Published in


Robotics will eat the world 🤖🍽🌍

Part 1: Industrial Robotics, why we love the space and which tech companies to watch out for

At Paua Ventures we are passionate about technologies that will radically transform the future we will live in (for the better 🍀). That’s why Robotics & Automation is one of our core focus areas, reshaping the way we produce, pack, transport and consume goods (and services). In recent years we have done some fantastic investments in the space (both into HW and SW startups) and mapped out the main market dynamics and active tech companies. This series is to share some of our thoughts on the space and send a little “hello world” to all the amazing brains in robotics out there. Enjoy reading and please get in touch! We are glad you exist and would love to receive feedback. 🥳 🙌🏻✨

Before we get started, let’s make sure we talk about the same things: Traditionally, an “industrial robot” means a big and heavy stand-alone industrial robotic cell, that is fenced off from humans for security purposes (~380,000 global shipments in 2018). But three newer categories are gaining traction: Firstly, collaborative robots (cobots), which are generally smaller, lighter, more affordable and designed to work safely in proximity with humans (~14,000 global shipments in 2018). Secondly, mobile robots (AGVs and AMRs = Automated Guided Vehicles and Autonomous Mobile Robots), used mainly for transportation purposes (~20,000 global shipments in 2018). And thirdly, exoskeletons, designed to enhance human performance (~7,000 global shipments in 2018).

Fig. 1: Classification of Industrial Robots in the wider sense (adapted from McKinsey & Company, 2019)

Now, why do we think this market is so exciting? 1) rapid growth 🚀✨ and 2) huge societal impact. 🍸🏝🚣🏼‍♂

Size & Growth. Few markets have shown such strong and sustained growth as industrial robots in the past decade. In recent years, annual installations have risen by a 19% CAGR, resulting in 422,271 new global installations in 2018, worth USD 16.5bn. This was only for the robotic hardware, while software/services and peripheral hardware roughly amounted to the same value each. End of 2018, the total global operational stock of robots was 2,439,543. Asia is the growth engine behind these numbers (two out of three new robotics installations happen in Asia), with China leading the pace (increasing operational stock of robots by almost 30% y-o-y in the past years). Europe and the Americas are slightly falling behind. While future global growth is estimated to continue strong (~13–15% y-o-y), I am personally convinced that in a few years we will see growth rates much higher than that…but more on that later in this post. 👀🔥 🚀

Fig. 2: Annual shipments (in thousands) of industrial robots worldwide. Source: IFR, McK

Societal Impact. Besides rapid growth, what is most exciting about robotics is the fundamental impact it will have on our future lives, potentially ridding us of most manual work. Despite a commonly different perception, even in “highly automated” industries such as automotive, the average degree of automation (=number of automated tasks / number of total tasks) in final assembly is only around 8–10%. And most other industries are far below that. On the other hand, up to 85% of all tasks are generally estimated to be “automatable”. So in reality the industrial robotics party is just getting started! Climbing this curve of automation holds tremendous opportunity for increasing global productivity and welfare (“lives of luxurious leisure”), if we as a society are able to share its benefits equally. While we at Paua Ventures are optimistic about this (🥃 = half full), there are/were people out there predicting “ever-increasing inequality” (e.g. Prof. Stephen Hawking). At least in the short/medium-term we are likely to see the following societal effects of robotisation, which will be challenging but hold great opportunities for bright entrepreneurs like you. 👾🏆💸

(a) A big reduction of manual labour: a recent BCG survey of 1300 global executives expects total employment in their respective companies to go down anywhere between 5–20% in the coming 5 years until 2025, where blue-collar labour is expected to take almost the entire hit. Other studies suggest a milder decline of roughly 10% of the global manufacturing jobs in the coming decade. In any case, it will mean dozens of millions of people that need to re-invent themselves professionally.

