Cloud computing may be a key driver for the growth and spread of robots, according to findings by ABI Research. In its Commercial and Industrial Robotics report, ABI presents a forecast for 2030 where cloud computing-fueled robotic services revenue could rise to $157.8 billion, up from $3.3 billion in 2019.

Rian Whitton, senior analyst at ABI Research, says future generations of robot deployment may be predominantly mobile and run on Wi-Fi and cellular network data connections. That could mean a need to handle hefty amounts of data for robots to operate autonomously.

Whitton says cloud providers AWS, Google Cloud, and Microsoft Azure have been collaborating with robotics developers on this. He describes this space as robotics-as-a-service teamed with software-as-a-service. This pairing could help advance mobile robotics, which currently is associated with single units that have been deployed in logistics or factory work, according to Whitton. “They’re not really deployed in service operations like retail or real estate to any significant or great degree,” he says.

Putting more robots in the field is challenging, Whitton says, because they amass huge amounts of data through the sensors used to localize themselves and navigate their environments. He estimates a robot might process or create 500 billion gigabytes (exabytes) of data a week, potentially per day, depending on circumstances. “The more robots deployed, the more expensive it gets,” Whitton says. “As a technology, it doesn’t really scale very effectively at the moment.”

That is where he says cloud computing and infrastructure could help reduce processing costs for data while increasing the capacity for that data. It would shift the burden from the robot’s onboard technology and onsite computer resources to the cloud. This may lead to larger fleets and ecosystems of robots, Whitton says. “In the future, you might have a mall where you have robots doing cleaning, robots doing security work, robots doing material handling at the backend of stores,” he says. “You might have loads of robots doing very different things. It’s important to orchestrate the data between these different systems.” Some degree of cloud computing will be necessary to facilitate that, Whitton says.

Insight and analytics that can be derived from cloud computing could also be important. Cloud services and analytics providers are developing software, but that could take data from robots and turn that into valuable insights for the end user, he says. “It’s about creating the necessary architecture to scale robots up from individual units to broader systems.”  

Whitton says cloud services providers are expanding their robot-centric portfolios such as with AWS RoboMaker, a robotics platform for developers that applies cloud computing services to simulation software and motion control for robots. Microsoft has been doing its own work on autonomous systems. “Service providers definitely see the opportunity here, but I think a big problem is the culture of where robots are currently deployed,” he says. “A lot of it is in factories, which may be very skeptical about shifting to cloud systems to manage their data because of the security issues.”

Such reluctance might ease, Whitton says, as Microsoft Azure and AWS increase their presence in the manufacturing sector. Further, newer robotics companies seem to embrace the cloud, he says, by emphasizing the robotics-as-a-service model. In such an arrangement, the developer leases out equipment while the customer uses it on a monthly subscription.

A number of technology challenges such as lack of interoperability could still hinder the momentum cloud computing might offer to robotics, Whitton says. Legacy business models and industry inflexibility could compound the dearth of interoperability. “Having a factory where you use different robots from different vendors is very difficult, especially from a data orchestration point of view,” he says.

Standardized data formats will be necessary, Whitton says, for cloud computing to work in such a setting. Open source software might help with that issue. Some vendors are building common platforms that can offer interoperability between different pieces of hardware. “A good example of that is Ready Robotics, which is building a common control platform for industrial robots,” Whitton says.

Latency and speed requirements can also be issues for cloud computing, he says. A core goal for cloud computing is high velocity data streaming alongside real-time analytics at scale with no manual tuning, Whitton says. Currently available wireless connectivity, he says, does not offer the necessary low latency to establish such a perfected environment. “There are significant speed limits imposed on the amount of data that can be processed,” Whitton says. The rise of the 5G wireless network is expected to significantly reduce latency of data streaming while increasing the potential speed of data processing, he says.

If consolidation comes to robotic software vendors and cloud software vendors, Whitton says companies such as AWS and Microsoft Azure might create their own robotics-as-a-service platforms. Robotics companies might also follow suit developing platforms where robots from different developers and vendors can plug into a service system that is accessed by end users in a range of markets. “We are really at the beginning of this,” Whitton says. “You’re likely to see it start to manifest around 2025-2030.”

For more content on robotics and the enterprise, follow up with these stories:

Robotics Will Prevail Over Other Tech Sectors Post-Pandemic

The Robots Are Still Coming

The Art and Science of Robot Wrangling in the AI Era