Political campaigns increasingly rely on digital ads to engage potential voters, with platforms like Meta offering advertisers detailed targeting options based on user demographics, behaviors, and interests. A lesser-known feature of social media advertising is the ad delivery algorithm, which sets ad prices through auctions and may deliver ads to ‘relevant’ audiences without explicit direction from advertisers. This automated process raises the question of whether it may introduce systemic biases into digital political advertising. To examine the pricing of political ads, we conducted a pre-registered study in collaboration with three Dutch political parties. We placed 135 identical ads on the parties’ official Facebook and Instagram accounts, targeting nine different audiences during the 2022 municipal elections. All ads ran simultaneously with identical settings, daily budgets, texts, and images, ensuring any differences in pricing and delivery were due to the advertiser and target audiences. We anticipated that parties would pay less when ads targeted ‘relevant’ audiences. However, we observed pricing differences between parties and audiences that did not always align with our expectations. For instance, one party paid 9.24% to 10.74% less per 1,000 users. Additionally, lower-educated citizens, women, and younger individuals (18–24) were more expensive to reach. These results have important implications. The unequal costs create an unfair and unequal playing field, while the exclusion of certain groups from ad delivery may limit their access to election information. Our findings suggest that simply restricting targeting capabilities, without addressing the role of ad delivery algorithms, would leave pricing disparities unchecked and without transparency.