Résumé
We review the current status of Early Dark Energy (EDE) models proposed to resolve the ``Hubble tension'', the discrepancy between ``direct'' measurements of the current expansion rate of the Universe and ``indirect measurements'' for which the values inferred rely on the $\Lambda$CDM cosmological model calibrated on early-universe data. EDE refers to a new form of dark energy active at early times (typically a scalar-field), that quickly dilutes away at a redshift close to matter-radiation equality. The role of EDE is to decrease the sound horizon by briefly contributing to the Hubble rate in the pre-recombination era. We summarize the results of several analyses of EDE models suggested thus far in light of recent cosmological data, including constraints from the canonical {\it Planck} data, baryonic acoustic oscillations and Type Ia supernovae, and the more recent hints driven by cosmic microwave background observations using the Atacama Cosmology Telescope. We also discuss potential challenges to EDE models, from theoretical ones (a second ``cosmic coincidence'' problem in particular) to observational ones, related to the amplitude of clustering on scales of $8h$/Mpc as measured by weak-lensing observables (the so-called $S_8$ tension) and the galaxy power spectrum from BOSS analyzed through the effective field theory of large-scale structure. We end by reviewing recent attempts at addressing these shortcomings of the EDE proposal. While current data remain inconclusive on the existence of an EDE phase, we stress that given the signatures of EDE models imprinted in the CMB and matter power spectra, next-generation experiments can firmly establish whether EDE is the mechanism responsible for the Hubble tension and distinguish between the various models suggested in the literature.