BACKGROUND To explore the role of bevacizumab and a chemotherapy‐free approach, the authors evaluated the combination of bevacizumab, trastuzumab, and paclitaxel (HAT) and the regimen of trastuzumab and bevacizumab (HA) with the addition of paclitaxel after progression (HA‐HAT) as first‐line treatment for patients with human epidermal growth factor receptor 2 (HER2)‐positive metastatic breast cancer. METHODS In a noncomparative phase 2 trial, patients were randomized between HAT and HA‐HAT. The primary endpoint was the progression‐free rate at 1 year (1‐year PFR). In the HA‐HAT group, progression‐free survival (PFS) was separately established for HA (PFS1) and HAT (PFS2). RESULTS Eighty‐four patients received HAT (n = 39) or HA‐HAT (n = 45). The 1‐year PFR was 74.4% (95% confidence interval [CI], 61.8%‐89.4%) and 62.2% (95% CI, 49.6%‐89.4%) in the HAT and HA‐HAT arms, respectively. The median PFS was 19.8 months (95% CI, 14.9‐25.6 months) in the HAT arm and 19.6 months (95% CI, 12.0‐32.0 months) in the HA‐HAT arm. In the HA‐HAT arm, the median PFS1 was 10.4 months (95% CI, 6.2‐15.0 months), and the median PFS2 was 8.2 months (95% CI, 7.0‐12.6 months). The number and severity of adverse events were comparable between the arms. CONCLUSIONS Both HAT and HA‐HAT have promising activity in patients with HER2‐positive metastatic breast cancer. In particular, starting with only targeted agents and delaying chemotherapy is worth further exploration. Cancer 2016;122:2961‐2970 . © 2016 American Cancer Society .