A Chinese satellite equipped with a robotic “octopus arm” has passed a key refueling test in low Earth orbit (LEO), according to state-run media. The achievement highlights China’s continued leadership with this particular technology, which NASA has not yet caught up with.
The experimental spacecraft will eventually deploy a giant balloon in LEO, which could help solve another important issue surrounding satellite “megaconstellations” like SpaceX‘s Starlink network.
Hukeda-2 is a demonstration satellite intended to test new technologies in LEO. Its most notable attachment is an octopus-like robotic arm that “can curl, twist and wrap around objects to work in tight, complex spaces, with a nozzle-like tip at one end designed to line up and connect with a target port,” according to the South China Morning Post.
The arm is made of a series of spring-like tubes threaded with cables attached to a motor, allowing it to bend in almost any direction and make the small adjustments needed to dock with another satellite while both spacecraft are traveling at speeds of around 16,800 mph (27,000 km/h).
On March 24, Chinese state media reported that Hukeda-2’s robotic arm had successfully completed its first refueling test. It was initially unclear if this test involved another satellite. However, photos have since confirmed that the robotic arm instead docked with a target port located on Hukeda-2 itself.
This is the biggest milestone for satellite refueling since June 2025, when China’s Shijian-25 satellite successfully connected to and refueled the Shijian-21 satellite, which had previously run out of fuel. This encounter occurred in a higher, geosynchronous orbit, around 22,000 miles (33,500 km) above Earth’s surface, and was the first confirmed case of satellite-to-satellite refueling, according to Live Science’s sister site Space.com.
Refuel, reuse, recycle
When satellites run out of fuel, they can no longer maintain their altitude and are slowly pulled back toward Earth, before eventually burning up in the upper atmosphere. By refueling them, operators can keep the same spacecraft in orbit much longer, thus making them more cost-effective and sustainable, reducing the need to launch replacements. This has been a big aim for NASA and Western companies for several years, but has remained out of reach so far.
China will likely attempt to use similar refueling spacecraft to service its rapidly expanding Qianfan, or “Thousand Sails,” constellation, which launched its first batch of satellites in 2024 and is set to rival Starlink in the coming years. (There are currently about 108 active Qianfan satellites in orbit, with plans to deploy 15,000 by 2030.)
SpaceX, on the other hand, does not appear interested in sustaining its active satellites. Instead, the company favors repeated launches of new spacecraft with its reusable Falcon 9 rocket to keep costs down.
Another issue with expanding constellations is that the number of dead spacecraft waiting to fall back to Earth is rising fast and taking up valuable space that could be occupied by new satellites. In an attempt to fix this, Hukeda-2 will deploy an 8-foot-wide (2.5 meters) balloon at the end of its mission, which will increase atmospheric drag and speed up its return to Earth.
If this works, future Chinese satellites could be deployed with similar devices that allow them to fall back to Earth without initiating a final deorbiting burn. However, it is unclear when Hukeda-2 will deploy its test balloon.
Although quick deorbiting of satellites is prudent, it is unlikely to fix the greater issue of overcrowding that is expected to occur in LEO in the coming decades, especially if SpaceX’s controversial plan to launch 1 million orbital data centers comes to fruition.
Recent research has also revealed that satellite reentries release high levels of metal pollution in the upper atmosphere, which is likely triggering issues we are still not fully aware of.
“What goes up must come down,” University of Regina astronomer and vocal megaconstellation critic Samantha Lawler recently wrote for Live Science.
