Robotics markets: up, down or just sideways?

Amazon this week introduced its Sparrow robotic arm for repetitive warehouse tasks just days after FedEx powered down Roxo, its last-mile delivery robot, and Amazon said it was ending field tests of its home delivery robot called Scout.

What gives? Is robot technology ascending or not?

There are fits and starts with every technology, but robots as a massively broad category are clearly on an upward trajectory even if there are not investors for every startup or customers for every robot product and type or research dollars to spend on every robot segment—warehouse, home care, hotel reception, robotaxi or what have you.

Some robotics market segments appear to be doing better than others, but even then it is really hard to quantify success for the overall industry. For every major company like Amazon dropping a field test, an ambitious smaller company like Starship appears to have branched out to tackle a new customer segment. In addition to delivering cookies to college students on all-nighters, Starship Technologies wants to move your spare machine parts from the inventory warehouse over to the blast furnace.

From a market perspective, Sparrow is an industrial robot, a segment that Amazon clearly believes deserves investment. In June, Amazon also unveiled its first fully autonomous robot, Proteus, to operate alongside warehouse workers. At the Delivering the Future event in Massachusetts on Thursday, Amazon also showed off dozens of mobile Amazon Hercules robots.

So, clearly, there is a long-term raison d’etre behind the industrial robot segment, especially to replace human workers in short supply. Consumer-facing robots, especially home delivery, may not be as popular, or valuable, or worthwhile, at least for now.  Amazon’s Scout wasn’t meeting customer needs in every aspect, a company spokesperson said.

RELATED: Amazon ends Scout robot tests amid slower growth

Over at FedEx, the feeling about Roxo’s pullback was that the company wanted to put more emphasis on near-term opportunities, according to a spokeswoman.

When companies abandon field tests of robots – which can incur substantial costs in research dollars and take up engineering staff time -- they are essentially balancing the tradeoff between robot costs and their near-term usefulness.  Matt Beane, an assistant professor in technology management at University of California, Santa Barbara, told The Wall Street Journal  there have thus far not been good examples of successful robotic system deployments in “low-margin, public-facing settings”—such as last mile delivery and restaurant automation.

Investors care about such decisions to put a tech trial on hold or in the trash, but big companies like Amazon and FedEx can protect their losses with gains in other segments, hopefully keeping investors happy.

 Research, however, is research. There is no engineer worth his or her salt who will say a scrubbed test or a lackluster result was a failure.  All of it gets chalked up to experience and what can be done better. Consider the Wright Brothers and their development of hundreds of different wing shapes to produce lift. Or Edison’s famous light bulb comment, “I haven’t failed – I’ve just found 10,000 that won’t work.” (Sorry to preach here.)

Most of us would bang our heads on the wall, and maybe that’s what investors do when a startup has to pull back or go out of business. Engineers, instead, make notes on what didn’t work in hopes there will soon be funding to move ahead.

So, some setbacks in robotics are obviously happening but it’s best to be precise about what is working and what is not. There is undoubtedly a lively debate on how to spend dollars on robotics research inside Amazon: engineering team A making industrial robots moves ahead; meanwhile, team B at work on delivery robots moves back. 

Financial reporting and analysis on the robotics segment is frustrating just because it is such a broad field and losses in one field of robotics can be made up in another.  A good example is how Nvidia CEO Jensen Huang has described the autonomous vehicle as just another type of robot. Given the vast investment in self-driving car technology by multiple global players, his is a prescient way to view the robotic reality, since many different types of robots will require Nvidia GPUs and other chips in various components working in tandem with multiple sensors.

Qualcomm has a similar vision, with its next-gen RB6 Platform for faster processing for AI tasks, applicable to a multitude of robot types.

RELATED: Qualcomm boosts robotics with next-gen RB6 Platform

When Ford recently pulled back on its Argo AI effort, it was not a condemnation of fully autonomous driving technology, but more a recognition that Ford would not be creating that tech used in robotaxis and the like and would be leaving it to others.  At least for now.  That’s not a bash on robotics at all.

RELATED: Self-driving vertigo: Mobileye, Ford and Tesla safety

As for delivery robots and their future, maybe it is enough for smaller players to stay in the game for the foreseeable future with or without Amazon or FedEx.   Allied Market Research puts the global delivery robot market at $30 billion in 2030, up from $3.5 billion in 2020.  Even if a nasty recession crops up and that projection is reduced by half (not likely), there will be market segments that want robotic delivery at airports or on university campuses, even if not so much in suburban neighborhoods.

In robot delivery, take startup Ottonolmy.IO, which raised seed funding in August of $3.3 million after showing up in January at CES in Las Vegas fresh on news of robot delivery success at Cincinnati Airport .

RELATED: What it’s like to head up a robotics startup at CES 2022

Matt Hamblen

On college campuses, Starship Technologies appears to be going strong with its delivery robots.  First launched in 2014, the company now claims it has made 4 million autonomous deliveries at numerous colleges; it just launched service at Ball State University in November. In early 2022, it raised $100 million in 30 days. In October, Starship announced a partnership with Grubhub on college campus deliveries in the US.

 The company has lately begun promoting its delivery robots for industrial uses, including moving spare parts, samples and mail on large campuses and manufacturing sites.

Note to robot analysts: Starship may have scored a delivery robot customer segment crossover, from deliveries of late-night snacks on college campuses to mail and parts deliveries at your friendly neighborhood steel mill.