BMW Founds IDEALworks in Bid to Dominate Mobile Material Handling

In November, the BMW Group founded IDEALworks (IW)—a fully-owned subsidiary—headquartered in Munich. The stated aim is for IW to become a leading supplier of autonomous robotics solutions in the logistics sector, with the name “IDEAL” standing for Industry-Driven Engineering for Autonomous Logistics.

BMW has been partnering with Nvidia to develop mobile robots for internal use in its factories, primarily around automated material handling at the last mile, for a number of years. IW builds on this internal development and expands the scope to include autonomous robots in the logistics sector, suggesting that BMW is targeting couriers, 3PLs, retailer, and e-tailers.

This is exceptional in that, while major automotive vendors have frequently bought mobile robots, and in the case of Toyota, have sold automated forklifts via the use of technology provided by automation specialist like BlueBotics and Kollmorgen, there is little precedent for such a large player developing its own robotics division and selling outside of its traditional market. In the case of Amazon, the company is acquiring robots to develop in-house.

At the moment, BMW plans to sell into logistics. However, the company has been using the technology stack behind IW for its own production facilities since 2017. The robot deployed is referred to as the Small Transport Robot (STR) and is equipped with a Lips 3D camera and Sick sensors for safety. All robots rely on the Nvidia AGX hardware and make significant use of Nvidia’s SDK. Among many other things, this is a major endorsement of Nvidia as a key part of the future robotics ecosystem, as the company faces increased competition from Qualcomm and Intel. At least when it comes to the highest performance robots for industrial applications, BMW considers Nvidia the best choice.

At present, five IW deployments exist in BMW’s seat production, with more on the way. The company planned to integrate up to eight robots by the end of 2020. The IW.hubs are responsible for running the seat assembly lines in 10-minute cycles with pre-assembled backrests, heads, and center armrests. A key question, of course, is when the current business case for using IW.hubs to this extent will pay off. With the investment of four hubs, four employees can be relieved from mundane and repetitive loading and unloading tasks, so that they can focus on their core competencies. In a two-shift operation with eight hubs, the purchase will pay for itself after only 1 year, based on an average wage cost of EUR€75,000 per year and employee, marking a fast ROI for this use case.

Two things can be inferred from such a dramatic announcement. One, this is a validation of the mobile robotics industry. A large automotive vendor clearly sees the viability behind AMR technology and wishes to pursue the opportunity through a dedicated solution and service provider. BMW’s annual revenue in 2019 was US$126 billion dollars, while the robotics market’s value in the same year was US$27 billion worldwide, with the mobile robot market making up US$11.8 billion of that revenue.

Two, having such a large player entering the market with the intention of selling beyond its market is also a major challenge to the incumbent vendors, a throwing down of the gauntlet. AMR developers usually feed into companies like BMW, and don’t plan to compete with them. Earlier in the year, BMW argued that internal development of a mobile robot solution was necessary because it was not satisfied with the offerings of the market. As BMW plans to expand its offerings to logistics companies, it is clear that the company also believes this is a part of the market that is underserved.

ABI Research is soon to release a competitive assessment on the major AMR developers in industrial manufacturing, including BlueBotics, Balyo, Fetch Robotics, MIR, and Seegrid, to name a few. These companies have achieved thousands of deployments and have a wide range of customers. But in IW they face a major challenger with in-built advantages, the most notable being the ability to use the entirety of the BMW supply chain as a testing ground for its pilots.

That being said, IW won’t just be able to dominate this space. AMR vendors can take that the IW.hub is itself a rudimentary under-ride vehicle with no unique sensor solution. The use of stereoscopic cameras and 2D lidar is becoming fairly mainstream in AMRs, and even their use of Convolutional Neural Networks (CNNs) and autoencoders is not unique, as MIR has deployed this capability with the use of stationary cameras. Though IW likely will develop more robot sub-types like placing and sorting systems, the current roster doesn’t add anything unique to the market. What is more, due to the use of advanced processors and sensors, the IW.hub is reported to cost US$75,000 per robot, more than the price of comparable systems like a Fetch or MIR robot. The company does stress the use of simulation tools on its website, but any other player can use the Nvidia Isaac to get access to the same tools.

The main advantage for IW is therefore resources and having both the physical and digital playground offered by BMW’s infrastructure. This leads to the question of whether the ability to develop an excellent AMR is becoming much less of a differentiator than it might have been even 12 months ago? With the availability of excellent 3D cameras, increasingly affordable lidar, more processing hardware, and an open-source ecosystem anchored by ROS and SDKs like Nvidia Isaac, can any team of half-decent roboticists enter the market? This does seem to be the case, with hundreds of companies popping up and average sales prices decreasing markedly to less than US$40,000 per system. What is more, service models and flexible sales strategies like leasing are lowering the barriers to entry for even the more expensive systems like autonomous forklifts.

AMR developers can, of course, refute the notion of commoditization through product development and further innovation across hardware and software. This might take the form of mobile manipulation or automating new subcategories of vehicles like Very-Narrow-Aisle (VNA) forklifts, or through developing better semantic understandings of the environment through CNN advances. Yet the increasing availability of SDKs and open source material makes the benefits of innovation short-lived, and it is remarkably difficult to maintain a steady stream of advances in product innovation while hoping to scale your robotic fleet. It is therefore essential that AMR vendors master the scaling of their robots, and they should focus particularly on building partnerships with BMW’s competitors, not merely to compete with the resources of IW, but also to become essential to those manufacturers who might also consider developing their own roster of mobile robots.