Mycobacterium tuberculosis (Mtb) can infect macrophages as single or aggregated bacilli, where aggregate infection of macrophages was shown to have a substantially higher probability to result in macrophage death. Given that the response of macrophages to Mtb infection may determine the infection trajectory, it is important to understand the macrophage response to infection with Mtb aggregates. Here I investigated the early transcriptional response of monocyte derived macrophages (MDMs) to Mtb aggregate infection. I found that Mtb aggregates elicited the highest TNF-α and pro-inflammatory response relative to single Mtb bacilli. Additionally, aggregate-mediated MDM death was dependent on infection with live Mtb aggregates. I also investigated macrophage acidification in response to infection with Mtb aggregates and found that acidification, per Mtb bacillus, decreased as aggregate size increased. This suggests that Mtb aggregates have an advantage over single bacilli due to a weaker host response per mycobacterium. I also quantified Mtb aggregate number in human lung tissue sections using custom digital image analysis pipelines and developed a convolutional neural network (CNN) model, HyRoNet, to automate and expand the analysis. I found that Mtb aggregates occurred often, but not exclusively, in association with the granulomatous cavity surface. Together, these observations suggest a potentially important role for Mtb aggregation in the pathogenesis of Mtb.