Case Study: Pong Yuen achieves 6x AI training performance and faster time-to-market with NetApp

Pong Yuen uses AI and the metaverse to improve decision making

Pong Yuen Holdings, founded in 1991, is a specialist in AIoT and machine‑learning solutions that builds metaverse simulations to help clients predict outcomes and manage risk—such as virtual highways for training autopilot models and simulated building fires for detection and evacuation. To support these compute‑ and data‑intensive use cases, the company needed a high‑performance, reliable AI infrastructure that could speed analytics, training and inference while handling large datasets, many model versions, and reducing time‑to‑market.

Pong Yuen adopted NetApp ONTAP AI with NVIDIA DGX A100 hardware and the NetApp AI Control Plane, leveraging NetApp Snapshot and FlexClone and all‑flash storage to simplify deployment and data management. The solution eliminated complex design work, enabled near‑instant dataset and model cloning (seconds instead of hours), delivered multi‑GB/s throughput and markedly faster training (up to sixfold), lowered space/power and training costs, and provided a secure, scalable platform for faster, more resilient ML operations.

Pong Yuen

Anri Kitami

Software Engineering Specialist