplying a number of links and reading recommendations at the end of this article. One thing however is important to understand and should be uncontroversial: Mass deployment of robots (and other new technologies) is crucial and necessary (though possibly not sufficient in itself) to drive productivity in the face of adverse demographics and aging societies. Let’s not forget that productivity growth is largely responsible for humanity leaving behind the Malthusian trap and continuously improving living standards since the dawn of the industrial revolution. This enabled inclusive and broad participation in economic activity by the vast majority of people and hence a continuous increase in consumption, which in turn fueled investment and on-going economic growth in a virtuous cycle (the environmental impact notwithstanding). If a disruptive technology now shut out of economic participation vast numbers of people (as unnecessary), consumption would drop and economic growth would collapse. This alone makes a radical rethink of how economic activity is understood and governed necessary and it needs to happen quickly -before the impact takes too high a toll.
This briefing will consist of three parts: In the first part we’ll examine how the robotics industry is evolving and is itself being disrupted. In the second part we’ll take a look at how mass robot deployment will affect manufacturing and at robots as potential “service platforms”; And in the third we’ll look at robots as service providers outside of manufacturing.
Part 1: Understanding robot disruption
In another article we discussed from both theoretical (Clayton Christensen and others) and practical angles the topics of disruption and servitization and in a further one on digitization, we argued that disruptive technologies can be identified, among other things, by very rapid growth rates over a period of time. If we look at sales of conventional industrial robots, they have been growing strongly, but not at rates which could be deemed disruptive. According to the International Federation of Robotics (IFR), global sales of industrial robots increased by 8% to almost 240,000 units in 2015 –an all-time high for the third year in a row. However most of the growth is coming from the Chinese automotive and electronics industries, where rapidly rising production, adverse demographics and increasing labor costs are causing a surge in deployment.
Source: IFR Statistics Dept.
Robot density in China is low with about 36 robots per 10,000 manufacturing workers (2014 figures) compared with 164 in the US, 292 in Germany, 314 in Japan and a whopping 478 in South Korea. The average global robot density is approximately 66 units. The automotive industry has by far the highest rate of robot adoption, with density in Japan at 1,414 industrial robots per 10,000 employees, followed by Germany with 1,149, the United States with 1,141 and South Korea with 1,129 units. In China, robot density in the automotive industry has increased considerably in the last several years, but is still moderate at 305 units per 10,000 employees. As production and labor costs increase (Chinese auto production accounts for approximately 30% of global production), payback periods for robotic installations are reducing and return on investment (ROI) is increasing. Robot shipments into automotive are therefore increasing very rapidly and will continue to do so.
In fact, given its size, the robotics industry overall and its key products have primarily evolved and are configured to support the automotive industry and its particular requirements in the global competitive context. Today automotive accounts for 40% of global shipments, electronics (also due to size) for 23%, with other industries far behind. This means that due to limitations of the technology, cost and the effort required to structure the environment to suit the tool, traditional industrial robots have penetrated only an estimated 10% of total manufacturing industry.
For comparison purposes, robotics density in manufacturing, including electronics, but excluding automotive is 365 robots per 10,000 employees in South Korea, followed by Japan with 211, Germany with 161, Sweden with 142 and Taiwan with 138 units. All other countries have lower robot densities, while emerging markets have robot densities significantly below 30, including Russia, Brazil and India.
The rate of robot deployment is a function of expected investment payback or ROI. In China this has primarily been driven by rapidly rising labor costs and declining availability of labor. So for example Foxconn, which assembles the iPhone for Apple and is China’s largest private employer, announced in 2011 that it would deploy up to 1 million robots over a 3-year period to replace workers. Five years later, deployment is still a fraction of what was planned, plagued by premature assumptions regarding robotic capabilities. Nevertheless the calculations were based on the fact that Chinese wages had increased by 500% since 2000 and an expectation of 18% p.a. increases for the foreseeable future. Against this, robot costs had remained steady or were declining. Newer calculations by Barclays Equity Research published in the Financial Times (FT), show that the trend in expected returns is, in fact, not widely off the mark. Wages growth rate has averaged a more modest 12% p.a. over the past 5 years, however robot costs, have been declining faster than expected (mainly because of scale, local manufacturing as well as technological change). As a consequence, the payback period for a robot investment has declined from 5.3 years in 2010 to 1.5 years in 2016, a very rapid change which is a first indication of almost disruptive acceleration, resulting in a tripling of shipments in three years .
