On 30 June 2026, a telescope on a mountain in Chile started a movie it will spend the next ten years filming [source: Rubin Observatory, 2026]. The NSF–DOE Vera C. Rubin Observatory formally began the Legacy Survey of Space and Time (LSST), and the headlines that followed were the kind astronomy rarely gets: the biggest camera ever built, millions of galaxies in a single frame, thousands of new asteroids, a hunt for dark matter and dark energy. Six weeks in, it is worth doing something the coverage mostly didn't — separating what this machine has already measured from what it is merely expected to find, and separating a genuinely new kind of telescope from the space-science stories it keeps getting filed next to.
Because the confusion is easy. When people hear "new observatory + galaxies + cosmology," they think of the James Webb Space Telescope. But Rubin is close to Webb's opposite. Webb stares at a tiny patch of sky and takes a deep, detailed spectrum of a single object — one exoplanet's atmosphere, one distant galaxy. Rubin does the reverse: it photographs the entire southern sky, over and over, to catch anything that moves or changes. It trades depth on one target for breadth across all of them. Understanding that difference is the key to understanding what its first results do — and don't — mean.
The machine, in numbers that are actually measured
Start with the hardware, because that is the part that is real and finished, not projected. At the heart of Rubin sits LSSTCam, the largest digital camera ever built for astronomy, at 3,200 megapixels — 3.2 gigapixels [source: SLAC, 2025; NOIRLab, 2024]. Its focal plane is made up of 201 individual custom-designed CCD sensors [source: BNL, 2024]. The camera is roughly the size of a small car and weighs about 2,800 kg (6,200 lb) [source: SLAC, 2025], and its front lens, at 1.57 metres (5.1 ft) across, is the largest high-performance optical lens ever fabricated [source: BNL, 2024]. It rides on the 8.4-metre Simonyi Survey Telescope at Cerro Pachón, in the Chilean Andes [source: SLAC, 2025; Rubin Observatory, 2026].
The number that matters most, though, isn't the pixel count — it's the field of view. A single Rubin image covers a patch of sky about 45 times the area of the full Moon [source: SLAC, 2025]. That width is the whole point. Because each shot is so wide, the telescope can photograph the entire southern sky roughly every three nights, taking a new image about every 40 seconds and around 1,000 images a night [source: Rubin Observatory, 2026]. Over the decade, it will return to each patch of sky about 800 times [source: Rubin Observatory, 2026]. Add it up and it produces on the order of 20 terabytes of data per night [source: SLAC, 2025].
None of those are promises. They are specifications of a camera that has been built, installed, and switched on. This is the layer of the Rubin story that is fully verified: the machine exists and performs to spec.
What it has already done
The verified column has real science in it, too. When Rubin released its first-look images in June 2025, a little over ten hours of test observations captured millions of galaxies and Milky Way stars [source: SLAC, 2025]. In those same early frames, the system flagged 2,104 previously unseen asteroids, including seven near-Earth objects — in ten hours [source: SLAC, 2025]. For context, that is not the sky being unusually full of asteroids; it is what happens when a wide, fast, repeat-imaging survey looks carefully at ground most telescopes sweep past.
Then, in February 2026, Rubin switched on the part of the system that makes it different from a very good camera: the alert stream. On the night of 24 February 2026, the observatory issued roughly 800,000 alerts — automated notifications that something in the sky had changed — each generated within about two minutes of the exposure [source: Rubin Observatory, 2026]. Those alerts flag supernovae, variable stars, active galactic nuclei, and moving Solar System objects, and they flow out to community software "brokers" that sort and classify them [source: Rubin Observatory, 2026].
Here is where the first careful distinction matters. Rubin's alert system is designed to issue up to about 7 million alerts a night at full tilt [source: Rubin Observatory, 2026]. That is a design capacity — a ceiling the pipeline was built to handle. The 800,000 it produced on its first alert night is the demonstrated figure. Both numbers are true; they are not the same kind of true. One describes what the plumbing can carry, the other what has actually flowed through it. Reporting that blurs them — and some did, quoting even higher round numbers — turns an engineering spec into a discovery.
The asteroid story: a real count and a much bigger forecast
Nowhere is the gap between measured and expected clearer than with asteroids, and it is worth walking through because it is the observatory's most concrete near-term promise.
The measured part is genuinely impressive. Using only early commissioning and optimization data — about one million observations gathered over roughly a month and a half — Rubin discovered more than 11,000 new asteroids and submitted them to the International Astronomical Union's Minor Planet Center [source: Rubin Observatory, 2026; University of Washington, 2026]. The finds arrived in accelerating bursts: 73 from the early Commissioning Camera in late 2024, then 1,514 during the June 2025 First Look, then about 11,000 as the survey ramped up over the following summer [source: University of Washington, 2026]. Among them were 33 previously unknown near-Earth objects and roughly 380 trans-Neptunian objects in the distant Solar System [source: University of Washington, 2026]. This haul required a new asteroid-discovery software architecture, built by University of Washington researchers, because Rubin's observing rhythm doesn't match the surveys the old tools were written for [source: University of Washington, 2026].
