Cefiderocol may represent a therapeutic option for carbapenem-resistant Acinetobacter baumannii (CRAB) infections, but clinical data are limited. This is an observational retrospective study conducted in the University Hospital of Pisa including consecutive patients with CRAB infections (January 2020 to August 2021). Patients were divided in two study groups according to the antibiotic treatment received: cefiderocol- and colistin-containing regimens. The primary outcome was the 30-day mortality. A Cox regression analysis was performed to identify factors independently associated with 30-day mortality. A propensity score analysis using inverse probability of treatment weighting (IPTW) was also performed. A total of 124 patients were included: 47 (37.9%) received cefiderocol, while 77 (62.1%) colistin-containing regimens. Overall, 79 (63.7%) patients had a bloodstream infection (BSI), 35 (28.5%) a ventilator-associated pneumonia (VAP) and 10 (8.1%) other infections. Thirty-day mortality was higher in patients receiving colistin- compared to those who received cefiderocol-containing regimens (55.8% versus 34%, P = 0.018). This difference was confirmed in patients with BSI, but not in those with VAP. On multivariable analysis, septic shock, SOFA score, and age were independently associated with 30-day mortality, while cefiderocol therapy was protective in an IPTW analysis (Hazard ratio 0.44, 95% confidence interval 0.22–0.66, P < 0.001). Nephrotoxicity was more common in the colistin group. Microbiological failure occurred in 17.4% of patients receiving cefiderocol versus 6.8% of those receiving colistin (P = 0.079). Among 8 cases in the cefiderocol group who experienced microbiological failure, 4 (50%) developed resistance to cefiderocol. Cefiderocol represents a promising therapeutic option in patients with severe CRAB infections. Randomized clinical trial in this specific patient population should confirm our findings.