Part 1: Rise of the Autonomous Machines

How we got here

Over the past decade, advancements in artificial intelligence (the “brains”), sensing technology (the “senses”) and modular components (the “body”) combined to make robots smarter, nimbler and significantly more capable than their predecessors. 

Most importantly, they are now truly autonomous.

We like to think we’ve played a small part in this evolution. For the past seven years, we’ve invested in innovations around sensor technology, data processing, artificial intelligence and modular components — all the necessary components to create a truly autonomous machine. We see the potential for automated machines in a growing list of industries, now that they are truly automated and small enough to drop into existing building footprints (or wander freely in office parking lots and hospital corridors). 

Equipped with advanced sensors and AI, today’s autonomous machines can fly around, drive around and move around in the real world among human counterparts. This opens a world of possibilities for robotics companies.

A convergence of global trends is driving these advancements:

  • Mobile computing power has increased dramatically, while the price of electricity needed to operate devices has dropped.
  • The global supply chain for phones is driving down the prices of mobile CPU/GPU, sensors, accelerometers and depth sensors — the components that allow robots to be aware of their surroundings. This is driving the technology that makes robots intelligent. Whereas robots from a decade ago were 90% hardware, there is much more development happening around the software.
  • Accessibility of modular components (motors, wheels, actuators, etc.) is driving down hardware costs.
  • Rising global demand is driven by a number of factors, including labor shortages, rising labor costs and COVID-19.
  • The maturation of e-commerce (especially in the wake of the pandemic) and the resulting proliferation of logistics companies to support the industry are helping to drive demand even more. 

As a result businesses across a wide array of industries can now make a case for investing in robotics.

Thanks, carmakers

We have electric car companies to thank for much of the progress in automation. 

Unable to constrain edge cases the way a robotics maker could for, say, a warehouse, car companies must solve their edge cases before they have anything to sell. Companies like Waymo, Cruise, Zoox and Tesla have advanced the ball significantly on this front by focusing on the underlying technology that allows machines to learn and react to their environments. 

All of that technology works beautifully for a robotics maker creating a parking lot sweeping robot, for example. 

Armed with open-source software, modular bodies, and sensors developed to keep pedestrians safe from moving cars, today’s automated machines are capable of addressing a huge array of problems. 

Heavy industry

We invested in Everest Labs this year for precisely this reason. We saw that today’s autonomous machines make for a much stronger business case in a recycling center than “robots” from even a few years ago.

Most recycling centers still use human labor, which comes at a particularly high cost due to the nature of the work (the three Ds: dirty, dangerous, dull). Yet, the alternatives have come with sticker shock of their own, requiring significant upfront capital investments in both the machines and retrofitting of plants to accommodate them. On top of that, much like a car plant’s “dumb” robotic arm, these robots could not function alongside human workers.

Everest Lab saw the opportunity to leverage new technology and a new business model to drastically lower the expense and complexity of automating assembly lines while also allowing it to drop into an assembly line next to workers.

The company developed a much simpler, smaller and smarter machine that closely emulates the tasks performed by humans on an assembly line. Crucially, the machine senses its external environment and can react to unpredictable human movements.

All of this allows recycling centers to pay a modest monthly fee for the service and quickly and seamlessly integrate it with its existing plant (we’ll address the evolution of this RaaS business model in Part 2 of this series).

E-commerce

We’re seeing a similar transformation in e-commerce, an industry that’s matured in recent years and exploded in 2020 as a result of COVID-19.

The proliferation of logistics companies tasked with delivering goods to consumers opens a new market for autonomous machines.

The founders of our portfolio company Osaro saw the opportunity early and got to work on the brains of a simple-looking machine that could pick and sort through packages alongside human counterparts, much like Everest Labs’ work on an assembly line. The company pioneered the use of AI software in industrial systems to expand the range of applications for robotics, automating the process of accurately seeing, picking and placing objects.

The seemingly simple task of picking up different sized packages and putting them where they need to go on a factory floor is a deceivingly complex problem for roboticists to crack.

Today Osaro’s systems are deployed in grocery, cosmetic and other commercial sites helping them automate the most repetitive tasks in their distribution centers and factories.

Service sector

The global pandemic is also driving demand for robots in the service and healthcare industries.

As The Wall Street Journal reported recently, hotels are eager to invest in robots to reduce human-to-human contact of their staff with customers. This includes everything from delivering towels to cleaning rooms. 

Hospitals, too, are eager for more options to reduce contact by deploying robots to deliver supplies between rooms.

Steve Cousins, CEO of our portfolio company Savioke, has been working on this problem for more than a decade. In his previous role as CEO of Willow Garage, he helped develop the open-source Robot Operating System (ROS) that powers the majority of robots on the market. His company built a relatively simple, affordable and adaptable delivery robot using ROS and off-the-shelf hardware that’s now deployed in dozens of hospitals and hotels across the world. 

Glen Kegley, chief operating officer of Hutchinson Health in Hutchinson, Minnesota, said this about the robot (which is dubbed “Spencer”): “I think that Spencer will become even more functional as we move on because there are so many things that he can do that we have even attempted to do yet.”

As the startup ecosystem continues to advance these technologies, we expect more and more executives to share Kegley’s enthusiasm for what autonomous machines can do for them.