The direct epitaxy of III-V lasers on Silicon (Si) substrates has been considered for decades as an important objective for the realization of integrated photonics chips. However, the difference in crystal structure, lattice-constant, thermal expansion coefficient as well as issues related to the reactivity of the Si surface have made this topic extremely challenging. For that reason, the direct epitaxy of lasers having good performance and decent lifetime on Si was considered as an almost unattainable dream, and many labs and companies have thus developed ways to circumvent this issue using hybridization techniques. However, major progress has been achieved recently thanks to a better understanding of the defect generation mechanisms, and systematic studies of their suppression [1], or at least the drastic reduction of their density into the active region of the device. In this presentation, we will discuss the challenges related to the growth of III-Vs on Si as well as the progress made lately in that field. We will then discuss the current state-of-the-art of mid-IR lasers grown on Si [2]. GaSb diode lasers (DLs) with GaInAsSb/AlGaAsSb quantum well active zones emitting near 2.3 µm exhibit record threshold current-densities near 200 A.cm-2. Their threshold current density is however sensitive to the threading-dislocation density that generate non radiative recombination [3,4]. In contrast, InAs/AlSb quantum cascade lasers emitting near 8 µm demonstrated performances similar to those on native InAs substrates, thanks to the short radiative lifetimes and the unipolar nature of the radiative process [5]. Finally, at intermediate wavelength (3.5 µm), InAs/GaInSb type-II interband cascade lasers also proved to be relatively insensitive to dislocations, which is ascribed to their particular band structure, their performance being similar to those achieved on native GaSb substrates [6]. These results open the route to active photonic integrated circuits operating in the mid-infrared for sensing applications [7].