Standard Physics (YC W26) — Fast physics simulations for rapid design and real time control

In their words

Standard Physics (formerly Inviscid AI) builds AI that does real physics. Most AI tools approximate physical behavior and get it wrong in ways engineers can’t see. Ours run actual numerical solvers, so the results are accurate enough to make real engineering decisions on. We have three products built on one engine. Reynolds runs computational fluid dynamics for you. An engineer describes the flow in plain language, and Reynolds writes the OpenFOAM case, builds the mesh, runs the solver, and returns the figures, the numbers, and a case they can rerun. It’s open source. Archimedes searches the design itself. Swarms of agents try thousands of layouts in parallel, simulate each one through Reynolds, keep what the physics rewards, and redraw the search space when the best answer was never in the original plan. We sell it as outcome-based engineering: customers pay for a better design, not for compute. Inviscid AI (now a product!) points the same engine at a running building. It holds a data center within its thermal limits in real time, computed continuously from live telemetry. Our platform optimizes airflow patterns and ventilation strategies to eliminate dead zones, improve air distribution, and reduce the load on mechanical systems. On the energy side, we minimize HVAC power consumption, reduce cooling costs, and lower overall operational expenses while maintaining optimal thermal comfort and indoor air quality. We work with hardware startups and engineering teams on data center cooling, drones, energy systems, piping, and electronics. Reynolds earns engineers’ trust in daily use, and Archimedes and Inviscid AI turn that trust into large contracts.

Founders, titles and bios as the company published them. Photographs are our own copies of the ones on that record, so they do not break when the original link expires.