Background: The development of anti-HER2 antibody–drug conjugates opens new therapeutic options for patients with breast cancer, including patients with low expression of HER2. To characterise this new breast cancer subtype, we have compared the clinical and molecular characteristics of HER2-low-positive and HER2-zero breast cancer, including response to neoadjuvant chemotherapy and prognosis. Methods: In this pooled analysis of individual patient data, we evaluated a cohort of 2310 patients with HER2-non-amplified primary breast cancer that were treated with neoadjuvant combination chemotherapy in four prospective neoadjuvant clinical trials (GeparSepto, NCT01583426; GeparOcto, NCT02125344; GeparX, NCT02682693; Gain-2 neoadjuvant, NCT01690702) between July 30, 2012, and March 20, 2019. Central HER2 testing was done prospectively before random assignment of participants in all trials. HER2-low-positive status was defined as immunohistochemistry (IHC) 1+ or IHC2+/in-situ hybridisation negative and HER2-zero was defined as IHC0, based on the American Society of Clinical Oncology/College of American Pathologists guidelines. Disease-free survival and overall survival data were available for 1694 patients (from all trials except GeparX) with a median follow-up of 46·6 months (IQR 35·0–52·3). Bivariable and multivariable logistic regression models and Cox-proportional hazards models were performed based on a predefined statistical analysis plan for analysis of the endpoints pathological complete response, disease-free survival, and overall survival. Findings: A total of 1098 (47·5%) of 2310 tumours were HER2-low-positive and 1212 (52·5%) were HER2-zero. 703 (64·0%) of 1098 patients with HER2-low-positive tumours were hormone receptor positive, compared with 445 (36·7%) of 1212 patients with HER2-zero tumours (p<0.0001). HER2-low-positive tumours had a significantly lower pathological complete response rate than HER2-zero tumours (321 [29·2%] of 1098 vs 473 [39·0%] of 1212, p=0·0002). Pathological complete response was also significantly lower in HER2-low-positive tumours versus HER2-zero tumours in the hormone receptor-positive subgroup (123 [17·5%] of 703 vs 105 [23·6%] of 445, p=0·024), but not in the hormone receptor-negative subgroup (198 [50·1%] of 395 vs 368 [48·0%] of 767, p=0·21). Patients with HER2-low-positive tumours had significantly longer survival than did patients with HER2-zero tumours (3-year disease-free survival: 83·4% [95% CI 80·5–85·9] vs 76·1% [72·9–79·0]; stratified log-rank test p=0·0084; 3-year overall survival: 91·6% [84·9–93·4] vs 85·8% [83·0–88·1]; stratified log-rank test p=0·0016). Survival differences were also seen in patients with hormone receptor-negative tumours (3-year disease-free survival: 84·5% [95% CI 79·5–88·3] vs 74·4% [70·2–78.0]; stratified log-rank test p=0·0076; 3-year overall survival: 90·2% [86·0–93·2] vs 84·3% [80·7–87·3], stratified log-rank test p=0·016), but not in patients with hormone receptor-positive tumours (3-year disease-free survival 82·8% [79·1–85·9] vs 79·3% [73·9–83·7]; stratified log-rank test p=0·39; 3-year overall survival 92·3% [89·6–94·4] vs 88·4% [83·8–91·8]; stratified log-rank test p=0·13). Interpretation: Our results show that HER2-low-positive tumours can be identified as new subgroup of breast cancer by standardised IHC, distinct from HER2-zero tumours. HER2-low-positive tumours have a specific biology and show differences in response to therapy and prognosis, which is particularly relevant in therapy-resistant, hormone receptor-negative tumours. Our results provide a basis for a better understanding of the biology of breast cancer subtypes and the refinement of future diagnostic and therapeutic strategies. Funding: German Cancer Aid (Deutsche Krebshilfe).