Next generation of CMB experiments will require a large number of detectors (few tens of thousands) in order to tackle the challenging detection of primordial polarization B modes. Furthermore, high resolution experiments are needed for a detailed study of high redshift objects including clusters of galaxies, proto-clusters and dusty galaxies. Within this context Kinetic Inductance Detectors (KIDs) are a serious alternative to bolometers at millimetre wavelengths. Indeed, KIDs are naturally multiplexed and compact allowing us to construct arrays of thousands of detectors. Furthermore, KIDs present short time constants (below 1 ms) and have been demonstrated to be background limited on ground based observations. The NIKA camera, made of two matrices (200 KIDs each) operated at 140 and 240 GHz, has been installed successfully at the IRAM 30 m telescope in Pico Veleta, Granada. NIKA has provided the first ever scientific quality astrophysical observations with KIDs. In particular RXJ1347.5-1145, a massive intermediate redshift galaxy cluster at z = 0.4516 undergoing a merging event, has been successfully mapped at 12 arcsec resolution by NIKA. NIKA is a general purpose camera and it can be also used for other astrophysical objectives including for example observations of high redshift galaxies and proto-clusters, and detailed intensity and polarisation mapping of star-forming regions in the Galaxy. NIKA is a prototype of the NIKA2 camera that should be installed in 2015 at the IRAM 30 m telescope. NIKA2 should have 2 frequency bands at 150 and 250 GHz with about 5000 detectors in total and polarisation capabilities. NIKA2 will be well-suited for in-depth studies of the Intra Cluster Medium in intermediate to high redshift clusters and the follow-up of clusters and proto-clusters newly discovered by the Planck satellite. Finally, we discuss the possibility of including KIDs in the next generation of CMB satellites as for example PRISM.