How Toyota scaled SAP HANA with Intel Xeon 6 on AWS

Toyota uses SAP as its enterprise resource planning (ERP) solution across 14 North American facilities and roughly 30,000 employees. To keep up with the growth of manufacturing units, it migrated SAP to AWS, which provided the scalability that it needed as its database size grew from 6 TB in 2020 to 24 TB in 2024. However, even following the migration, Toyota encountered CPU limits that could overwhelm its architecture, forcing manufacturing and finance teams to revert to manual workflows and slow production while engineering teams worked to resolve disruptions.

With complex workloads requiring more memory and processing power, Toyota needed to upgrade its digital infrastructure. “We needed to stabilize our architecture so that SAP could run efficiently and teams could have confidence in it,” says Ashish Patel, senior manager of SAP technology at Toyota. “We needed a server with more CPU cores and faster memory speeds.”

With the need for higher-performing hardware, Toyota approached AWS and Intel about testing the newest Intel-based instances on Amazon Elastic Compute Cloud (Amazon EC2), which offers secure and resizable compute capacity to support demanding SAP workloads. AWS and Intel had been codeveloping Amazon EC2 High Memory U7i Instances—which are purpose-built to run large in-memory databases—and gave Toyota an opportunity to preview the new instances nearly a year before public availability.

Amazon EC2 U7i Instances deliver significant performance gains compared to the U-1 instances that Toyota had previously relied on, including 135 percent more compute performance, 115 percent higher memory performance, and up to 45 percent better price performance. The SAP S4/HANA application layer runs on Amazon EC2 U7i Instances combined with Amazon EC2 R6i Instances, memory-optimized instances powered by Intel Xeon Scalable processors. Together, both instances delivered the power and performance Toyota needed.

AWS Enterprise Support—which helps businesses accelerate innovation, strengthen security, and streamline cloud operations—worked alongside Toyota to build a thorough, step-by-step migration checklist for all North American production sites to minimize operational disruption and maintain resilience. As part of this plan, Toyota scheduled downtime at each facility to migrate the overarching system, and the teams performed rigorous pre- and post-testing measures, as well as system and organization-controls compliance checks.

“It was outstanding work from the Toyota and AWS teams to come together and make the migration successful,” says Patel. “The migration was done flawlessly.”

Toyota migrated its SAP workloads in under 2 weeks, including testing. Using Intel-powered AWS instances, the company improved performance, removed operational bottlenecks, and delivered better price performance. The finance team’s book-closing process has dropped from 72 hours to under 24 hours—a roughly 66 percent reduction—while the engineering team has seen 1.5-times performance gains over the previous architecture.

Looking ahead, Toyota is evaluating Amazon EC2 R8i and R8i-flex Instances as well as Amazon EC2 X8i instances, memory-optimized instances powered by custom Intel Xeon 6 processors that are available only on AWS. These instances represent Intel’s latest-generation custom silicon and the next step in an architecture that’s designed to scale with Toyota’s ambitions. 

Toyota is also using Kiro, an agentic coding service built on AWS, to solve challenging operational problems with agentic AI, including SAP operational health checks and enhanced development capabilities. The company is also supplementing its operational teams using AWS DevOps Agent, an autonomous AI agent that resolves and proactively prevents incidents.

With its new SAP architecture and future advances, Toyota’s finance and manufacturing teams work with confidence that their digital tools will remain reliably available. “Using these Intel-powered instances on AWS, we know our system is stable,” says Patel. “If an alert comes along, my team has plenty of time to address it.”