Background Enucleation of pancreatic neuroendocrine tumours (pNETs) via robotic surgery has rarely been described. This study sought to assess the safety and efficiency of robotic surgery for the enucleation of small pNETs. Methods A comparison was conducted of enucleation of pNETs smaller than 2 cm by robotic or open surgery between January 2000 and May 2015. Propensity score matching was used to balance sex, age, BMI, tumour location and tumour diameter. Pathological results, safety‐related outcomes (postoperative pancreatic fistula (POPF) rate, estimated blood loss, and short‐term mortality and morbidity) and efficiency‐related outcomes (duration of surgery and postoperative length of hospital stay) were compared between the groups. Results A cohort of 120 patients with pNET were enrolled in the study (1 : 1 matched for open or robotic surgery, 60 per group). Ninety‐three patients (77·5 per cent) had a grade 1 tumour and 114 (95·0 per cent) had an insulinoma. Robotic surgery had a conversion rate of 5 per cent (3 of 60), and was not associated with an increased POPF rate (10 per cent versus 17 per cent after open surgery; P = 0·283) or grade III–V surgical complications according to the Dindo–Clavien classification (3 versus 10 per cent respectively; P = 0·272). Estimated blood loss was reduced with the robotic approach (32·5 versus 80·0 ml in the open group; P = 0·008), as was duration of surgery (117 versus 150 min; P < 0·001). Length of hospital stay after surgery was similar in the two groups (12·0 versus 13·5 days respectively; P = 0·071). Conclusion Robotic surgery for enucleation of pNETs smaller than 2 cm did not increase POPF or major complication rates, and reduced the duration of surgery and estimated blood loss, compared with open surgery. Registration number: NCT02125929 ( https://www.clinicaltrials.gov/). Robot beats surgeon