China Wants to Charge Drones From Orbit. The Numbers Do not Add Up.

China says it has figured out how to recharge drone swarms mid-flight — from orbit. The technology is real. The gap between the demo and what a military operator would actually need is enormous. And Beijing has a plan to close it. Gizmodo first reported the Sun Chasing project on May 21, 2026.

Xidian University researchers, led by Duan Baoyan — an academician at the Chinese Academy of Engineering — have built a ground-based wireless power system that sent 143 watts to a moving drone at 30 meters distance, according to Xinhua. A separate test achieved 20.8% DC-to-DC transmission efficiency — turning one-fifth of direct-current input into a microwave beam aimed at a target 100 meters away, with 88% of the beam landing on target. The researchers also confirmed they can power multiple moving drones simultaneously from a single transmitter — effectively a space-based charging pad in embryo. The roadmap runs to 2030 for an orbital demonstration and 2050 for a commercial gigawatt-scale space solar station, backed by a state that has already built the Great Firewall, a lunar base, and a space station.

The 20.8% efficiency figure is what got picked up everywhere. Here is what it actually means. In the 100-meter ground test, the system converted roughly one-fifth of the DC power going into the transmitter into a microwave beam aimed at the receiver — and about 88% of that beam landed on target, per the China Daily report. Those are real numbers, and they represent an improvement over the 15.05% efficiency the same team achieved at the same distance in 2022.

But here is where the math stops being comfortable. The actual power delivered to the drone in the moving test was 143 watts. Not input power — watts received by the aircraft. A small military surveillance drone typically burns 10,000 to 15,000 watts to stay aloft. The gap between what this demo delivered and what a fielded system would need is roughly two orders of magnitude. And that is before accounting for what happens at orbital distances. At 400 kilometers — typical low Earth orbit altitude — even a perfectly focused microwave beam loses roughly 99.9% more intensity than at 100 meters due to path loss alone, before any atmospheric absorption is factored in. Wireless power transfer does not scale linearly with distance. It scales with the inverse square.

There is also a line-of-sight problem. A ground-based charging pad can only reach drones it can see. Buildings, terrain, and the curvature of the Earth all create shadows. From orbit, a satellite can only charge a drone when it passes directly overhead — the same constraint that limits solar panel output on every current satellite. The Chinese researchers' own roadmap acknowledges this: their next step is an actual orbital test, not because the ground demo is insufficient, but because it is obviously not the same thing.

The comparison to Caltech's 2023 space solar power demonstrator — which wirelessly transmitted power in orbit — is being used by several outlets to frame this as a China-versus-Caltech efficiency race. It is not. Caltech's Mappe used a tiny receiver chip illuminated by a transmitted beam to light up a couple of LEDs. The Chinese ground test used a full-size drone receiver at meaningful power levels. They are measuring different things. The Chinese system also does not yet have an independent efficiency verification from a non-Chinese laboratory.

The more direct comparison is the one the Chinese team itself is making: their 2022 baseline was 15.05% efficiency at 100 meters. The new result is 20.8%. That is a genuine improvement in a real program with state backing and a roadmap. An expert panel organized by Shaanxi Provincial Technology Transfer Center concluded the results had reached "world-leading level" — their panel, their words, their standard.

The story is not that China has cracked unlimited drone endurance. It has not. The story is that China has a working ground demonstrator, a funded orbital roadmap, and a state willing to spend across a 25-year horizon to find out if it works at scale. Whether it does — whether the physics cooperate, whether the orbital tests stay on the 2026-2030 schedule, whether the economics survive contact with actual engineering — is the real question. And that one does not have an answer yet.