Source: FT
For a good take on how robots are starting to revolutionize Chinese manufacturing and how they are affecting industry in emerging markets, the Financial Times has published an interesting series of articles.
Real disruption however in the robotics industry will not come from conventional industrial robots, but from new concepts and underlying technology, which are both reducing cost and enabling deployment of autonomous systems in a far wider range of applications, including – in addition to manufacturing- healthcare, agriculture, defense and security, energy, various consumer industries and services, transport and mobility, materials handling, industrial and infrastructure maintenance, construction, firefighting, pollution control, environmental protection and wildlife conservation, mapping and surveying and many more. Apart from recognizable “robots”, technology platforms include all types of drones, vehicles and other specialized machines.
In the manufacturing space, collaborative robots or “cobots” are disrupting the conventional robot paradigm: Cobots were invented by Northwestern University professors J. Edward Colgate and Michael Peshkin in 1996 (see Peshkin’s page at Nothwestern for a history and technological evolution). They have since been commercialized in various forms and the Robotics Industries Association (RIA) defines Cobots as follows:
Collaborative Robot – A robot designed for direct interaction with a human within a defined collaborative workspace.
Collaborative Workspace – A safeguarded space where the robot and a human can perform tasks simultaneously during production operation.
Collaborative Operation (Human-Robot Interaction or HRI) – A state in which purposely designed robots work in direct cooperation with a human within a defined workspace.
RIA has also published a primer on Cobots, including a description of the various models on the market (2013) and updated its review focusing on technological aspects a year later.
The main conceptual features of cobots, in contrast to conventional industrial robots, is firstly the trade-off between accuracy/precision and speed (lower) for dexterity and flexibility (higher) and therefore the ability to perform more tasks, i.e. cobots are “generalists” rather than “specialists” and, secondly (essentially a condition for the first feature), their ability to work safely alongside humans. They are therefore built with integrated multiple sensors and passive compliance or overcurrent detection as safety features and a designed-in ability to interact with humans -since no additional safety features are required (fences, switches, etc.).
They are “force limited” in the sense that the integrated sensors will feel external forces and, if this force is too high, the robot will stop its movement. Passive compliance is produced by mechanical components. If an external force acts on a joint, this joint will submit itself to this force. So, in case of a collision, the joint will move in the opposite direction avoiding any injury. Also, an overcurrent can be detected when a collision occurs. This is another safety feature, because the software can generate a security stop when it detects a current spike.
Taking advantage of advances in software, including machine learning, and to make human-robot-interaction meaningful, some cobots have been designed to be “taught” by demonstration rather than requiring significant programming. In this way they can be implemented easily, brought on-line fast and optimized along the way. To fulfill their generalist nature, the majority of collaborative robots can also be moved around the factory floor with ease in order to be set-up to perform another task at another station.
The new concept combined with new technology is enabling cobots to significantly bring down the cost of robotics applications. The combination of cost reductions in off-the shelf hardware and sensing technology, algorithms -including through standardization and open source middleware, such as Robot Operating System (ROS)- and advances in processing power, connectivity, machine vision, dexterity and other areas are resulting in a projected cost curve trajectory that has disruptive characteristics. Note also that for safety and mobility purposes, cobots must be significantly lighter than their conventional counterparts, which further reduces cost. They pay a price for that in accuracy and speed, but for the applications they are intended for , they make up for it through more task abilities and “intelligence”.
Comparison of Cobot UR10 by Universal Robots and Industrial Robot M-10A by Fanuc. Source: Robotiq Blog
Cobots in general have smaller payloads, slower speeds and lower accelerations. This means that the different structural components and motors can be smaller than a standard industrial robot. By having a frame that is lighter, the required electric motors are smaller and cheaper than heavy duty industrial strength robot motors.
