Democratising AI-powered digital twin adoption in manufacturing supply chains

Introduction

A digital twin is a virtual replica of a physical asset, process, or system that is continuously updated using real-time data from its physical counterpart. Until recently, digital twins were largely confined to large organisations with significant technical and financial resources. However, recent AI-enabled innovations in digital twin technology are rapidly reshaping its application across industries, moving from a niche, high-cost innovation to an increasingly accessible tool for operational resilience, sustainability, and strategic decision-making. Digital twins are now being reimagined as scalable, AI-enabled systems that SMEs can realistically adopt across production, warehousing, and logistics networks. Digital twins enable manufacturing organisations to simulate production lines, factory layouts, inventory flows, and logistics operations, allowing managers to test alternative scenarios, anticipate disruptions, and optimise performance before implementing changes in the real world. When enhanced with AI, digital twins evolve from static models into intelligent systems capable of learning from data, identifying inefficiencies, and supporting data-driven decision-making at scale.

Lowering barriers to digital twin adoption in manufacturing supply chains

Despite their potential, digital twins have seen limited adoption among manufacturing SMEs. High development costs, fragmented operational data, and the need for specialist modelling expertise have historically placed digital twins beyond the reach of smaller firms. These barriers are particularly significant given that SMEs form the backbone of manufacturing supply chains and are often the most exposed to operational disruptions, cost volatility, and climate-related risks. Recent advances in AI and generative AI are beginning to change this equation. New digital twin ecosystems are being designed to work with the types of data already used by manufacturing managers, such as spreadsheets, process maps, equipment lists, and basic factory layouts. Through natural language interfaces, users can now interact with complex simulations without requiring advanced technical skills. This democratisation allows SMEs to build and operate digital twins of factories, warehouses, and internal logistics processes using intuitive, low-cost, low-code/no-code tools rather than bespoke engineering solutions.

For manufacturing supply chains, this shift is transformative. It enables firms to simulate production bottlenecks, assess alternative facility layouts, test inventory policies, and evaluate the ripple effects of supplier delays or transport disruptions. Importantly, these simulations can also incorporate energy consumption, waste generation, and efficiency metrics, aligning operational optimisation with sustainability objectives.

Overcoming data integration challenges in supply chain digital twins

While the promise of digital twins is substantial, data integration remains a core challenge in manufacturing supply chains. Production systems, warehouse management platforms, transport tracking tools, sensors, and enterprise systems generate vast volumes of heterogeneous data. These datasets differ widely in structure, quality, and update frequency, making it difficult to integrate them into a single, reliable simulation environment.

Data quality issues such as missing records, sensor failures, or inconsistent formats require extensive preprocessing before they can be used for real-time simulation. This increases computational demands and can limit responsiveness in fast-moving manufacturing environments. Timing also presents challenges. Operational data often arrives continuously, while simulation models rely on discrete time steps or historical baselines. Aligning live manufacturing data with simulation timelines is technically complex but essential for ensuring that digital twins reflect real-world conditions accurately.

Governance and data security further shape adoption. Manufacturing supply chain data often includes commercially sensitive information related to suppliers, production capacities, and logistics routes. Ensuring appropriate data governance, access controls, and compliance is critical, particularly when digital twins are deployed across multi-tier supply networks.

OpenUSD, an open and extensible framework for describing and exchanging complex 3D scenes, plays a foundational role in addressing this challenge. By acting as a common data language, OpenUSD allows information from different tools, systems, and disciplines to be integrated into a single, coherent digital representation of manufacturing facilities and supply chain assets. This interoperability is essential for manufacturing supply chains, where digital twins must combine layout data, process flows, equipment specifications, and operational parameters across multiple functions and partners.

Building a high fidelity digital twin using NVIDIA Omiverse and OpenUSD

NVIDIA Omniverse platform builds on this foundation by providing a real-time simulation and collaboration platform for OpenUSD-based digital twins. For manufacturing supply chains, Omniverse enables high-fidelity simulation of factory operations, internal logistics, and material flows, while supporting real-time updates from operational data streams. The platform allows multiple stakeholders, including engineers, supply chain managers, and external partners, to interact with the same digital twin environment simultaneously, improving coordination and decision-making.

