Viral blips, where HIV RNA plasma viral load (pVL) intermittently increases above the lower limit of assay detection, are a cause for concern. We investigated a number of hypotheses for their cause. We assessed HIV RNA, and total and episomal HIV DNA from 16 individuals commencing antiretroviral therapy (ART) consisting of raltegravir and tenofovir/emtricitabine for 3 years, using two assays: a single-copy assay [SCA; lower limit of quantification (LLOQ), <1 copy/ml] and the Amplicor assay (LLOQ of 50 copies/ml). Two individuals exhibited viral blips. From week 20 onward, the period where ART had achieved its final suppressive levels, pVL ranged from <1 to 330 copies/ml, except for one individual at the final time. Both assays were 98% consistent (108/110) in assessing pVL <50 copies/ml, but the Amplicor assay registered 56% of samples (19/34) as below the LLOQ that were in the 50 to 1000 copy/ml range as quantified by SCA. pVL changes between successive time points did not correlate with changes in cellular infection as measured through either total or episomal HIV DNA. Changes in pVL were correlated (negatively) with changes in total CD4(+) T cell numbers (p=0.003), naive (CD45RO(-)CD62L(+)CD4(+)), natural regulatory (CD45RO(-)CD25(+)CD127(-)CD4(+)), activated effector (CD45RO(+)CD38(++)CCR5(+)CD8(+)), but not activated (CD38(+)HLA-DR(+)) CD4(+) T cells. Patients receiving stable, seemingly suppressive ART can have pVL near the 50 copy LLOQ at multiple time points. The high Amplicor assay error rate around this level implies that viral blips underrepresent pVL being more consistently above the LLOQ. Activation of latently infected cells is less likely to contribute to this phenomenon.