At 7:35 a.m. on August 19, 2026, at the Dongfeng Commercial Space Innovation and Test Zone in Jiuquan, Gansu, Zhuque-3 Yao-2 tore into the sky on a pale-blue plume. All eyes followed its trajectory westward toward Minqin County, 390 kilometers away. About 137 seconds after liftoff, the first and second stages separated. At around 7:41 a.m., a 66-meter-tall behemoth came plunging vertically from an altitude of more than 100 kilometers. Its four landing legs deployed cleanly, and the rocket stood upright on the landing pad of the Zhuque-3 landing site in Minqin County, Gansu Province. It had done it.
On December 30, 2011, Elon Musk wrote on Twitter that the endgame in the rocket technology race was all about China. Fourteen years later, China’s private commercial space sector answered with a complete land-based recovery. For the first time, a Chinese commercial rocket had literally stood on its own legs.
This was not China’s first rocket recovery. Just 40 days earlier, the Long March 10B had used a net to “catch” its first stage at sea. Zhuque-3 took a different approach: vertical recovery on landing legs. China has now demonstrated both.
To understand just how difficult today’s landing was, consider what happened 257 days earlier. At noon on December 3, 2025, Zhuque-3 Yao-1 lifted off from the same site. It was China’s first attempt to recover the first stage of an orbital-class rocket. The second stage successfully reached orbit, and the mission was otherwise a success. During the first stage’s return, everything went according to plan—supersonic reentry, aerodynamic gliding, guidance and control—until the final moments of landing, when an anomaly occurred during the engine braking burn. The vehicle caught fire in midair and crashed near the edge of the landing zone.
As Dong Kai, deputy chief designer of LandSpace, put it, “Everything that needed to happen happened. We were just missing that final push.” The failure did not shake the team’s confidence. If anything, it clarified the path forward. After the failed maiden flight, Zhang Changwu, founder and CEO of LandSpace, said simply: “Next time, we won’t let it crash again.” 257 days later, he kept his word.
The State Team’s “Catch” vs. the Private Sector’s “Stand”: What’s the Difference?
The state-led program has pursued a rocket recovery method pioneered in China: net-based recovery. On July 10, the Long March 10B made its maiden flight from the Hainan Commercial Space Launch Site. About six minutes after stage separation, the first stage returned vertically to an offshore recovery platform and was successfully captured by a net system.
It was the world’s first successful net-based recovery of a launch vehicle. Rather than using landing legs to stand on the ground, the rocket flew directly into a large net and was caught by a flexible arresting system.
Net-based recovery requires specialized infrastructure, but it eliminates heavy landing legs, reducing structural mass and minimizing the impact on payload capacity. The Long March 10B has a liftoff mass of about 760 tonnes and can deliver 16 tonnes to low Earth orbit in reusable configuration. By simplifying the vehicle’s onboard structure and reducing its weight, the net system improves payload capacity and overall economics.
Zhuque-3 takes a different route. It carries four landing legs throughout ascent, sacrificing some payload capacity in return. The margin for error during the final landing phase is extremely small, placing demanding requirements on guidance and control accuracy.
The advantages, however, are equally clear: the vehicle can land on both land and sea, giving operators greater flexibility in mission planning and supporting frequent, routine launches. Think about why SpaceX’s Falcon 9 can fly several times a week. The answer lies in its vertical landing-leg recovery architecture.
With this flight, Zhuque-3 has, for the first time in China, successfully completed the same basic recovery approach that SpaceX has developed and refined. The failure of Zhuque-3 Yao-1, in fact, helped validate the hardest parts of the process. The team collected critical data on supersonic reentry, aerodynamic gliding, and high-precision guidance and control.
LandSpace said its post-flight analysis confirmed high-precision guidance and control toward the designated landing point during both the reentry burn and aerodynamic-glide phases. The combined control strategy using the cold-gas attitude-control system and grid fins was also fully validated. The only part that had not yet been proven was the final touchdown.
From Yao-1 to Yao-2, Zhuque-3 underwent several key recovery-related upgrades. The number of engines used during the landing phase was reduced to simplify the system. The vehicle’s autonomous flight safety system was enhanced with a predicted-landing-point safety function. Return-flight control strategies were comprehensively optimized, while its ability to withstand and protect against the complex thermal and mechanical environment of reentry was significantly improved.
Behind these changes was a painstaking process of analyzing and repeatedly validating the data from the failed maiden flight. This time, the missing pieces were filled in.
The two approaches are not mutually exclusive; they can complement each other. Offshore net-based recovery is better suited to high-payload missions, while landing-leg recovery offers greater technological maturity and deployment flexibility. China has now become the only country in the world to have demonstrated both recovery approaches.
How Capable Is Zhuque-3? A Comparison with Falcon 9
Zhuque-3 is not a simple copy of SpaceX’s Falcon 9. The two vehicles are similar in overall scale. Zhuque-3 is 76.6 meters long, 4.5 meters in diameter, has a liftoff mass of about 660 tonnes, and produces roughly 900 tonnes of liftoff thrust. Falcon 9 Block 5 is 70 meters long, 3.7 meters in diameter, and has a liftoff mass of about 549 tonnes.
In terms of physical dimensions, Zhuque-3 is actually somewhat larger than Falcon 9.
