Background Long-lasting insecticidal nets (LLINs) were once fully effective for the prevention of malaria; however, mosquitoes have developed resistance to pyrethroids, the main class of insecticides used on nets. Dual active ingredient LLINs (dual-AI LLINs) have been rolled out as an alternative to pyrethroid (PY)-only LLINs to counteract this. Understanding the minimum community usage at which these novel nets generate an effect that also benefits non-net users against malaria infection is vital for planning net distribution strategies and mobilization campaigns. Methods We conducted a secondary analysis of a 3-year randomized controlled trial (RCT) in 84 clusters in North-western Tanzania to evaluate the effectiveness of three dual-AI LLINs: pyriproxyfen and alpha(α)-cypermethrin (pyriproxyfen-PY), chlorfenapyr and α-cypermethrin (chlorfenapyr-PY), and the synergist piperonyl-butoxide and permethrin (PBO-PY) compared to α-cypermethrin only nets (PY-only). We measured malaria infection prevalence using 5 cross-sectional surveys between 2020 and 2022. We assessed net usage at the cluster level and malaria infection in up to two children aged between 6 months and 14 years in 45 households per cluster and compared infection prevalence between net users and non-users with the different net types and usage levels. Findings A total of 22,479 children from 12,654 households were tested for malaria using rapid diagnostic tests in January 2020, 2021, & 2022 and July 2020 & 2021. In all surveys combined, 23% (5,062/22,479) of children reported not using a net the night before the surveys. The proportion of non-net users was highest in the later surveys. Across all study arms and at each time point, users of nets had significantly lower malaria infection than non-users. Overall, malaria prevalence was 52% (2649/5062) among non-net users and 32% (5572/11845) among users (of any net). Among non-net users, community-level usage of >40% of dual-AI LLIN was significantly associated with protection against malaria infection: chlorfenapyr-PY (OR: 0.44 (95% CI: 0.27-0.71), p=0.0009), PBO-PY (OR: 0.55 (95% CI: 0.33-0.94), p=0.0277) and pyriproxyfen-PY (OR: 0.61 (95% CI: 0.37-0.99), p=0.0470) compared with non-users in clusters with >40% usage of PY-only LLINs. There was weak evidence of protection against malaria infection to non-net users in the chlorfenapyr-PY arm when community-level usage was ≤40% (OR: 0.65 (95% CI: 0.42-1.01), p=0.0528) compared to those living in clusters with >40% usage of pyrethroid-only LLINs. The study was limited to non-users which were defined as participants who did not sleep under any net the night before. This might not capture occasional net usage during the week. Conclusion Our study demonstrated that at a community usage of 40% or more of dual-AI LLINs, non-net users benefited from the presence of these nets. Noticeably, even when usage was ≤40% in the chlorfenapyr-PY arm, non-users were better protected than non-users in the higher coverage PY-only arm. The greater difference in malaria risk observed between users and non-users across all study arms indicates that nets play a crucial role in providing personal protection against malaria infection for the people using the net and that net usage needs to be maximized to realize the full potential of all nets.