mRNA COVID-19 vaccines have hegemonized the world market, and their administration to the population promises to stop the pandemic. However, the waning of the humoral immune response, which does not seem to last so many months after the completion of the vaccination program, has led us to study the molecular immunological mechanisms of waning immunity in the case of mRNA COVID-19 vaccines. We consulted the published scientific literature and from the few articles we found, we were convinced that there is an immunological memory problem after vaccination. Although mRNA vaccines have been demonstrated to induce antigen-specific memory B cells (MBCs) in the human population, there is no evidence that these vaccines induce the production of long-lived plasma cells (LLPCs), in a SARS-CoV-2 virus naïve population. This obstacle, in our point of view, is caused by the presence, in almost all subjects, of a cellular T and B cross-reactive memory produced during past exposures to the common cold coronaviruses. Due to this interference, it is difficult for a vaccination with the Spike protein alone, without adjuvants capable of prolonging the late phase of the generation of the immunological memory, to be able to determine the production of protective LLPCs. This would explain the possibility of previously and completely vaccinated subjects to become infected, already 4-6 months after the completion of the vaccination cycle.