Public Cloud Vendors Getting Into Cloud Robotics Game

Subscribe To Download This Insight

4Q 2018 | IN-5290

Google announced its cloud robotics platform in October 2019. The promise of robot fleet management, path finding, and object tracking at a massive scale can now be delivered using the strength of a public cloud platform. This foresight dives into the efforts of public cloud vendors in cloud robotics and how they will impact the future of robotics industry.

Registered users can unlock up to five pieces of premium content each month.

Log in or register to unlock this Insight.


Google Cloud Now Offers Cloud Robotics Platform


Recently, Google announced its cloud-based robotics platform will be made available to developers in 2019. Developers can use the platform to collect, store, and process data from various sensors and integrate hardware and services from third parties. Users of the cloud robotics platform will have access to all key components of the Google Cloud Platform, including Kubernetes (Google’s cloud container technology), Cloud Bigtable (Google’s big data management platform), Cloud AutoML (Google’s cloud-based Artificial Intelligence [AI] development platform), and flexible automation services.

As one of the market leaders in public cloud platforms and AI, Google has demonstrated its ability to identify technological trends and deliver with great execution and implementation. Google is one of the early proponents of cloud robotics platforms, as they introduced the idea as early as 2012. However, the industry has been taking a long time to warm up to cloud robotics. Thanks to the advancement in cloud computing, several startups—such as CloudMinds, Neurala, Vicarious Systems, and—have started to push for the mass adoption of cloud robotics in recent years. While these startups have achieved some degree of success, the industry at large still has some reservations about cloud robotics technology.

A Sliver of Light at the End of the Tunnel


The promise of the platform that is able to perform predictive maintenance, optimize functionalities, and introduce AI capabilities to robots using a scalable, flexible, and highly secure public cloud is visionary and future proof. However, there are many concerns around its implementation; some of these concerns, such as cybersecurity and a fully data-driven robotics intelligence, are well justified. These have been highlighted in ABI Research’s recent executive foresight on cloud robotics. However, the capabilities and values of a robust cloud robotics platform are too obvious to be ignored. Industrial cloud platforms, such as ABB Ability, GE Predix, Siemens’s MindSphere and PTC’s ThingWorx, have actively integrated industrial robotics capabilities to their platforms. At the same time, other public cloud vendors, such as Amazon Web Services (AWS), Alibaba, and Huawei, have started to extend their services to robotics vendors via smart manufacturing platforms.

As such, Google’s announcement is no doubt a healthy dose of endorsement for the future development of cloud robotics. The company has always been very confident about its cybersecurity technology and is one of the key leaders in AI and machine learning. This allows Google to address the two aforementioned concerns of cloud robotics platform. In addition, developers will have the access to the Tensor Processing Unit (TPU), Google’s cloud-based AI chipset that is optimized for cloud-based training and inference of machine learning models.

Public Cloud and AI Vendors Are Pushing into Cloud Robotics


That being said, Google is not the only company that brings unique value propositions to cloud robotics. It is impossible to avoid discussing China every time Google is mentioned. Given its lack of presence in China, Google will definitely miss out on the advancement of robotics there. China is the world’s largest market for industrial robotics, and industrial robotics shipments to China are larger than that of the United States and Europe combined. As China continues its march toward smart manufacturing, the demands for smart robotics platforms will only increase, creating impetus for public cloud vendors to join the fray.

As such, the void has been filled up competently by Alibaba. Alibaba Cloud has focused on smart manufacturing and smart cities as its two main growth drivers and has close relationships with many Chinese manufacturers. In addition, Alibaba Cloud also supports commercial robotics via Cainiao Logistics, its logistics subsidiary, and has launched a slew of commercial robots in September this year—all leveraging Alibaba’s cloud-based facial recognition, natural language processing, and other AI capabilities.

As for the rest of the world, AWS hosts its own warehousing system under Amazon Robotics for all of Amazon’s warehousing and logistical needs. Not limited to commercial robotics applications, iRobot, a consumer robotics company, has been hosting its customer-facing applications and back-end Internet of Things (IoT) platform on AWS since 2015. AWS’s scalable infrastructure enables iRobot to store massive amount of real-time data and to expand its computing capabilities whenever special needs arise. Moving forward, iRobot aims to enhance the mapping technology inside its vacuum robots by combining data from extra sensors from other smart home devices and supportive cloud services.

At the other end of the spectrum, NVIDIA has launched NVIDIA Isaac, a virtual simulator for robots. This simulator leverages NVIDIA’s Graphic Processing Units (GPUs) to create a virtual environment for the training of robotics capabilities. Using data collected from sensors and actuators of industrial and commercial robotics, developers can perform machine-learning training and testing on their robots in simulation, before transferring the capabilities to physical robots. As compared to AWS, Isaac focuses solely on robotics training and simulation. Therefore, developers may have to rely on other solution providers for the processing and storage of other functions.

ABI Research would argue that the decision for any user to go with public cloud vendors should be based on the compatibility with existing industrial platforms, upfront costs, and cybersecurity requirements. Nonetheless, it is still critical for developers to evaluate the project scope (e.g., integration with other technologies, such as AR and digital twins), protocol adaptability, connectivity, and edge intelligence requirements of the robotics applications before deciding to go with a particular vendor. As more and more players are coming up with innovation solutions, the cloud robotics space is going to be become crowded and competitive.