There is a lack of evidence for the optimal type of anesthesia technique in patients ⩾ 90 years with acute ischemic stroke undergoing mechanical thrombectomy (MT) as this subgroup of patients was often excluded or under-represented in previous trials. We aimed to compare outcomes between general anesthesia (GA) and non-GA techniques in patients ⩾ 90 years with large vessel occlusion (LVO) undergoing MT. Our study included patients ⩾ 90 years with anterior circulation LVO, NIHSS ⩾ 6, ASPECTS ⩾ 5 consecutively treated with MT within 6 h after stroke onset in three thrombectomy capable centers between January 1st, 2016 and March 30th, 2023. Inverse probability weighting (IPW) was used to reduce bias by indication of the anesthesia type on study outcomes. We used a weighted ordinal robust logistic regression analysis to explore the primary outcome of modified Rankin Scale (mRS) shift at 90 days in GA versus non-GA treated patients. Secondary outcomes included 90-day mortality, symptomatic intracranial hemorrhage (sICH) and TICI score of 2b, 2c, or 3. We included 139 patients ⩾ 90 years treated with MT, 62 were in GA group and 77 in non-GA group. There was a significant shift for worse mRS scores at 90-day in non-GA treated patients (cOR 3.65, 95% CI 1.77-7.77, p = 0.001). The weighted logistic regression showed that non-GA technique was an independent predictor of 90-day mortality (OR 7.49, 95% CI 2.00-28.09; p = 0.003). Our study indicated that nonagenarians with acute ischemic stroke treated with MT without GA have a worse prognosis than their counterparts undergoing MT with GA. Further studies in larger cohorts are warranted to evaluate the optimal type of anesthesia in this patient population.