Agentic AI increasingly enters everyday consumption such as online food ordering. This study investigates consumer acceptance of Agentic AI for food-ordering by extending UTAUT2 with Autonomy, Perceived Control, and Time Pressure. Analyzing 396 responses via PLS-SEM, results show that Hedonic Motivation, Performance Expectancy, Effort Expectancy, and Perceived Control significantly drive Habitual Usage Intention. Autonomy enhances both Perceived Control and Performance Expectancy, contradicting assumed control-loss effects, while Time Pressure shows no significant moderation. Findings extend UTAUT2 to Agentic AI, guiding its task capability, information presentation, and user control in food ordering.