Mercedes and Microsoft Collaborate on Joint Data Platform

Microsoft and Mercedes-Benz recently announced a new collaboration that merges the automotive world with artificial intelligence and data analytics to provide real-time feedback. The goal of this newly formed partnership is to enhance the efficiency of Mercedes vehicle production, address supply chain issues, and dynamically allocate resources to prioritize electric models and luxury vehicles.

The new data platform MO360 is a true evolution of the digital manufacturing ecosystem of Mercedes-Benz. It will connect up to 30 passenger car factories worldwide with Microsoft Cloud. The platform is standardized on Microsoft Azure and provides the flexibility and power of cloud computing to deploy AI technology on a global scale while addressing cybersecurity and compliance standards across all regions, according to Microsoft’s statement. The data platform is already available to teams in Europe, the Middle East, and Africa and will be deployed in the USA and China.

Virtual Replica of the Production Process

– We are making the transition to electromobility even faster and more efficient. The data platform provides us with artificial intelligence technologies that make our production more efficient, our supply chain more resilient, and also supports sustainability at a higher level – said Jörg Burzer, a member of the board of Mercedes-Benz Group AG, to VentureBeat.

With the new MO360 platform, the German car manufacturer can create a virtual replica or digital twin of the entire vehicle production process, combining insights from assembly, production planning, factory logistics, supply chain, and quality management. Before production begins in the factory, virtual simulation and process optimization will accelerate operational efficiency and increase energy savings, the statement said.

– What is different is the real-time digital feedback loop created by the industrial metaverse – said Judson Althoff, Microsoft’s EVP and Chief Commercial Officer, to VentureBeat.

Most previous AI projects, even those very successful, have mostly been some kind of offline analytical exercise, he explained.

– In other words, let’s take a bit of machine learning, operate on a dataset, utilize some cognitive services to study the problem, then learn from it and implement something new or different based on what we’ve learned. In this scenario, you actually have a digital twin that teaches people, and then people in turn teach the model for continuous real-time learning – he concluded.