As artificial intelligence infrastructure scales rapidly, the complexity of designing and delivering large data centres is increasing. In response, digital simulation tools are beginning to play a more central role in how such facilities are planned and operated.
Jacobs, a global engineering and consulting firm, has introduced a Data Centre Digital Twin solution aimed at enabling developers and owners to plan, simulate and optimise gigawatt-scale AI data centres within a virtual environment. The approach brings together design, engineering and operational considerations into a single digital framework.
A virtual model for complex infrastructure planning
The solution has been developed using the NVIDIA Omniverse DSX blueprint and combines a standardised reference design for gigawatt-scale AI data centres with a highly detailed virtual environment.
Within this environment, advanced simulations model key systems including compute, power and cooling. The platform functions as a real-time virtual representation of a data centre, allowing stakeholders to assess performance and coordination before physical development begins.
It also enables secure integration of components both within and beyond the facility, covering aspects such as on-premises power configurations and indoor airflow simulations. By bringing together compute, power, cooling, water systems and site-specific inputs, the model provides visibility across multiple stages of a project.
Extending visibility across the asset lifecycle
The digital twin is designed to support decision-making from early planning through to virtual commissioning, operations and maintenance. This allows project teams to track and evaluate system performance at different stages, rather than relying solely on post-construction adjustments.
According to Jacobs, the increasing scale and cost of advanced facilities is driving the need for greater certainty prior to capital commitment. The digital model is intended to support this by enabling more detailed planning and performance analysis in advance of construction.
The first module of the solution is currently available for owners and operators, with additional functionality expected to extend its application across the full asset lifecycle, including design, construction and commissioning phases.
Modular approach targets wider industry adoption
A key component of the solution is its modular reference design, which supports repeatable deployment of large-scale facilities. This standardisation is intended to assist in the delivery of gigawatt-scale data centres in the physical environment.
Future development of the platform is expected to include configurations for smaller, 250MW facilities, enabling broader applicability across different project scales.
The digital twin forms part of the wider NVIDIA Omniverse DSX framework, which focuses on creating and operating large-scale AI data centre infrastructure through integrated simulation and modelling tools.
Jacobs is currently involved in delivering high-performance computing environments for organisations including Hut 8 and PsiQuantum, reflecting its role within advanced technology sectors such as data centres, semiconductors and specialised manufacturing.
