AI glossary · Basic · Robotics & physical AI · AI platforms & solutions · Manufacturing · Energy
Digital twin
A digital twin is a virtual model inside a computer that replicates a real object such as a factory, equipment, a city or a power grid. It continuously receives data from the real thing to track its state and is used to test things in advance.
About this term
- Level
- Basic · Technical terms you often meet in AI news
- Category
- Industry AI
- Industries
- Manufacturing, Energy
- Added
- 2026-10-10
How it works
Data such as temperature, operating status and output is received from sensors and operational systems attached to the real object and reflected in the virtual model. The virtual model mimics the real object's behavior using simulations based on physical laws or AI models trained on data. Operators first test "what happens if we change this setting" in the virtual space and apply the options that produce good results to the real object. The results after applying them come back in as data, so the model keeps being adjusted to stay in line with the real object.
The real–virtual loop
Keeping real and virtual in sync
- 1Collect real-world data
- 2Update virtual model
- 3Test virtually
- 4Apply to real object
- 1. Collect real-world data Sensors, operational systems
- 2. Update virtual model Track real-world state
- 3. Test virtually Try changing settings
- 4. Apply to real object Human review and approval
Why it matters
Changing a factory line layout or how a power grid is operated is costly and risky to test on the real thing. With a digital twin, such changes can be tested in advance without shutdowns or accidents. Because it shows equipment status at a glance in real time, it also serves as a foundation for predictive maintenance and energy use optimization. Its use is expanding beyond manufacturing to fields that deal with large systems, such as energy, cities and environmental management.
Things to know
A simple 3D view or a one-off simulation is not continuously connected to real-world data, so it should be distinguished from a digital twin. If the real object and the virtual model drift apart, the twin can lead to wrong conclusions, so sensor quality and model validation are key. Twins of infrastructure such as power grids and plants contain facility structure and operational information as is, so security and access control are important. When applying options found in the virtual model to real equipment, it is common to have a process in which people review and approve them according to safety regulations.
Example
A factory builds a digital twin of its entire production line, and before starting production of a new product, it changes the equipment order and work speed on the virtual line to find where bottlenecks occur. It picks the best layout, applies it to the real line, then compares actual operating data with the virtual model's predictions and readjusts the model.