• Mazda consolidates design and development data into a unified storage platform, establishing the foundation for a future AI and generative AI data lake
• Storage cost per unit drops 90% while total capacity expands from 4PB to 10PB

Dell Technologies announced that Mazda Motor Corp. is building a data foundation to accelerate AI-driven automotive development.The automaker has deployed Dell PowerScale, part of the Dell AI Data Platform, to unify its model-based development and CAD storage environments into a scalable infrastructure designed to evolve into a data lake supporting AI and generative AI workloads.

Why it matters
Mazda has relied on model-based development for decades, using mathematical formulas and physical models to move vehicles from concept to the road faster. As advanced driver assistance systems have grown more complex, simulation data has increased alongside them, expanding by several hundred terabytes annually. Managing that volume using a storage and tape system approach created retrieval delays, operational complexity and unsustainable costs.

Data collection is also one of the most significant barriers to effective AI adoption. Mazda’s legacy infrastructure was not built to consolidate or scale the rising volumes of engineering data required for modern AI workloads. By consolidating decades of engineering data on PowerScale, Mazda can surface new insights and establish a foundation for advanced AI capabilities – in an increasingly software-defined automotive industry. Dell’s prior supercomputer deployment for Mazda analysis workloads built the trust that made this next step possible.

An integrated platform for scale and performance
Mazda chose Dell PowerScale due to its performance and scale to handle two distinct requirements within a single scale-out NAS solution: large capacity for model-based development data and high performance for CAD workloads. 

Dell’s InsightIQ provides unified AIOps-driven observability across the environment, giving Mazda’s team a single view of storage health and utilization. Meanwhile, SnapshotIQ protects against accidental data deletion and SyncIQ supports inter-chassis replication for business continuity.

Economics that validate the decision
The infrastructure went into full operation in December 2025. Total storage capacity expanded from approximately 4PB to 10PB, eliminating dependence on tape media for data offload. Recurring issues around capacity shortages and performance degradation were resolved.

The consolidation reduced storage costs per unit by 90%^[i]. Operational management burden dropped significantly as capacity allocation inquiries and performance-related tickets decreased. Critical design and development data is now protected through automated replication.

The road ahead
Mazda plans to use its PowerScale, part of the Dell AI Data Platform, to support a data lake for AI workloads. The consolidated environment gives Mazda’s engineering teams direct access to decades of accumulated design and development data. That access is what makes it possible to train models, surface patterns and accelerate innovation across Mazda’s end-to-end vehicle development pipeline. Dell PowerScale is the foundation that makes that ambition possible.

Perspectives
“The volume and complexity of data in modern vehicle development demands infrastructure that can scale with your ambitions, not just your current workload,” said Yuichi Tetsumoto, staff manager, engineering systems, MDI and IT headquarters, Mazda. “PowerScale gave us the foundation to consolidate decades of engineering data and position Mazda to put AI and generative AI to work across our development pipeline.”

“Mazda has 30 years of simulation data shaping automobile design. The challenge was never storing that data, it was unlocking it,” said Travis Vigil, SVP, product management, infrastructure solutions group, Dell Technologies. “PowerScale delivers the performance to feed workloads like CAD today and the capabilities to power AI tomorrow, helping Mazda turn its data into fuel for what’s next.”