Very metal-poor massive stars hold the key to interpret high-redshift star-forming galaxies and the early reionization epoch, but also contemporary events such as gravitational waves. To study these objects in resolved environments, we need to resort to dwarf irregular galaxies far from the potential wells of M31 and the Milky Way, and therefore distant. While the archives, recently boosted by the ULLYSES and XSHOOTU programs, store a healthy dataset of massive stars in the Milky Way and the Magellanic Clouds, the number of observed targets with poorer metal content than the SMC (1/5 Zȯ) is dramatically small. This paper reviews the state of observations of very metal-poor massive stars, assessing what can be realistically learned about their physics and evolution with current instrumentation, and arguing whether or not near-future facilities can remedy the gaps in the knowledge that remain.

Winds of massive stars are an important ingredient in determining their evolution, final remnant mass, and feedback to the surrounding interstellar medium. We compare empirical results for OB star winds at low metallicity with theoretical predictions. Observations suggest very weak winds at SMC metallicity, but there are exceptions. We identified promising candidates for rotationally enhanced mass-loss rates with two component wind and partially stripped stars hiding among OB stars with slow but dense wind in the SMC. A preliminary analysis of these systems, derived parameters, and their implications are discussed. Finally, we briefly discuss the interaction of OB winds near black holes in X-ray binaries.

Understanding stellar birth requires observations of the clouds in which they form. These clouds are dense and self-gravitating, and in all existing observations, they are molecular with H2 the dominant species and CO the best available. When the abundances of carbon and oxygen are low compared to hydrogen, and the opacity from dust is also low, as in primeval galaxies and local dwarf irregular galaxies CO forms slowly and is easily destroyed, so it cannot accumulate inside dense clouds. Then we lose our ability to trace the gas in regions of star formation and we lose critical information on the temperatures, densities, and velocities of the material that collapses. I will report on high resolution observations with ALMA of CO clouds in the local group dwarf irregular galaxy WLM, which has a metallicity that is 13% of the solar value and 50% lower than the previous CO detection threshold and the properties derived of very small dense CO clouds mapped..