Following the COVID-19 pandemic, lung ultrasound (LUS) received growing attention as the need for a safe, point-of care, cost-effective, and widely available imaging solution was crucial. However, when performed with clinical scanners, LUS remains limited to the subjective and qualitative interpretation of artifacts and imaging patterns. This negatively affects accuracy and reproducibility. During the COVID-19 pandemic, we have developed a standardised LUS acquisition protocol and scoring system, aiming at minimizing the impact of several confounding factors. This approach has proven prognostic value and is capable of stratifying patients based on their risk of worsening. Moreover, the availability of an extensive and standardised dataset (772,780 LUS frames) allowed us to develop AI approaches capable of analysing LUS data automatically. Results from a multicentric study conducted on 220 patients will be presented. Next, beyond what can be achieved with clinical scanners, we are also pioneering the development of novel ultrasound solutions dedicated to the lung. To provide with a forward-looking view on the future of LUS, the latest results from numerical and in vitro- studies focused on unveiling the physical phenomena at stake will be discussed in detail, together with the latest results from the first clinical studies on quantitative lung ultrasound spectroscopy.