Since a large part of the collaborative robot has to be easily movable by hand, they have to be made out of light weight materials. This is a big difference from industrial robots with the same payload. In fact, for the same payload, a collaborative robot will always be lighter than an industrial robot. This means that less material is used to manufacture the robot, which results in a cheaper robot.
However the most significant cost reductions come from operations. Configuring the environment, programming, operating and otherwise supporting a conventional industrial robot can cost up to 5-fold its original purchase price. Cobots do not require the safety systems industrial robots do, nor do they need the environment to be configured to their operating process. Rather they conform to the existing operating environment. They need far lighter or no programming, and can be “trained” by the end-user, empowering operators on the shop floor: Essentially the person using the tool is able to shape it to his/her own requirements and way of working. Their operating costs are therefore drastically lower than those of conventional industrial robots. Combined with ease of deployment and use, this allows cobots to put the power of robotics within reach of low-volume manufacturers for the first time, radically expanding the market in a way reminiscent of when personal computers replaced mainframes.
Cobots are enabling robotics for the first time to be integrated in new areas within the automotive and electronics industries and to penetrate industries and processes that had been considered “robot proof”, where robot density has been low due to the adverse economics of traditional robotics applications. This applies to both large and small companies. The 90% of manufacturing industry that traditionally has been out of scope of robotics is now a very legitimate target. And rapidly evolving technology, such as machine learning and dexterity capabilities will compound the effect. The Boston Consulting Group (BCG) therefore forecasts that the percentage of tasks handled by advanced robots will rise from 8 per cent today to 26 per cent by the end of the decade, driven by China, Germany, Japan, South Korea and the US, which together will account for 80 per cent of robot shipments. As Hal Sirkin, a consultant at BCG’s Chicago office told the FT, the rapid expansion of automation could be compared to the difference between the “human learning curve” and Moore’s Law, which posited that computing power could double every 18 months to two years. “Even if you’re very good, humans can only double their productivity at best every 10 years. In contrast, researchers can push robots to double their productivity every four years”, he estimates. “Compounded over time, that makes a big difference”.
Currently, the technology being very new, cobot shipments account for just 5% of global industrial robot sales. However Barclays Equity Research and other analysts estimate that the cobot market could grow from just over $100m last year to $3bn by 2020, a CAGR of over 75%, which is a highly disruptive rate.
Barclays base their forecasts on current (2015) global sales of 4,100 units and an average selling price of $28,177 per unit; declining to a forecast $21,000 per unit in 2020, with annual unit sales of 150,000, thereby estimating a global market of $3.1 billion in that year. They further estimate unit prices continuing to decline by 3-5% per annum through 2025 to circa $17,500 per unit, and global sales of a staggering 701,000 cobot units in 2025 when they forecast a market size of around $12 billion. Source: Robotenomics.com
The industrial robot market is dominated by four major suppliers (ABB, Fanuc, Yaskawa and Kuka), while the two major disruptors at present appear to be Universal Robots of Denmark (acquired by Teradyne in 2015) and Rethink Robotics of Boston, MA, developer of the famous Baxter and Sawyer cobots – though a number of other start-ups and other players have also entered the market. Their technology, while it has similarities, is also conceptually different in important ways, including in the emphasis placed on the adaptability of the cobots. The incumbent suppliers have now all launched their own cobot products. In fact ABB acquired Gomtec in 2015, to expand its cobot portfolio. Another company making a cobot entry is Bosch, which became the first company, whose system, called APAS, was certified as an assistance system, i.e., safe to work alongside with human workers. As cobots improve technologically and even begin to compete with conventional industrial robots, it is an open question whether incumbents will prove sufficiently committed to the technology to allow it to cannibalize and gradually replace their traditional products or whether they will succumb to disruption “from below” (in the sense of Christensen).
An interesting account of the market potential and current cobot models can be found at: Why Co-Bots will be a huge innovation and growth driver for robotics industry – IEEE Spectrum Magazine 2015
In any case investors are giving cobots a vote of confidence: According to the FT, almost US $ 600 million in funding was raised for cobot start-ups in 2015, an increase of over 150% over 2014 (almost $1 billion for the broader robot category, incl. non-industrial robots). Patent filings covering robotics technology — one sign of the expected impact — have soared. According to IFI Claims, a patent research company, annual filings have tripled over the past decade.
