Botnets are a serious threat on the Internet and require huge resources to be thwarted. ISPs are in the best position to fight botnets and there are a number of recently proposed initiatives that focus on how ISPs should detect and remediate bots. However, it is very expensive for ISPs to do it alone and they would probably welcome some external funding. Among others, botnets severely affect ad networks (ANs), as botnets are increasingly used for ad fraud. Thus, ANs have an economic incentive, but they are not in the best position to fight botnet ad fraud. Consequently, ANs might be willing to subsidize the ISPs to do so. We provide a game-theoretic model to study the strategic behavior of ISPs and ANs and we identify the conditions under which ANs are likely to solve the problem of botnet ad fraud by themselves and those under which the AN will subsidize the ISP to achieve this goal. Our analytical and numerical results show that the optimal strategy depends on the ad revenue loss of the ANs due to ad fraud and the number of bots participating in ad fraud.