China's Long March 10B Catches Its Booster at Sea — What the Net Means for the Lunar Race

On July 10, 2026, at 04:15 UTC, a Long March 10B lifted off from Launch Complex 2 at the Wenchang Commercial Space Launch Site on Hainan island. Some ten minutes later, the rocket’s first stage hung from tensioned steel cables stretched across the deck of a ship in the South China Sea, its seven engines extinguished. China had become the second country in history to recover an orbital-class booster intact. It is also the first time anyone has done it with a net rather than landing legs. And it worked on the vehicle’s maiden flight — a combination no other orbital launcher has achieved.

How the Booster Came Back

After staging, the Long March 10B first stage executed a re-entry sequence that shares its logic with other reusable vehicles but differs in its final step. Three of the stage’s YF-100K engines reignited for a re-entry burn, decelerating the descending booster as it entered the thicker lower atmosphere. The engine count then dropped to one for a final approach and hover phase, during which the booster positioned itself precisely over the Linghang Zhe, a purpose-built recovery vessel stationed approximately 430 kilometres downrange from Wenchang.

Four metal hooks at the upper end of the first stage then engaged tensioned steel cables strung across the Linghang Zhe’s deck. The engine cut off. The booster hung from those cables while an automated clamping platform beneath it locked the stage in place for the return transit. From stage separation to confirmed catch: approximately six minutes.

The Linghang Zhe is a purpose-built platform, not a converted barge. It carries certification from the China Classification Society, making it the first sea-based rocket recovery structure in China to receive that class and statutory certification.

The Vehicle That Had to Work First

The Long March 10B is a two-stage vehicle, 63 metres tall with a 5-metre core diameter. At liftoff, it weighs approximately 760 metric tons, and its seven YF-100K engines generate a combined thrust of roughly 890 metric tons.

The YF-100K burns liquid oxygen and kerosene in an oxidizer-rich staged combustion cycle. In this architecture, a pre-burner runs a fuel-lean, oxidizer-rich mix to generate hot gas that drives the turbopumps, before the fully mixed propellants ignite in the main combustion chamber. The high chamber pressures this produces yield a sea-level specific impulse of approximately 300 seconds. Oxidizer-rich staged combustion at this scale was previously the exclusive domain of Soviet and Russian LOX-kerosene engine programmes. The YF-100K family changed that.

The second stage uses a single YF-219 engine burning liquid oxygen and liquid methane, producing 140 tonnes of vacuum thrust. The propellant split between stages is deliberate. Kerosene’s high volumetric density suits the first stage’s sea-level operation and pad handling. Methane burns without leaving coking deposits in the engine, simplifying post-flight inspection. Rapid reuse depends partly on what the refurbishment team finds inside the engine. A clean combustion environment compresses that schedule.

In reusable configuration, the Long March 10B delivers at least 16 tonnes to a 200-kilometre low Earth orbit. On this maiden flight, the second stage placed CX-26, an experimental satellite built and operated by China Satellite Network Group, into a target orbit at 800 kilometres altitude.

The Engineering Case Against Landing Legs

SpaceX lands the Falcon 9 on deployable legs, upright, atop a drone ship or ground pad. Booster B1067 flew for the 36th time on July 9 — the day before this launch — demonstrating where the legs-and-pad approach can go given a decade of iteration.

CALT chose a different architecture. Landing legs are structural mass the booster must carry from the pad on every flight, regardless of whether a recovery is attempted. Each kilogram of leg hardware is a kilogram that cannot be propellant or payload. By relocating the mechanical complexity of the catch from the rocket to the recovery vessel, the Long March 10B first stage reaches orbit carrying less inert mass.

The outcome: at least 16 tonnes to low Earth orbit in reusable mode, against Falcon 9’s approximately 17.4 tonnes. The gap is narrow. The trade shifts mass from rocket to infrastructure, which is a legitimate engineering decision when the infrastructure is stationary and the rocket flies repeatedly.

There is a corresponding cost. The recovery vessel must be positioned precisely in open ocean. The cable tension and hover guidance requirements are tighter than a pad landing, where the rocket targets a fixed, instrumented surface. The Linghang Zhe has worked exactly once. The Falcon 9 has been recovered several hundred times.

The Rocket That 10B Is Clearing the Way For

The Long March 10B is not the vehicle China intends to use for its crewed Moon landing. It is the vehicle designed to prove that the recovery architecture works before any human depends on it.

China’s crewed lunar vehicle, the Long March 10A, is a larger three-stage system intended to carry two separate payloads under Project 921: the Mengzhou crewed spacecraft and the Lanyue lunar lander. Each crewed lunar landing requires two CZ-10A launches. Mengzhou and the Lanyue lander stack rendezvous in lunar orbit; the crew transfers to Lanyue for the descent, then returns to Mengzhou for the journey home. The stated target is a crewed Moon landing before 2030.

Two launches per mission places a hard constraint on how reliably the CZ-10A first stage can be recovered and returned to service. A recovery architecture that requires months of structural repair between flights cannot support a sustainable lunar cadence. A series of CZ-10B recoveries and reflights under actual orbital mission conditions gives CALT the data it needs before the stakes include crew.

CASC has stated it intends to refly the July 10 first stage before the end of 2026. If that holds, it establishes that the stage sustained no damage during recovery requiring extended repair, and begins to define what a realistic turnaround cycle looks like under operational conditions rather than test assumptions.

One Catch. Not Yet a Programme.

SpaceX recovered its first Falcon 9 booster on December 21, 2015. The first refly did not happen until March 30, 2017 — fifteen months later. The booster that flew for the 36th time on July 9, 2026 is the product of over a decade of iteration, a refurbishment process revised across hundreds of landings, and an industrial supply chain built specifically around high-cadence reuse.

China caught one booster in a net. It is the only time the Linghang Zhe has been used for this purpose.

A successful refly before December 2026 does not close that gap. What it closes is the first and most basic question the gap contains: whether the July 10 stage is genuinely reusable, or whether the maiden flight catch was a demonstration that will take months to turn into a second flight. Those are different claims, and only one of them has programme-level consequences for the 2030 deadline.

The answer begins when the inspection team opens the recovered stage. What they find there determines whether July 10 was the start of China’s reusable rocket era, or the first data point in what will prove to be a longer and more complicated story.