Despite the widespread use of seasonal influenza vaccines, there is urgent need for a universal influenza vaccine to provide broad, long-term protection. A number of factors underpin this urgency, including threats posed by zoonotic and pandemic influenza A viruses, suboptimal effectiveness of seasonal influenza vaccines, and concerns surrounding the effects of annual vaccination. In this article, we discuss approaches that are being investigated to increase influenza vaccine breadth, which are near-term, readily achievable approaches to increase the range of strains recognized within a subtype, or longer-term more challenging approaches to produce a truly universal influenza vaccine. Adjuvanted and neuraminidase-optimized vaccines are emerging as the most feasible and promising approaches to extend protection to cover a broader range of strains within a subtype. The goal of developing a universal vaccine has also been advanced with the design of immunogenic influenza HA-stem constructs that induce broadly neutralizing antibodies. However, these constructs are not yet sufficiently immunogenic to induce lasting universal immunity in humans. Advances in understanding how T cells mediate protection, and how viruses are packaged, have facilitated the rationale design and delivery of replication-incompetent virus vaccines that induce broad protection mediated by lung-resident memory T cells. While the lack of clear mechanistic correlates of protection, other than haemagglutination-inhibiting antibodies, remains an impediment to further advancing novel influenza vaccines, the pressing need for such a vaccine is supporting development of highly innovative and effective strategies.