Planetary Vision

What if we had a virtual telescope larger than the planet?

Jan 15, 2026

This text is an excerpt from Planetary Vision, an article by Peter Galison published in Volume 2025 of Antikythera: Journal for the Philosophy of Planetary Computation.

Read the Full Article

The Event Horizon Telescope is a collaboration of telescopes and people spanning Earth: It took a planet to see a black hole.

Back in 1969, when astronauts first set foot in the Mare Tranquillitatis, the telescope with the best resolution in the world was the 200-inch aperture Hale Telescope on Palomar Mountain in San Diego County. Hale could just about make out a football-stadium-sized feature on the Moon. The first black hole imaged, Messier 87* (M87*), is larger than our solar system, but so far away (fifty-five million light-years, and therefore tiny in the sky) that to resolve the image required an aperture the size of Earth: not 100 inches, but 500 million inches. This is what it took to resolve an image of M87*—to pick out not a stadium on the Moon as seen from Earth but the straps on Neil Armstrong’s lunar boots. Building such an instrument was the goal of the Event Horizon Telescope (EHT), set even before its formal establishment in 2015.

Planetary in scale

Sadly, we couldn’t build an instrument with an 8,000-mile aperture. Instead, the EHT linked existing radio telescopes all over Earth to synthesize a single, virtual telescope the size of our planet. In fact, we needed not only that Earth-sized virtual machine; we also required the rotation of Earth on its axis as part of the instrument. In a fundamental way, Earth was not so much the object of inquiry (as when mapping a world atlas or globe) but itself a part of the instrument (the telescope).

No single radio telescope, not even the very largest in the world, could possibly image a black hole on its own. It requires a network of integrated telescopes, brought together across the whole of Earth: In the western United States, Mexico, Chile, Hawaii, Spain, the South Pole, and more recently Greenland, France, and Korea. Using advanced computational methods, this extended network provides a skein over Earth itself: a virtual, planet-sized telescope.

Planetary in collaboration

A second meaning of “planetary” in the EHT: there was a human and institutional structure that covered the world. More than 200 scientists and engineers participated in the effort that led to the first image of a black hole, distributed over more than thirty countries and regions. This large-scale collaboration was vital so that the team could attend to the needed range of tasks, from designing the instrument to operating telescopes on 15,000-foot mountain peaks, from correlating the astronomical data on supercomputers all the way through the production of images and their theoretical interpretation.

Planetary interdependence

Though the third planetary aspect of the M87* image is subtle, it is extremely important. We were not just sampling a myriad of different pieces, the way you might send out a hundred photographers to take pictures of a park, with each photographer snapping images of a distinct square on a grid, each of their images a perfectly valid picture. The park photographers can stitch their images together aggregatively, making a large-scale montage out of constituent images. Instead, the EHT needed the assembly of data from the telescopes integratively to mean anything at all; data from any single device signified nothing.

Indeed, the basic unit of the EHT is not a particular telescope or even all the telescopes taken one by one. Instead, it is the baseline: the correlation between two different telescopes. These baselines created a set of frequency-capturing devices, the way a violin string pulled taut between the pegbox and the bridge vibrates in the presence of the notes G, D, A, or E. (A string fixed only to the bridge would not vibrate; it would only flap in the wind.) Were you to have a sufficient number and variety of such string lengths—if you could measure how much each one was vibrating and when—then it would be possible to reconstruct a song. So it is visually: Each pair of telescopes added information about the image. Put together a sufficient number of telescope pairs with a variety of distances between each duo, and the image of the black hole can be constructed: by integration, not aggregation.

For that first, now famous image of M87* released in April 2019, we had telescopes on five sites. Correlating the signals falling on each pair of telescopes was vital, because a single telescope captures vastly more noise than signal; it was essentially a bucket collecting all the light that fell on it from any source in view. The signal—light coming from around the black hole—could be extracted only by correlating the results of two telescopes at a time, so if some unwanted light fell into one and not another, that junk would be eliminated in the process of comparing the two telescopes.

Only with these pair-wise correlations could we go from the 5 petabytes of data collected in the observing campaign to the tiny fraction of those data relevant for the black hole image.

The three senses of a planetary vision were all at play here: the world-spanning array of telescopes, the intercontinental distribution of scientists and institutions, and the fundamentally correlative nature of the instrument. The EHT experiment was planetary in scope and essence. It used the size of Earth, the spin of Earth; it engaged institutions and people from dozens of countries and regions around the planet. It processed the data by correlating these various telescopes to produce the fundamental data that undergirded the whole image-making project. Together, the three realms of planetary vision led to the image of the supermassive black hole at the center of the distant galaxy M87, and the smaller, but still supermassive black hole at the core of our galaxy, the Milky Way.