<p>Embodied Artificial Intelligence (AI) refers to AI systems in which intelligence is embedded in physical systems and emerges through interaction with its environment. Embodied AI (partly referred to as embedded AI also) acts in the real world through continuous cycles of sensing, decision-making, actuation, and learning. Embodied AI participates in operations and moves beyond supporting decision-making-support to a constitutive element of value creation. Firms must redesign what activities are performed, how they are linked, and who controls them. Embodied AI implies a double loop: a closed learning loop inside the adopting firm, where embodied AI transforms situated use into operational feedback and workflow changes, and an external learning loop across the ecosystem of technology providers, component suppliers, software firms, platform orchestrators, and users. Data generated through physical use travels beyond the adopting firm. Our conceptual study provides pioneering theoretical accounts of embodied AI in management research and focuses on its implications for business models. We develop two complementary models. First, a transition model shows how business models shift from asset-based and episodic logics toward adaptive, data-driven systems. Second, an integrative embodied AI business model grid explains how changes are generated through reconfiguration of value activities, interdependencies, and governance across actors and technologies. We derive nine propositions specifying how embodied AI transforms business models. We further show four systemic tensions: Openness versus control, scaling versus local fit, automation ambition versus reliability constraints, and monetization versus trust. Using agriculture as a revealing template, carves out how embodied AI reshapes business models in traditional industries moving from product performance toward continuous workflow optimization, lifecycle-based orchestration, and recurring, trust-based monetization.</p>