For twenty years small satellites have been built as cubes, and not for any reason to do with spaceflight. Four spacecraft shaped like dinner plates have now spent six months in orbit testing the alternative.
The cubesat standard was written in 1999 by two university professors who wanted a cheap, repeatable way for students to fly hardware. The unit they defined was a ten centimetre cube, and the reason it stuck is that it fitted a spring-loaded dispenser bolted to the side of a rocket. Build to the box and you get a ride.
It worked extraordinarily well. Thousands of cubesats have flown, and an entire industry of components, radios and deployers grew around that ten centimetre cube.
But nothing about a cube is good for a spacecraft. It is a compromise with a dispenser, and every satellite built to that shape carries the compromise into orbit with it.
The Aerospace Corporation asked what the shape would be if the dispenser were not the constraint. The answer is flat.
Here is what happened
Four flat satellites launched in December and have now flown for six months. The DiskSats went up on 18 December 2025 from NASA's Wallops Flight Facility aboard a Rocket Lab Electron, to a 550 kilometre orbit at 45 degrees inclination.
Each one is a disc a metre across and 2.5 centimetres deep. A 17-kilogram spacecraft with roughly the footprint of a small table top and the thickness of a paperback.
They were flying edge-on within two weeks. Attitude control checkout completed quickly and the discs settled into the orientation the whole design depends on.
The power numbers are the headline result. Each spacecraft generates in excess of 100 watts with up to 282 watt hours of stored energy, which for a 17-kilogram satellite is an unusual ratio.
The hard test is still to come. The remaining demonstration involves progressively lowering at least one vehicle into very low Earth orbit while continuing electric propulsion and orbit control tests. That descent has not yet happened.
How it works
Edge-on flight is the entire idea. The disc flies through its orbit like a frisbee held sideways into the wind, presenting its thin rim to the direction of travel and almost no cross-section to the residual atmosphere.
The same shape solves power at the same time. While the rim faces forward, the broad flat face is turned toward the Sun. One geometry gives minimum drag and maximum collecting area, which is why the wattage is high for the mass.
Very low orbit is the prize. A few hundred kilometres up, a camera sees more detail for the same optics and a radio signal makes a shorter round trip. Everything you want from a satellite improves as you descend.
Drag is what stops everyone. There is still a thin trace of atmosphere down there, and it pulls a conventional satellite out of orbit within months. Fighting it means carrying propellant, which means mass, which means more drag.
Low drag plus high power makes the propulsion work. Electric thrusters produce very little push and need a lot of electricity. A shape that both reduces the force to be countered and increases the power available is what turns very low orbit from a brief visit into somewhere you can stay.
Why it matters
An assumption nobody examined is being tested with hardware. The cubesat standard was a convenience that hardened into a rule, and this is the first serious challenge to it with spacecraft actually flying rather than rendered.
Very low orbit is the last unused piece of near space. It offers sharper imaging and lower latency to anyone who can survive there, and survival is a shape problem before it is a propulsion problem.
It changes the imaging economics. Resolution improves as you get closer, so a cheaper telescope at 300 kilometres can match an expensive one at 550. That matters for the commercial observation market we described in a crowded sky.
It helps rather than worsens the debris problem. Satellites in very low orbit are dragged down naturally within a few years, so the same atmosphere that makes the region difficult also cleans it up.
Form factor is becoming a design variable again. Alongside the standard docking ports we covered in the port that could end disposable satellites, this is part of a wider shift from building satellites around launch constraints to building them around what they need to do.
The honest catch
Six months and four spacecraft is a beginning, and the part everyone wants to see has not been attempted.
They are still at 550 kilometres. The whole case for the shape rests on very low orbit, and the descent is future work. Until a DiskSat has flown low for a sustained period, the drag argument is a design claim supported by good early telemetry.
All figures come from the operator. The Aerospace Corporation designed, built and flies these spacecraft and reported the results, at a conference rather than in a peer-reviewed journal.
Flat is awkward everywhere except in orbit. A one-metre disc needs a launch adapter nobody has standardised, and the entire cubesat supply chain of dispensers, structures and components is built around a different shape.
Attitude control is unforgiving. Edge-on flight only works if the spacecraft holds that orientation continuously. Lose attitude control at very low altitude and the disc becomes the worst possible shape rather than the best.
Nobody has flown a payload on one yet. These are demonstrations of the bus, not of an imaging or communications mission, and payload integration on a 2.5 centimetre deep spacecraft is its own engineering problem.
EDITOR'S TAKE
What makes this worth attention is not the disc, it is the question the disc answers. For twenty years the shape of a small satellite has been set by the geometry of the thing that pushes it out of the rocket, and almost nobody noticed, because the standard worked and standards that work stop being examined. Flatten the spacecraft and two separate problems, drag and power, resolve into the same geometry, which is usually the sign that a design has found the grain of the physics rather than the grain of the paperwork. The number that will settle this is altitude. Four spacecraft holding formation at 550 kilometres is a competent demonstration of a bus. One spacecraft holding station below 300 kilometres for months, generating its own power and fighting its own drag, is a new category of orbit. Watch for that descent, and watch whether anyone flies a camera on one.
Quick questions
Why have small satellites always been cubes?
Because of the container, not the physics. The cubesat standard was defined in 1999 as a ten centimetre cube so that student projects could share rides to orbit cheaply, riding inside a spring-loaded dispenser attached to a rocket. The shape suited the dispenser, the dispenser suited the rocket, and once thousands of satellites had been built to that specification an entire supply chain of structures, radios and deployers assumed it. The cube was never chosen because it is aerodynamically or electrically good. It is a compromise with the delivery mechanism that hardened into an industry standard, which is why challenging it requires solving the launch adapter problem as well as the spacecraft one.
What is very low Earth orbit, and why does anyone want to fly there?
It is the band a few hundred kilometres up, below where most satellites operate. Everything a satellite does gets better closer to the ground: a camera resolves more detail for the same lens, a radio link has less delay and needs less power, and a smaller antenna will do. There is also a housekeeping benefit, because anything that fails at that altitude is pulled down by the atmosphere within a few years instead of staying up for centuries. The obstacle is that same thin atmosphere. It creates drag that removes an ordinary satellite within months, so operating there means either carrying a great deal of propellant or presenting so little cross-section that the drag stops mattering.
Has the flat design actually been proven?
Partly. Four DiskSats have flown for six months, established the edge-on orientation within two weeks, and reported strong power generation for their mass, which validates the basic engineering. What has not happened is the test that the whole concept exists for: descending into very low Earth orbit and staying there. Until one of these spacecraft has held a low altitude for a sustained period while managing its own drag, the central claim remains a well-supported design argument rather than a demonstrated capability. It is also worth noting that the results were reported by the organisation that built and operates the satellites, and presented at a conference rather than peer reviewed.
Sources
SpaceNews: reporting on the DiskSat demonstration and the planned descent into very low Earth orbit.
Small Satellite Conference 2026: DiskSat, On-Orbit Performance and Lessons Learned From the Inaugural Flight of Two-Dimensional Satellites, the technical paper behind this piece.
The Aerospace Corporation: the programme page, with dimensions, mass and power specifications.
Frontier Signal explains frontier technology in plain English. Operator-reported results should be independently verified. This is general information, not investment advice.