Together, NVIDIA Omniverse and OpenUSD significantly reduce the technical complexity and cost of building industrial-grade digital twins. They enable manufacturing firms, including SMEs, to move away from isolated, bespoke models towards scalable digital twin ecosystems that can evolve as supply chains grow, reconfigure, or face new risks. This interoperability is particularly important for supply chain resilience, where insights depend on understanding how changes in one part of the system propagate across production, storage, and distribution.

Digital twins and climate resilience in manufacturing supply chains

Climate change is intensifying risks across manufacturing supply chains, from heat-related productivity losses on factory floors to disruptions in warehousing, storage conditions, and outbound logistics. Digital twins provide a powerful mechanism for understanding how these risks interact with manufacturing operations over time. By integrating operational data with environmental and energy-related inputs, digital twins allow firms to explore how temperature extremes, energy constraints, or resource scarcity may affect production efficiency, equipment reliability, and supply continuity.

Within manufacturing environments, AI-powered digital twins can simulate heat stress impacts on machinery performance, workforce productivity, and energy demand, enabling proactive adaptation strategies. Similarly, warehouse and distribution twins can model how extreme temperatures influence storage conditions, cold-chain integrity, and material handling efficiency. This capability allows manufacturing supply chain managers to test mitigation measures such as process redesign, energy-efficient equipment, revised shift patterns, or alternative logistics configurations before disruptions occur.

Conclusion: an optimistic outlook for manufacturing supply chains

As artificial intelligence, open standards, and advanced simulation platforms continue to mature, digital twins are moving beyond experimental tools towards becoming essential infrastructure for manufacturing supply chains. For firms operating in climate-exposed and highly competitive environments, the ability to simulate, predict, and optimise supply chain performance in near real time is rapidly becoming a strategic necessity rather than a technological advantage.

Digital twins enable continuous monitoring of manufacturing operations, support rigorous “what-if” scenario analysis, and allow organisations to anticipate future disruptions before they materialise. The convergence of AI, generative AI, OpenUSD, and platforms such as NVIDIA Omniverse is lowering long-standing cost and complexity barriers that have historically limited adoption. At the same time, advances in cloud computing and GPU efficiency are making real-time or near-real-time supply chain simulations financially viable, extending the use of digital twins beyond planning phases into day-to-day operational decision-making.

Collaboration between academic institutions, industry partners, and government bodies is also playing a critical role in accelerating this transition. By combining research expertise, real-world operational data, and enabling policy and funding frameworks, research-led initiatives are demonstrating how AI-enabled digital twins can be developed as scalable, interoperable platforms grounded in real manufacturing contexts rather than isolated pilot projects.

For manufacturing supply chains, the question is no longer whether digital twins will become essential infrastructure, but how quickly organisations can adopt and govern them responsibly, inclusively, and at scale.

Acknowledgement
All authors are affiliated with the Digital Twin Futures Lab at Middlesex University Dubai. The authors acknowledge the support of the Dubai Future Foundation’s Research, Development and Innovation (RDI) Grant Initiative for the project “Cognitive Twins: Leveraging Artificial Intelligence and Omniverse-Powered Digital Twins for Transforming Dubai’s Manufacturing Sector” (Grant No. 2025/DRDI0176), which provided the foundation for this article.

Researchers

Sreejith Balasubramanian, Middlesex University Dubai; Leanne Braganza, Middlesex University Dubai; Huan Nguyen, Middlesex University London; Roberto Revetria, University of Genoa; and Michael Wagner, SyncTwin.

Dr Sreejith Balasubramanian is an Associate Professor and Head of the Centre for Supply Chain Excellence at Middlesex University Dubai. He can be contacted at s.balasubramanian@mdx.ac.ae.