Cosmic microwave background (CMB) experiments aim to better characterize the faint intensity and polarization anisotropies, in order to further constrain cosmological parameters, to better understand the physics of neutrinos, and to detect a possible signature of cosmic inflation. Existing detectors used to measure the CMB are already operating at or near the fundamental photon noise limit. To increase sensitivity, arrays containing tens or hundreds of thousands of detectors will be required for the next generation of experiments. Microwave kinetic inductance detectors (MKIDs) are a candidate to replace current detector technologies because they may be better suited for large arrays. I will introduce the physics of MKIDs then will discuss my group’s work to develop MKIDs for future CMB experiments, including multi-chroic devices in which each pixel is sensitive to two polarizations in two spectral bands.