The primary use case of Newton Physics is reinforcement learning, with 1000s of similar environments. Even if each environment would have sequential actions, you run many envs in parallel.