Background: Tumor budding is a strong independent prognostic factor in primary colorectal cancer (CRC). The prognosis of stage IV CRC is still poor and there is need for biomarkers to facilitate the clinical management. Data on tumor budding in colorectal cancer distant metastases (CRDM) are still missing. Therefore, the aim of this study was to analyze its role in tumor progression. Methods: 73 stage IV CRC with clinico-pathological data were retrieved from a cohort of 331 CRC patients, surgically treated between 2002 and 2013 at the University Hospital of Bern. Tumor budding was visualized immunohistochemically by pan-cytokeratin staining. Tumor budding was defined as intra- and peri-tumoral budding in the primary tumor (ITB = buds in the tumor center; PTB = buds at the tumor margin), intra- and perinodal budding (INB and PNB) and intra- and peri-metastatic tumor budding (IMB and PMB). Overall tumor budding was defined as tumor buds independent of the center and margin (OTB, ONB, OMB). The tumor bud count was assessed by calculating the mean number of tumor buds in 10 high power fields. For survival analysis a cut off of 10 tumor buds was applied to subdivide the tumors with low and high numbers of tumor buds. Results: In lymph nodes and CRDM, the tumor bud count was lower compared to primary tumors (PNB: 13; INB: 18; ONB: 21 vs PMB: 10; IMB: 21; OMB: 22 vs PTB: 23; ITB: 24; OTB: 33). For PTB/PNB and OTB/ONB, there were significant differences between primary tumor and lymph node metastasis (PTB/PNB: p < 0.001; OTB/ONB: 0.008) and primary tumor and distant metastases (PTB/PMB: p < 0.001; OTB/OMB: 0.007). The presence of PMB was predicted by INB, PNB and V1 (p = 0.032, p = 0.001 and p = 0.003). A trend towards a worse prognosis in patients with a high OMB number was observed. Conclusions: Tumor budding is present in lymph node and distant metastases of CRC, but its number is lower compared to the corresponding primary tumor. As tumor budding in local and distant metastases seems to be a predictor of tumor progression, these novel data should be considered as a basis for further analysis in large and multicentric clinical trials.