(b) This will lead to a big skill gap and re-skilling opportunity in our economies. A good example are the Adidas “speed factories”, which went live in Germany and the US in 2016/2017 (and are currently being moved to Asia). Using a high degree of robotisation in an industry that is traditionally manufacturing manually, one of these factories employs only 160 highly skilled workers versus a thousand (or so) low skilled workers in a typical factory in Vietnam or China.

(c) Further, developed nations will be able to re-shore many of their manufacturing activities as differences in global manufacturing costs will even out (=a robot costs the same $$$ in Vietnam as it does in Germany). This in turn will lead to a disproportionately high impact on manual labor in developing countries and a shift from global to regional supply chains.

Now, let’s go back to why I am convinced we will see a real tsunami of robotics eating the world, with explosive growth well beyond the often predicted 13–15% CAGR? 💨️🌊🏄🏻‍♂️

Firstly, we have to understand that we are still in the Stone Age of automation and are barely getting started from a very low base with tremendous room to grow. Globally we have an average of only 99 robots installed per 10,000 manufacturing workers and only five countries are holding 74% of all the installed base (China, Japan, Korea, Germany and the USA). As competitive pressure will increase, less industrialised countries/industries will (have to) catch up quickly.

Secondly, catalytic events like COVID-19 🦠 will reshape the global manufacturing ecosystem. Corporates will strive to a) reduce their dependency on overly complex and intertwined global supply chains, and b) reduce their reliability on “fragile” human labor (=robots don’t need social distancing). Hence, once the initial dust of the virus shock has settled and investment budgets free up again, we will inevitably see a wave of re-shoring and automation.

Thirdly, and most importantly, robots are essentially still too expensive and too “stupid”, allowing for only very few profitable areas of application (=mainly highly standardised mass-production in the automotive and electronics industries). But we are facing a perfect storm of breakthrough inventions that will soon unlock a tipping point, where robotics will become significantly cheaper and more versatile than most of manual work. Despite not including data from recent years, the chart below illustrates this well: annual EU & US patent applications for advanced manufacturing technologies (including robotics) have only started to take off in the past decade. As these exponential R&D efforts will hit commercial maturity, an unprecedented wave of new (and profitable) robotics use-cases will be unlocked. 🎯🚀🔥

Fig. 3: Annual patent applications for specific advanced manufacturing technologies. Source: OECD

Next, let’s dive deeper into the main areas of required innovation in robotics. In my view, today there are five major robotics technology gaps that need to be filled in order to trigger a “robotics eats the world” scenario. 🤖🍽🌍

1 Affordable & versatile hardware: In the past decade the average costs of a robotics installation have come down by roughly 40%, while its productivity has grown by about 5% per year. Still, required CAPEX will have to further decrease significantly, in order to allow for a positive business case beyond mass-production use cases (=tapping into the massive SMB market). To give you a feeling, today an average robotics cell (including peripheral systems) for automating a machine-tending process, would cost somewhere around EUR 100,000. Using a cobot could bring this number down to roughly EUR 70,000–80,000 (while introducing certain performance limitations). In my view, however, to make a really compelling case for mass robotisation, the CAPEX for a fully process-ready robotics cell will have to fall below EUR 50,000. Also, new business models will be needed to help foster market adoption (e.g. “robot as a service” or rental models). At the same time, mechanical capabilities and physical properties have to improve (less weight, higher reach, faster speed, better modularity, better safety systems etc).

2 Ease of setup and teaching: Today, around 70% of the total lifetime costs of a robotics cell are generated by services revolving around setup, programming and teaching. Typically, this is done by a myriad of small to medium-sized system integrators, taking weeks to months to set up, program and certify a new robotics application. Hence, we are not only facing a costs & time problem here but also an inflexibility challenge. If production processes change, highly skilled personell needs to painfully re-program every single robot movement and go through most of the process again. To make things even worse, these highly skilled engineers are extremely scarce. Hence, the limited engineering capacities of system integrators are more and more becoming a bottleneck for robotisation growth. Consequently, non-expert tools for robot teaching will be a key element to unlock widespread robotics adoption. 🚨 Spoiler alert: Our portfolio company Wandelbots is one of the global leaders in trying to crack this nut. 🥜

