KinetIQ

KinetIQ is Humanoid's proprietary four-layer AI framework for end-to-end orchestration of humanoid fleets, with a single cross-embodiment model controlling robots of different morphologies across simultaneous timescales. System 3 is an agentic fleet orchestrator that assigns fleet-level goals and reacts within seconds. System 2 is robot-level reasoning via an omni-modal LLM operating on a second-to-subminute timescale, decomposing goals into sub-tasks. System 1 is a vision-language-action model handling low-level execution (5-10 Hz predictions, 30-50 Hz action chunks). System 0 is reinforcement-learning whole-body control at 50 Hz, trained solely in simulation and covering both bipedal and wheeled platforms. Introduced in February 2026 and demonstrated in a multi-robot showcase at NVIDIA GTC; a reinforcement-learning approach called KinetIQ Ascend was announced in June 2026. The company reports a 42% throughput gain in a machine-feeding application, with robots running at 1.5x human-demonstration speed.

$270M raised; $152M Series A at a $1.35B valuation (July 2026)