And now the crucial qualifier, stated plainly in the primary source: none of the newly found near-Earth objects pose a threat to Earth, and the largest is about 500 metres across [source: University of Washington, 2026]. That sentence does two things. It answers the question everyone actually asks, and it models the honesty this whole subject needs. "Rubin found 33 new near-Earth asteroids" is a headline that can read as alarming; the finding is that Rubin can find them, not that any of them is coming.
Then there is the forecast column, which is where most of the eye-popping numbers live — and which is not yet data. Over its full ten years, the survey is expected to discover nearly 90,000 more near-Earth objects, to roughly triple the total number of known asteroids, and to increase the count of known trans-Neptunian objects by nearly an order of magnitude [source: University of Washington, 2026]. Today only around 40% of the mid-size near-Earth objects — those at least 140 metres across, big enough to devastate a region — have been catalogued, and Rubin is expected to push that fraction up substantially over the decade [source: University of Washington, 2026]. These are the planetary-defense payoffs, and they are real goals with a real instrument behind them. But 90,000 is a projection; 11,000 is a measurement. The first is what Rubin should do; the second is what Rubin has done.
The big science is a ten-year question, not a first-year answer
The observatory is named for Vera C. Rubin, whose measurements of how galaxies rotate provided some of the strongest early evidence that most of the universe's matter is invisible [source: NOIRLab, 2024]. So it is fitting that the survey's headline scientific goals are the deepest ones: to probe dark matter and dark energy, the unseen mass and the mysterious acceleration that together make up most of the cosmos.
But this is precisely where expectations need the firmest hand. Rubin attacks dark matter and dark energy indirectly, by mapping billions of galaxies and measuring the faint, systematic ways their apparent shapes are distorted by the gravity of intervening matter — an effect called weak gravitational lensing — and by tracking how the universe's structure has changed across cosmic time. Doing that requires the full statistical weight of the whole survey. It is, by design, a ten-year measurement. The first images, however beautiful, and the first alerts, however fast, do not contain the answer. Nothing about dark energy has been concluded, and nothing could be this early. The machine that will eventually make that measurement has been demonstrated to work; the measurement itself has not begun to resolve.
The same "expected, not yet delivered" caveat applies to the survey's other superlatives. Over ten years Rubin is projected to catalogue on the order of 20 billion galaxies and billions of stars, amassing roughly 500 petabytes of data [source: SLAC, 2025]. Those totals describe the finished library, not the current shelf. It is entirely reasonable to be excited by them — and important to file them under forecast.
Why this is a different kind of telescope
Step back and the reason Rubin resists easy comparison comes into focus. Most famous observatories are built to look hard at few things. Webb takes exquisite spectra of individual targets; a classic large telescope points where an astronomer tells it to and drinks in photons from one object for hours. Rubin is built to look repeatedly at everything in its half of the sky. Its unit of discovery isn't the deep exposure; it's the difference between tonight's sky and last week's.
That design is what makes its natural products asteroids, supernovae, variable stars, and alerts — things defined by motion and change — rather than the deep portrait of a single galaxy. It is also why the honest way to read Rubin is over time. A wide-field, time-domain survey doesn't announce its biggest results on day one; it accumulates them, night over night, as the same patches of sky are compared again and again. The first images are the trailer. The movie is the point, and it has ten years to run.
What to watch
So six weeks into the survey, the honest scorecard has three columns, and keeping them apart is the whole discipline. Measured and done: the largest astronomical camera ever built, a sky-covering cadence, 11,000 real asteroids, 2,104 in the first images, 800,000 alerts in a single night. Designed and demonstrated but not yet at scale: an alert pipeline built for up to 7 million notifications a night. Expected but not yet delivered: 90,000 more near-Earth objects, a tripling of the asteroid catalogue, 20 billion galaxies mapped, and the long, patient measurement of dark matter and dark energy.
Three things will tell you the survey is delivering on the third column rather than just the first. First, whether the asteroid count climbs from thousands toward the projected tens of thousands as the years accumulate — the near-term promise that is easiest to check. Second, whether the alert stream, running near its design capacity, actually yields new classes of transient events rather than just more of the same. And third — the one worth real patience — whether the weak-lensing and structure measurements, years from now, sharpen what we know about dark energy. Until then, the useful posture is the one the evidence supports: marvel at the machine, take the early detections seriously, and hold the ten-year promises exactly as what they are — promises from an instrument that has, at least, proven it can see.