3 Intelligence & Skills: After painfully hard-coding a robotics application, it still remains largely rigid and inflexible. If, for example, the shape or position of the parts to be picked by a robot are changing over time, in most cases this would mean “game over” for an industrial robot. Luckily there are hundreds of very bright minds working at the forefront of Machine Learning and AI to solve this lack of “intelligence”. This ranges from giving the robot a general understanding of its surrounding and context awareness to pre-training and productising very concrete skills (e.g. pick&place, machine tending, painting, welding, gluing etc.). Also the current momentum in industrial cloud computing will serve as an accelerator pedal here, allowing for new robotic “skills” to get bulk-deployed to the installed base at the push of a button. Today, only around 1-5% of all industrial robots are estimated to be connected to the cloud.

4 Periphery & Integrations: What often gets overlooked, is that the robotic arm is only part of the story. A robotic cell always comprises additional hardware elements such as grippers & tools, conveyor systems, security systems, 2D and 3D cameras etc. Most of these peripheral systems come from different vendors, don’t have standard interfaces and don’t easily talk to each other. Hence, another reason for spending lots of $$$ on expensive and complex systems-integration projects before a robotics application is up and running. Creating standards and modularisation will be key to allow for a “consumerized” plug&play experience across different systems providers.

5 Operating System & Certification: The last challenge is an overarching one. Today, there is no “MS Windows” for industrial robotics, which would allow all individual pieces of above puzzle to communicate with each other. Despite some open source efforts, such as ROS2 (=Robotics Operating System 2), no real industry software standard has yet been found. This is surely a mammoth task, but definitely one of the most important gaps to be filled. Furthermore, another overarching challenge is posed by rigid and painful legal certification processes ahead of any go-live of a robotics installation. In Germany, for instance, legal certification, documentation and industrial safety processes after setting up a robotics application might delay the go-live another 2–6 weeks. This is one of the main reasons why only about 1–2 out of every 10 cobots sold in Germany are actually used in a “robot+human” collaborative environment. In reality, one of the main driving forces behind cobot adoption are the significantly lower & quicker safety/certification requirements. Hence, local regulators will need to simplify and streamline legal requirements considerably, if they want to give their respective industries a chance to stay competitive in the global marketplace. Please listen, dear regulators 🙏🏻!

To conclude and show the buzzing action in the field, I have assembled an Industrial Robotics Market Map with some of the most prominent tech companies striving to fill above technology gaps 🔧🤚🏻. Please note, that in some cases the boundaries are not 100% clear cut as companies might be active in multiple areas at once.

Fig. 4: Market map of key tech companies in robotics (PAUA Ventures)

If you feel your company is missing on this map, please help me fix this via a quick typeform. 🕵🏼‍♂️🤖

If you have any thoughts and comments on this piece, I would be happy to hear from you -> 👋🏻📧

Else, many thanks for your interest and please stay tuned for the next parts of this series, diving deeper into Cobots and Service Robotics!




We invest in early stage B2B software Start-ups

Recommended from Medium

EV Startups Launching 200-Mile Battery Cars in 2024: Are You Kidding Me

2022 Investment Themes: Data Assembly Line, Rise of “x”Ops, and Old Economy Transformation

Investments in E-commerce: A Closer Look at the Top Investors

Case Study: Top/Bottom-ESG Long/Short Trading Strategies

Why is it the right time to invest in AI, in India? — A look back Analysis in India.

The Ultimate Dreamforce Survival Guide

How to Stop Burning $412 A Week In Shipping Fees

Did You Know: PeopleSoft Financials Can Be Made Fully Compliant With Making Tax Digital

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Georg Stockinger

Georg Stockinger

Partner @ Paua Ventures

More from Medium

They pulled a Fast one on us, again

So Why Venture Capital…. & The journey so far

Why we invested in Sano Genetics

Corporate venture funding models: Determining the sweet spot between risk and control