And the robotics disruption or, better, explosion is projected to continue indefinitely. According to the FT quoting Steve Jurvetson, a Silicon Valley investor and a director at Elon Musks’s Tesla and Space X companies, which have relied heavily on robotics:
“There isn’t a single mechanical or physical thing a human will be able to do better than a robot… Everything gets better over time; This is happening in almost every hardware product: They are becoming minimal vessels for software.”
Recommended Reading:
Better than Human: Why robots must and will take our jobs – Wired 2012
A world where everyone has a robot: Why 2040 could blow your mind – Nature 2016
New approach trains robots to match human dexterity and speed – New York Times 2015
Google is using machine learning to teach robots how to grasp random objects – Techcrunch 2016
Terradyne CEO on Universal Robots Acquisition – Robotics Business Review 2015
Seven Robots You Need to Know – Financial Times 2016
Robot to sit the national college entrance exam in 2017 – China Daily 2016
The future of employment: How susceptible are jobs to computerization? A paper by Carl Benedikt Frey and Michael A. Osborne, both University of Oxford 2013
Rise of the Robots: Technology and the Threat of Mass Unemployment: A book by Martin Ford, Financial Times and McKinsey Business Book of the Year Award 2015
Publisher’s summary:
Intelligent algorithms are already well on their way to making white collar jobs obsolete: travel agents, data-analysts, and paralegals are currently in the firing line. In the near future, doctors, taxi-drivers and ironically even computer programmers are poised to be replaced by ‘robots’. Without a radical reassessment of our economic and political structures, we risk the very implosion of the capitalist economy itself. In The Rise of the Robots, technology expert Martin Ford systematically outlines the achievements of artificial intelligence and uses a wealth of economic data to illustrate the terrifying societal implications. From health and education to finance and technology, his warning is stark – all jobs that are on some level routine are likely to eventually be automated, resulting in the death of traditional careers and a hollowed-out middle class. The robots are coming and we have to decide – now – whether the future will bring prosperity or catastrophe.
Race against the Machine: A book by Erik Brynjolfsson and Andrew McAffee, both MIT – 2011
Excerpt:
We wrote this book because we believe that digital technologies are one of the most important driving forces in the economy today. They’re transforming the world of work and are key drivers of productivity and growth. Yet their impact on employment is not well understood, and definitely not fully appreciated. When people talk about jobs in America today, they talk about cyclicality, outsourcing and off-shoring, taxes and regulation, and the wisdom and efficacy of different kinds of stimulus. We don’t doubt the importance of all these factors. The economy is a complex, multifaceted entity.
But there has been relatively little talk about role of acceleration of technology. It may seem paradoxical that faster progress can hurt wages and jobs for millions of people, but we argue that’s what’s been happening. As we’ll show, computers are now doing many things that used to be the domain of people only. The pace and scale of this encroachment into human skills is relatively recent and has profound economic implications. Perhaps the most important of these is that while digital progress grows the overall economic pie, it can do so while leaving some people, or even a lot of them, worse off.
And computers (hardware, software, and networks) are only going to get more powerful and capable in the future, and have an ever-bigger impact on jobs, skills, and the economy. The root of our problems is not that we’re in a Great Recession, or a Great Stagnation, but rather that we are in the early throes of a Great Restructuring. Our technologies are racing ahead but many of our skills and organizations are lagging behind. So it’s urgent that we understand these phenomena, discuss their implications, and come up with strategies that allow human workers to race ahead with machines instead of racing against them.
Titos Anastassacos is Managing Partner at Si2 Partners, a consultancy helping clients leverage services to win in industrial markets
Further articles on this and other topics can be found in the Si2 Partners Resources Page and the Si2 Knowledge Center
If you are a Service professional (manager, practitioner, consultant or academic) in an industrial setting join our group Service in Industry on Linkedin
Trackbacks/Pingbacks