Objectives We assessed whether changes in community viral load (CVL) over time were associated with the rate of new HIV diagnoses (NDs). Methods HIV‐1‐positive individuals referred to our institute and permanently residing in our province were considered for inclusion in the study. A total of 861 HIV‐infected adults with at least one HIV RNA measurement (12 530 measurements in total) between 2008 and 2014 were included. Viraemia copy‐years were calculated from all HIV RNA values for each patient using the trapezoidal rule; multiple CVL indicators were considered. Total NDs and recent infections (< 1 year) were analysed separately. The association between NDs and CVL was tested by means of mixed Poisson models, with CVL as a fixed effect and year as a random effect. Results The incidence of NDs was 2.28 per 100 000 residents in 2008 and 2.52 per 100 000 residents in 2014. Total numbers of NDs and recent infections did not vary significantly over time (P for trend 0.879 and 0.39, respectively). Mean HIV RNA decreased from 31 095.8 HIV‐1 RNA copies/mL in 2008 to 21 231.5 copies/mL in 2014 (P < 0.001); a downward trend was always observed regardless of the CVL indicator considered. Depending on the indicator, there were some differences in CVL by patient characteristics. The most substantial contributors to CVL appeared to be male individuals, men who have sex with men (MSM), non‐Italians, and untreated subjects (all P < 0.05). The relative risk of ND increased among Italians and MSM with an increasing proportion of subjects having an undetectable HIV RNA, and decreased in the same population with increasing levels of CVL. Conclusions In our setting, CVL represented a good marker of access to care and treatment; however, reduced CVL did not coincide with a reduction in the rate of NDs.