Coming Out of the Cage: The Evolution of Automation and Robotics

Manufacturing efficiency gains have played a key role in economic growth throughout the 20th century, and since the 1970s, much of the growth in output can be attributed to more efficient manufacturing and assembly lines driven by automation and robotics. The availability and sophistication of sensors, data and processing power are driving robotics to become more mobile, autonomous and allow them to work alongside people. These applications range from warehouse fulfillment, last-mile logistics, factories, and food preparation – they make pizza! The estimated US market size for these applications in 2016 was $11.1B. This is expected to grow by nearly 12% CAGR through 2025 to $34B.

Traditional industrial robots: Working in the cage

An example of traditional industrial robotics in a closed environment within a factory. (Source: Fanuc Robotics)

Since the 1970s, traditional industrial robots have been commonplace in large-scale manufacturing. They are fixed in place within factories working in caged environments away from humans, where they work at a rapid speed. They dominate in heavy manufacturing sectors like the automotive industry, where 65% of total industrial robotics applications can be found today (according to quoted figures from our interviews with both GE Ventures and Rethink Robotics). In this large-scale manufacturing environment, these robots quickly handle repetitive tasks such as assembly, as well as heavy lifting and movement of payloads.

Despite their obvious benefits in terms of speed, strength and precision, these traditional robots have limitations. They are very expensive, costing up to $100,000-$150,000 for initial deployment, and require extensive coding and dedicated software engineers for programming and on-going maintenance, which adds significant additional costs to what is already an expensive factory installation.

They are also very inflexible and only capable of simple repetitive tasks. This lack of flexibility means that they are not well suited to small production runs, or any kind of agile, iterative manufacturing process. Lastly, they are dangerous – in factories where they are in use, the environment must be caged, with no humans allowed in the area when these robots are in use.

Cobots working among us

Advancements in technologies such as software, vision systems, sensors and Programmable Logic Computers (PLCs) have enabled improvements in flexibility, control, safety, autonomy and mobility. Traditional industrial robots are not in danger of extinction – they will continue to play an important role in heavy manufacturing. Collaborative robots, or cobots, are an emerging technology able to collaborate alongside humans on manufacturing and assembly lines, working on dangerous or repetitive tasks and supplementing the labor force. These new smaller devices don’t just repeat stored sets of motion, they interact with their environment and ‘learn’ how to adapt to changes.

New approaches to interfacing with these robots allows for a much simpler process of ‘teaching’ them how to carry out actions compared to large legacy industrial robots, which require extensive machine-level programming. This improved interface has been enabled by innovations in the mapping of behavior trees of complex actions – as Steve Taub, Managing Director at GE Ventures, pointed out to us in an interview. Many of the leading cobot developers have recruited programmers from the video game industry, who are experts in this type of intuitive application programming interface.

The price point of these cobots is considerably lower than large, fixed-in-place industrial robot systems. Lower cost sensors and accelerometers, driven by their proliferation in modern mobile phones, contribute to their cost often running around $25,000-45,000 per unit, 25% of their predecessors.

Within this market today, the supplier’s strategy is based around targeting specific segments, gaining market share and quickly expanding – there is a lot of low hanging fruit in smaller manufacturing businesses. From moving objects from one line to another, machine tending, pick & place, and quality evaluation, cobots are well suited to meet broad sets of manufacturing needs.

Rethink Robotics
Rethink Robotics’ two cobot models, Sawyer and Baxter. (Source: Rethink Robotics)

One of the larger startups in the cobot space is Boston-based Rethink Robotics, which has raised nearly $150 million in paid-in-capital to date. Jim Lawton, COO at Rethink, points to cloud-based technologies and the network effect as important differentiators for the company. The ‘brain’ of robots is no longer always sitting in the controller box wired to a robot – it is now a hybrid shared brain between local and cloud-based CPU. This allows significant increases in computing power, which is needed due to the huge amounts of IoT data being collected by the robots. When this collective learning is extrapolated across a network of robots, there is a significant effect in iterative improvement and access to information.

Rethink’s go-to-market strategy is based around a distributorship model, which accounts for about 95% of the robots currently sold. The robots are delivered with identical hardware and software for every customer, with easily taught software built in. This standardization plays a significant role in allowing the company to provide a competitive pricing model. This structure has enabled the company to expand quickly, with the initial focus on US markets followed by rapid expansion in European markets, where there is more openness to technology innovation.

Chinese expansion has provided a different set of challenges and opportunities. First, labor rates are rising rapidly but are still relatively low when compared to Western standards. Second, Chinese robotics companies are rapidly gaining on US counterparts, with a supportive government incentives program driven by the Made In China 2025 strategy, in which 70% of domestic goods will be made in China by 2025. Despite these issues, China remains a huge market for innovative robots – it is not unusual for China to account for around half of Rethink’s total monthly unit sales. However, with the recent escalation in tensions and tariffs, it remains to be seen how US/Chinese trade relationships will continue.

OmniRobotic and Robots-as-a-Service

Although cobots are much cheaper than their larger industrial counterparts, many smaller manufacturing shops lack the liquidity to invest significant sums in this type of fixed asset. Rather than simply being priced out of the market, there is an emerging shift in business models to allow smaller manufacturers access to cutting-edge factory robotics. In the same way that SaaS services have revolutionized IT, Robotics as-a-Service (or RaaS) can help manage cost via rental or temporary acquisition of hardware.

To get a better idea of the dynamics of this RaaS model, we spoke to Founder & CEO of Montreal-based cobot developer OmniRobotic. The company’s fixed painting robot is able to ‘learn’ and adapt to complex industrial processes, in which the robot interprets real-time continuous imagery of its environment to autonomously spray paint items on an assembly line.

The robots will be available to rent as part of a RaaS model, in which small, niche manufacturers can rent and install a vision system and robot over a weekend, and immediately test the robot in use – already at a cheaper rate per hour than labor costs for a similar set of tasks. This not only reduces expenditure for the manufacturers, but also allows increased flexibility of operations, being able to scale up or down depending on the current production needs. The company is currently testing proof of concept, after which it is hoping to roll out commercially towards the end of 2018 and raise a Series A round of funding early in 2019.

The combination of cheaper, smarter robots and innovative business models are allowing smaller manufacturers to compete, as well as enabling robotics usage in new market segments. For example, farmers, who are traditionally risk-averse and capital-poor, can access advanced robotic technologies via rental services. Pricing for agricultural RaaS services might be based upon the amount of fruit picked ($/Kg) or the area of land de-weeded, meaning that the pricing of robotics can easily be integrated into cost structuring for manual laborers.

Now that they are out of the cage…

The ability of these devices to work alongside people, instead of in cordoned off zones opens up a wide range of applications. Warehouse and logistics robots are allowing factories to automate large parts of their fulfillment processes and are attracting significant amounts of capital. In September 2018, GreyOrange, a Singaporean supply chain robotics company, announced $140 million from Peter Thiel’s Mithril Capital Management and two others in the largest industrial robotics round of funding to date. Chinese service robot developer Yunji Technology also raised $120 million in a Series B round earlier in 2018.

The ability to collect and process data in real time will continue to expand applications, replace labor, and improve efficiency and safety. We are also beginning to see robots moving into the services sector, a topic for further exploration.