Health insurance costs from tobacco have been estimated from tobacco's contribution to annual or lifetime costs for heart disease, emphysema, and selected cancers. Because health plans seldom identify tobacco users, there are few published studies that compare users with nonusers. This study gathered data on paid claims from a large group's indemnity plan (4108 users, 16,723 nonusers) from Jan 1 to Nov 30, 1988. Tobacco users had more admissions per 1000 (124 v 76), days per 1000 (800 v 381), a longer average length of stay (6.47 v 5.03 days), higher average outpatient payments ($122 v $75), and higher average insured payments ($1,145 v $762). Tobacco use is correlated with other high-risk behaviors; thus, cost and utilization differences are not solely due to its effects. Nevertheless, tobacco users add to employer costs for health insurance as well as for absenteeism, workers' compensation, and life insurance. Employers may use these data to reduce costs by not hiring tobacco users, adding surcharges for their health insurance, and strongly encouraging cessation. Issues of equity are discussed in terms of coerciveness and intrusiveness.