Why Now?

In 2026, the problem we work on has been named by others. On 21st September 2026, NVIDIA wrote that safety for physical AI must extend across every layer, from the sensing a machine relies on to the actions it takes, and that validating it at scale needs simulation and synthetic data. NIST is treating physical AI deployment as a measurement and validation problem. Research published in 2026 separates trust in a digital representation of a system from the integrity of the data behind it.

*None of this tests or endorses our product. It is recognition that, before an automated system acts on a claimed physical state, someone has to ask whether that state could actually be true. That is the question Aequus asks, and we say which reading cannot be trusted.