Impaired type I interferons (IFNs) production or signaling have been associated with severe COVID-19, further promoting the evaluation of recombinant type I IFNs as therapeutics against SARS-CoV-2 infection. In the Syrian hamster model, we show that intranasal administration of IFN-α starting one day pre-infection or one day post-infection limited weight loss and decreased viral lung titers. By contrast, intranasal administration of IFN-α starting at the onset of symptoms three days post-infection had no impact on the clinical course of SARS-CoV-2 infection. Our results provide evidence that early type I IFN treatment is beneficial, while late interventions are ineffective, although not associated with signs of enhanced disease. Author summary: Type I interferons are major antiviral effectors produced by the host in response to viral infections. Importantly, delayed or impaired type I IFN signalling response has been shown to correlate with severe COVID-19. These observations provided further impetus to test the administration of exogenous type I IFN as a treatment against SARS-CoV-2 infection in patients. However, studies using MERS-CoV or SARS-CoV infected mice demonstrated that type I interferon treatment was beneficial when administered early, but was ineffective and even caused deleterious immunopathology when administered at later stages of infection. It is therefore crucial to understand how the timing of the type I IFN treatments modulates their efficacy and safety against SARS-CoV-2. In this preclinical study using the SARS-CoV-2-infected Syrian hamster model, we showed that intranasal type I IFN treatment was beneficial only when administered before the onset of symptoms. Importantly, late treatment was ineffective but was not associated with deleterious effects. This study provides important information to interpret clinical trials showing no to modest effects of type I IFNs in COVID-19 patients.