Three enterprise programs are worth looking at together because they apply the same underlying idea to very different operations.
In March 2026, PepsiCo announced a collaboration with Siemens and NVIDIA that uses Siemens Digital Twin Composer and NVIDIA Omniverse to support factory and supply-chain planning. The system brings 3D models, operational data, simulation, and AI into a shared environment before changes reach a physical facility.
General Motors announced its work with NVIDIA in March 2025. Its manufacturing teams use Omniverse with production-line digital twins for virtual testing of material handling, transport, and precision welding processes.
Unilever took a different route. Its 2025 program with Collective World and NVIDIA uses physically accurate 3D product assets as a common source for product imagery and marketing content across markets.
What connects them
PepsiCo is applying the model to factory and supply-chain planning. GM is testing production systems virtually. Unilever is using product twins to support content production. The business case changes, but the common requirement is trustworthy 3D data that people can inspect, test, and reuse.
For our own digital-twin work, Unreal Engine often provides the visual and interactive layer while enterprise systems remain responsible for product, operational, or analytics data. That separation keeps the experience useful without turning the real-time application into another system of record.