To evaluate post-nephrectomy outcomes and predictors of cancer-specific survival (CSS) between patients with localised sarcomatoid renal cell carcinoma (sRCC) and those with Grade 4 RCC (non-sRCC), as most sRCC research focuses on advanced or metastatic disease with limited studies analysing outcomes of patients with localised non-metastatic sRCC. A total of 564 patients with localised RCC underwent partial or radical nephrectomy between June 1988 to March 2019 for sRCC (n = 204) or World Health Organization/International Society of Urological Pathology Grade 4 non-sRCC (n = 360). The CSS at every stage between groups was assessed. Phase III ASSURE clinical trial data were used to externally validate the CSS findings. The Mann-Whitney U-test and chi-squared test compared outcomes and the Kaplan-Meier method evaluated CSS, overall survival (OS) and recurrence-free survival. Clinicopathological features associated with RCC death were evaluated using Cox proportional hazards regression. The median follow-up was 31.5 months. The median OS and CSS between the sRCC and Grade 4 non-sRCC groups was 45 vs 102 months and 49 vs 152 months, respectively (P < 0.001). At every stage, sRCC had worse CSS compared to Grade 4 non-sRCC. Notably, pT1 sRCC had worse CSS than pT3 Grade 4 non-sRCC. Negative predictors of CSS were sarcomatoid features, non-clear cell histology, positive margins, higher stage (pT3/pT4), and use of minimally invasive surgery (MIS). ASSURE external verification showed worse CSS in patients with sRCC (hazard ratio [HR] 1.63, 95% confidence interval [CI] 1.12-2.36; P = 0.01), but not worse outcomes in MIS surgery (HR 1.39, 95% CI 0.75-2.56; P = 0.30). Localised sRCC had worse CSS compared to Grade 4 non-sRCC at every stage. Negative survival predictors included positive margins, higher pathological stage, use of MIS, and non-clear cell histology. sRCC is an aggressive variant even at low stages requiring vigilant surveillance and possible inclusion in adjuvant therapy trials.