Amyotrophic Lateral Sclerosis (ALS) is a fatal neurodegenerative disease characterized by the progressive loss of motor neurons. While several inherited pathogenic mutations have been identified as causative, the vast majority of cases are sporadic with no family history of disease. Thus, for the majority of ALS cases, a specific causal abnormality is not known and the disease may be a product of multiple inter-related pathways contributing to varying degrees in different ALS patients. Using unsupervised machine learning algorithms, we stratified the transcriptomes of 148 ALS decedent cortex tissue samples into three distinct and robust molecular subtypes. The largest cluster, identified in 61% of patient samples, displayed hallmarks of oxidative and proteotoxic stress. Another 20% of the ALS patient samples exhibited high levels of retrotransposon expression and other signatures of TDP-43 dysfunction. Finally, a third group showed predominant signatures of glial activation (19%). Together these results demonstrate that at least three distinct molecular signatures contribute to ALS disease. While multiple dysregulated components and pathways comprising these clusters have previously been implicated in ALS pathogenesis, unbiased analysis of this large survey demonstrated that sporadic ALS patient tissues can be segregated into distinct molecular subsets.