During most of their life, stars fuse hydrogen into helium in their cores. Mixing of chemical elements in the radiative envelope of stars with a convective

Context. Massive stars are predicted to excite internal gravity waves (IGWs) by turbulent core convection and from turbulent pressure fluctuations in their near-surface layers. These IGWs are extremely efficient at transporting angular momentum and chemical species within stellar interiors, but they remain largely unconstrained observationally. Aims. We aim to characterise the photometric detection of IGWs across a large number of O and early-B stars in the Hertzsprung–Russell diagram, and explain the ubiquitous detection of stochastic variability in the photospheres of massive stars. Methods. We combined high-precision time-series photometry from the NASA Transiting Exoplanet Survey Satellite with high-resolution ground-based spectroscopy of 70 stars with spectral types O and B to probe the relationship between the photometric signatures of IGWs and parameters such as spectroscopic mass, luminosity, and macroturbulence. Results. A relationship is found between the location of a star in the spectroscopic Hertzsprung–Russell diagram and the amplitudes and frequencies of stochastic photometric variability in the light curves of massive stars. Furthermore, the properties of the stochastic variability are statistically correlated with macroturbulent velocity broadening in the spectral lines of massive stars. Conclusions. The common ensemble morphology for the stochastic low-frequency variability detected in space photometry and its relationship to macroturbulence is strong evidence for IGWs in massive stars, since these types of waves are unique in providing the dominant tangential velocity field required to explain the observed spectroscopy.

Almost all massive stars explode as supernovae and form a black hole or neutron star. The remnant mass and the impact of the chemical yield on subsequent star formation and galactic evolution strongly depend on the internal physics of the progenitor star, which is currently not well understood. The theoretical uncertainties of stellar interiors accumulate with stellar age, which is particularly pertinent for the blue supergiant phase. Stellar oscillations represent a unique method of probing stellar interiors, yet inference for blue supergiants is hampered by a dearth of observed pulsation modes. Here we report the detection of diverse variability in blue supergiants using the K2 and TESS space missions. The discovery of pulsation modes or an entire spectrum of low-frequency gravity waves in these stars allow us to map the evolution of hot massive stars towards the ends of their lives. Future asteroseismic modelling will provide constraints on ages, core masses, interior mixing, rotation and angular momentum transport. The discovery of variability in blue supergiants is a step towards a data-driven empirical calibration of theoretical evolution models for the most massive stars in the Universe.Stars born with masses larger than approximately eight times the mass of the Sun play a significant role in the evolution of galaxies. They are the chemical factories that produce and expel heavy elements through their wind and when they end their lives as supernovae and form a black hole or neutron star 1-3 . However, the chemical yields that enrich the interstellar medium and the remnant mass strongly depend on the progenitor star's interior properties 4 . The detectable progenitors of supernovae include blue supergiant stars, which are hot massive stars in a shell-hydrogen or core-helium burning stage of stellar evolution. Stellar evolution models of these post-main sequence stars contain by far the largest uncertainties in stellar astrophysics, as observational constraints on interior mixing, rotation and angular momentum transport are missing. These phenomena are further compounded when coupled with mass loss, binarity and magnetic fields 1-3 . Across astrophysics, from star formation to galactic evolution, it is imperative to calibrate theoretical models of massive stars using observations because they determine the evolution of the cosmos.A unique methodology for probing stellar interiors is asteroseismology 5 , which -similarly to seismology of earthquakes -uses oscillations to derive constraints on the structure of stars.The study of stellar interiors of low-mass stars like the Sun has undergone a revolution in