Rationale: The incidence of influenza-associated acute respiratory failure is unknown. Objectives: We conducted this study to estimate the population-based incidence of influenza-associated acute respiratory failure hospitalizations. Methods: This is a cohort study from January 2003 through March 2009 using hospitalization databases for Arizona, California, and Washington from the Healthcare Cost and Utilization Project and influenza surveillance data for regions encompassing these states. Acute respiratory failure requiring mechanical ventilation was defined by ICD-9-CM code. We used negative-binomial regression modeling to estimate the incidence of influenza-associated events. Measurements and Main Results: The incidence of influenza-associated acute respiratory failure was 2.7 per 100,000 person-years (95% CI 0.2, 23.5), and during the influenza season, 3.8% of all respiratory failure hospitalizations were attributable to influenza. Compared with adults aged 18-49 years, the incidence rate ratio (IRR) for influenza-associated acute respiratory failure was lower among children aged 1-4 years (0.9) and 5-17 years (0.3). However, the IRR was higher among adults aged 50-64 years (4.8), 65-74 years (10.4), 75-84 years (19.9), and 85 years and older (33.7). Results were similar with more sensitive and specific outcome definitions and in a sensitivity analysis using only Arizona-specific outcome and surveillance data. Conclusion: Our data indicate that influenza was an important contributor to respiratory failure hospitalizations during 2003-2009. Physicians should consider influenza testing and empiric antiviral therapy for hospitalized patients with severe acute respiratory disease during periods of influenza activity. Influenza has a greater effect on respiratory failure in the elderly, for whom better prevention measures are needed.