EDGES sits at the Murchison Radio-astronomy Observatory, on Wajarri Yamaji country some 800 kilometres north of Perth. The Wajarri name for the place, Inyarrimanha Ilgari Bundara, means 'sharing sky and stars,' and the choice is no accident. This is a legally protected radio quiet zone, where mobile phones, transmitters and stray electronics are restricted across a radius that now stretches hundreds of kilometres. The signal EDGES hunts is staggeringly faint, the whisper of hydrogen gas from when the cosmos was young, drowned a thousand times over by the glow of our own galaxy and by any human noise nearby. To hear it at all, you need somewhere on Earth where the radio sky is nearly perfectly dark.

Before the first stars, the universe was a fog of cold hydrogen, dark and featureless. When stars finally ignited, their ultraviolet light reached out and altered that gas, changing how it absorbed and emitted radiation at a characteristic 21-centimetre wavelength through a subtle effect called Wouthuysen-Field coupling. Stretched by the expansion of the universe across more than thirteen billion years, that signal should arrive at Earth not as light but as a low radio hum near 78 megahertz. What EDGES reported was a broad absorption dip centred exactly there, a shadow in the static at the redshift where theory says the first stars should announce themselves. It looked like the universe turning on the lights, recorded by a box on the ground in the Australian outback.

The signal was a puzzle from the start, because it was wrong in an interesting way: the dip was roughly twice as deep as any standard model predicted. Something had made the early hydrogen colder than it should have been. One bold explanation was that the gas had been chilled by colliding with dark matter, the invisible substance that outweighs ordinary matter five to one. That possibility electrified physicists. If real, EDGES was not only listening to the first stars but catching dark matter in the act of touching ordinary matter for the first time. It was the kind of result that, if it held, would rewrite two fields at once. That was exactly why everyone wanted to see it confirmed.

Confirmation has not come. In 2022 a rival Indian experiment called SARAS 3 reached comparable sensitivity, looked for the same dip, and found nothing. Its analysis suggested the EDGES detection was most likely an artifact rather than a message from the cosmic dawn. Suspicion has fallen on the edges of the 30-metre metal screen laid around the antenna: if the correction for reflections off that ground plane was even slightly off, it could fake an absorption trough indistinguishable from a real signal. The EDGES team has continued refining its instruments, and the question is not yet closed. For now the dip near 78 megahertz hangs in a strange limbo, possibly the oldest sound ever recorded, possibly the hum of the apparatus that recorded it.