Sona Comstar accelerates global EV innovation through cloud migration

As EVs reshape the automobile industry, Sona Comstar set out to expand beyond automotive components and develop electric drivetrain systems for the next generation of vehicles. The company was one of the first in India to develop advanced electric drivetrain systems, becoming a preferred supplier to a major US EV-only manufacturer. It formalized this direction under its EPIC strategy—Electric, Personalized, Intelligent, Connected—and went public on the National Stock Exchange in India as it prepared to scale its EV systems business.

To support this next stage of growth, Sona Comstar needed to transform the IT systems behind its business and manufacturing operations. Its aging on-premises infrastructure created recurring scaling challenges, while data silos limited plans for automation, data management, analytics, and AI. “Our systems had to integrate easily, and scale as needed to support our design engineers,” says Rajesh Rajendran, group head, Business Analytics & Chief Information Security Officer at Sona Comstar.

Sona Comstar adopted a phased hybrid multi-cloud strategy, choosing AWS to support critical workloads across its business, engineering, and innovation environments. The company started with its SAP enterprise resource planning (ERP) system before expanding its cloud roadmap to include its Siemens Teamcenter product lifecycle management (PLM) system, high-performance computing (HPC) environment, and AI Workspace. “Our mindset was to start gradually and build strong,” comments Rajendran. “AWS offered the breadth of services, partner network, and cloud specialists we wanted to deliver our strategic goals.”

With support from AWS, Sona Comstar migrated its critical ERP system to AWS in just 5 weeks. During the migration, the company moved its core SAP S/4HANA system to Amazon Elastic Compute Cloud (Amazon EC2) with Auto Scaling instances, along with 15–16 ERP-related applications and third-party network security and backup systems. The project also required managing complex changes across multiple functions and bringing the wider SAP-dependent ecosystem live within the same 5-week window. “We completed the migration without disruption to our operations,” says Rajendran. “It gave us a lot of confidence in our cloud strategy, encouraging us to press ahead with the PLM move to AWS.”

Building on that momentum, Sona Comstar began advancing the next phase of its cloud strategy across engineering workloads. As part of this, the company planned the migration of its Siemens Teamcenter PLM system and began piloting a hybrid HPC model on AWS. Design engineers will run pre-processing tasks on premises, such as defining design geometry and materials, while solver software runs compute-heavy simulations to verify designs on Amazon EC2 hosts, with Amazon Simple Storage Service (Amazon S3) for object storage. The pilot is helping Sona Comstar evaluate how cloud-based compute can support more flexible engineering workflows as its EV systems business grows.

Sona Comstar is also building an OT data lake on AWS to centralize shop-floor data for future analytics and automation initiatives. Alongside this, the company built its AI Workspace on AWS to support compute and infrastructure needs across teams, from traditional machine learning to generative AI projects. The environment uses Amazon Aurora PostgreSQL serverless for its data layer and Amazon Elastic Container Service (Amazon ECS) for application hosting, with generative AI capabilities powered by Amazon Bedrock and Anthropic models. One use case supports maintenance and service engineering teams through a chatbot that helps troubleshoot technical issues with machinery on the factory floor. Engineers can ask questions through a web interface, and the chatbot searches relevant manuals to find answers. “We’ve been able to scale the solution across three business units thanks to its flexibility,” states Rajendran.

By migrating to AWS, Sona Comstar has modernized the IT foundation supporting its business, engineering, and manufacturing operations. Improved SAP performance has increased availability to 99.5 percent, while compute resources can now scale within 30 minutes. This helps teams provision capacity faster and support critical workloads across its SAP, PLM, HPC, and AI environments as business demands evolve.

 

Sona Comstar has also improved support for maintenance engineering teams using its generative AI-driven chatbot. Engineers can now access troubleshooting information from digitized manuals in under 5 minutes, compared with a couple of hours previously spent searching through pages of paper documents or PDFs. By helping teams find answers faster, the chatbot can reduce possible disruption on the factory floor and support quicker maintenance response.

The move to AWS is also helping Sona Comstar evaluate how cloud-based HPC could support faster engineering innovation. In pilot testing, design iterations can run in under 3 days, compared with weeks on the legacy HPC environment. “By using HPC on AWS to evaluate multiple designs and optimize existing ones faster, we can accelerate our time to market,” says Rajendran.