Digital twin
vs. simulation
The two terms are often used as if they meant the same thing. They do not. A simulation is a physics model of a generic system. A digital twin is a simulation that is kept in sync with one specific real machine through live data.
Simulation
A simulation answers "what would happen if?" for a model of a system. The model can be generic: a Unitree G1 robot in general, not your G1 with its worn left knee. You set up a scenario, run it, read the result. The physics simulators described elsewhere in this hub (see What is robotic simulation?) are simulations in this sense.
Digital twin
A digital twin is a simulation tied to one particular physical thing, continuously updated with data from it. The international standard for manufacturing, ISO 23247, defines a digital twin as "a fit-for-purpose digital representation of an observable manufacturing element with synchronization between the manufacturing element and its digital representation", as quoted by the US National Institute of Standards and Technology.[1] The key word is synchronisation.
IBM's definition adds the second property: a twin "uses real-time data to accurately reflect its real-world counterpart", and unlike a one-off simulation it stays live and can feed decisions back to the physical system.[5] Siemens describes its twins as combining "physics-based simulation with real-time operational data".[6] NVIDIA's glossary makes the same point from the other direction: simulation is one component of a twin, not the whole of it.[2]
Side by side
| Simulation | Digital twin | |
|---|---|---|
| Represents | A generic model | One specific asset |
| Data | Set up by the engineer | Streamed from sensors on the real asset[5] |
| Runs | A scenario, then stops | Continuously, alongside the asset[5] |
| Feedback | To the engineer | To the asset and its operators[5] |
In robotics
Two uses are common. First, a twin of a whole facility: NVIDIA's Mega blueprint, adopted by industrial companies in 2025, lets fleets of robots be tested in a twin of a warehouse before the real ones are switched on.[3] Second, twins of individual objects built from ordinary camera footage, so that a robot can practise on a copy of the exact item it will handle.[4]
A twin is only as truthful as the physics inside it. A factory twin that models motion but not motor temperature or battery charge will predict a shift the real robots cannot complete. That is where the extra physics from Invisible Force fits.
References
- Guodong Shao, US National Institute of Standards and Technology. Manufacturing Digital Twin Standards (quoting ISO 23247). September 2024.
- NVIDIA. What is a digital twin? (glossary). 2026.
- NVIDIA blog. Industrial ecosystem adopts Mega NVIDIA Omniverse Blueprint to train physical AI in digital twins. March 2025.
- Melnik et al., arXiv 2504.13159. Digital Twin Generation from Visual Data: A Survey. April 2025, revised February 2026.
- IBM. What is a digital twin?. August 2026.
- Siemens. Digital twin: the living blueprint. 2026.
Every link was opened and checked in September 2026. Links open in a new tab.