DraftKings Says It Uses Machine Learning to Target Likely Losing Bettors

DraftKings disclosed on September 19 that it applies machine learning to customer betting behavior in order to direct promotional offers toward bettors identified as likely to lose. According to former employees, parallel work on using similar techniques to flag customers at risk of problem gambling appears to have been delayed or shelved.

The disclosure arrives during the sportsbook's most important stretch of the year. The 2026 NFL season opened September 9, and Citizens recently raised its price target to $37 from $36 on an Outperform rating, citing expectations for record season betting activity. The company's second quarter showed 73% year-over-year customer growth at an 8% lower acquisition cost, with a rapidly scaling predictions segment.

The stock closed at $24.33 on September 16, down 1.26% on the session and below an average analyst forecast near $34. The gap between operating momentum and share price has narrowed the bull case to execution, and the targeting disclosure introduces a regulatory and reputational variable that was not previously in the model. The next scheduled catalyst is the November 5 earnings report.