Anaemia is highly prevalent at the time of intensive care unit discharge and is persistent for a high proportion of intensive care unit survivors. Whether anaemia is a driver of impaired recovery after critical illness is uncertain. The aim of this study was to test the hypothesis that, in adult intensive care survivors, anaemia at the time of intensive care unit discharge independently predicts decreased days at home-90. This retrospective cohort study was conducted in a tertiary intensive care unit in Perth, Western Australia. All patients aged ≥ 16 years, discharged alive from their index intensive care unit admission and without documented treatment limitations were included. Median (IQR [range]) age of the 6358 participants was 61 (46–72 [16–95]) years and included 3385 (53.2%) unplanned admissions. Intensive care unit discharge with a haemoglobin concentration < 100 g.l-1 occurred in 2886 (45.4%) patients, a threshold that identified a cohort with significantly lower days at home-90 (median (IQR [range]) 80 (64–85 [0–90]) days vs. 85 (77–88 [0–90]) days (median difference 5 days, 95%CI 4.4–5.5, p < 0.0001). The association followed a severity-response relationship with more severe anaemia predicting lower days at home-90. When accounting for prespecified covariates including admission haemoglobin concentration and red blood cell transfusion, anaemia at intensive care unit discharge remained a significant predictor of decreased days at home-90, relative risk 0.96 (0.93–0.98), p < 0.002. These findings support the need for interventional trials investigating whether this risk is modifiable.