Payload capacity puts the two vehicles in broadly the same class. Zhuque-3 can deliver 21.3 tonnes to low Earth orbit in expendable configuration and 18.3 tonnes when recovering the vehicle downrange. Falcon 9 can deliver roughly 17–18 tonnes to LEO in recovery mode and about 22.8 tonnes in expendable mode.
What has industry watchers particularly excited, however, is the question of cost. The fundamental promise of reusability is to bring launch costs down. Zhang Changwu, LandSpace’s founder and CEO, has said that Zhuque-3 is expected to bring launch costs below RMB 20,000 per kilogram.
What does that mean? Conventional expendable liquid-fueled rockets can cost tens of thousands of dollars per kilogram to launch. Falcon 9 is currently priced at roughly $3,000 per kilogram, or about RMB 21,000. Zhuque-3’s cost target is therefore already in the same ballpark as Falcon 9.
What is the secret to lowering costs? Technologically, Zhuque-3 is following a development path similar to that of the next generation of large reusable rockets: a stainless-steel airframe powered by liquid oxygen and methane. This differs from the aluminum-lithium alloy structure and liquid-oxygen/kerosene propulsion used by earlier Falcon 9 vehicles.
Zhuque-3 uses stainless steel as its primary structural material, making it China’s first large liquid-fueled launch vehicle to use stainless-steel propellant tanks. High-strength stainless steel costs only RMB 30–40 per kilogram, roughly one-fifth the cost of aluminum alloy. It is not only cheaper, but also resistant to high temperatures and corrosion, making it well suited to large-scale industrial production. SpaceX has likewise adopted stainless steel for Starship.
Zhuque-3 is powered by liquid-oxygen/methane engines. Methane burns cleanly with little to no coking, meaning recovered first-stage engines do not need to be completely disassembled for overhaul. Fuel accounts for only 1%–3% of launch costs. The more times a vehicle is reused, the more its cost is spread across individual launches.
By design, Zhuque-3’s first stage is intended to be reusable at least 20 times. Falcon 9 currently holds the record for the highest number of flights by a single booster, at 34.
On December 30, 2011, Musk wrote that SpaceX had pushed Boeing, Lockheed Martin, Europe’s Ariane program, and Russia’s Proton and Soyuz programs to the brink, and that the endgame of the race would be determined by China.
At the time, China’s private commercial space industry barely existed. Fourteen years later, China not only has Zhuque-3, a vehicle positioned against Falcon 9, but has also demonstrated two different rocket-recovery technologies. Musk’s prediction is now being tested by time.
China’s “Double Insurance”: While Others Choose a Route, China Has Pursued Both
There is another notable aspect of Zhuque-3’s success: China has become the first country in the world to demonstrate both offshore net-based recovery and land-based landing-leg recovery for rockets.
Globally, SpaceX in the United States has pursued landing-leg recovery, while Europe and Japan have largely been following the same path. China, by contrast, has now validated two fundamentally different technological approaches in parallel.
The Long March 10B’s net-based recovery is an approach developed in China. The rocket eliminates heavy landing legs and instead uses foldable hooks mounted near the bottom of the vehicle, which are captured by a flexible arresting net.
Zhuque-3’s landing-leg recovery follows a technological path similar to that of SpaceX’s Falcon 9. Pursuing both approaches in parallel gives China greater strategic flexibility in reusable launch vehicles. If one approach encounters a bottleneck, the other can provide an alternative; different mission requirements can be matched with different recovery systems.
This is more than simply having “one more option.” It increases the redundancy and resilience of the broader industrial system.
The two approaches are being advanced primarily by China’s private commercial space sector and state-led space programs, respectively, creating a combination of competition and coordination. Zhuque-3 has already been included on the supplier list for China’s national satellite internet constellation and has secured contracts for batch launches of low-Earth-orbit satellites. It also won a launch-service contract to deploy 18 satellites in a single launch.
The private sector’s share of, and influence within, China’s commercial launch market is rising rapidly.
From “Coming Back” to “Flying Again”: How Far Away Is True Reusability?
The rocket has been recovered, but there is still a significant distance to go before it can be considered truly reusable.
Dong Kai, deputy chief designer of Zhuque-3, put it candidly: compared with SpaceX’s Falcon 9, LandSpace is currently around Level 6 on a nine-level scale—having completed laboratory validation and only just entering Level 7, the stage of physical flight validation.
He offered a metaphor from Chinese cultivation fiction: if Falcon 9 is a powerful master who has reached the “Spirit Transformation” realm, Zhuque-3 is, at best, in the “Core Formation” stage. Only once it has actually flown again after recovery will it have reached the “Nascent Soul” stage.
The next challenges are straightforward but critical: Can the recovered vehicle fly again? What condition are the airframe, propellant tanks, and critical welded structures in? How much service life remains in the engines? For a high-frequency launch system, inspecting a recovered rocket for one week versus three months makes a world of difference.
Dong Kai said the team is working to reuse the recovered rocket as soon as possible. LandSpace plans to fly the recovered vehicle again within six months.
But the significance of Yao-2’s success extends beyond the technology itself. The flight validated the complete process of returning a first stage from high-speed flight to a land-based landing site. It achieved China’s first recovery of a launch vehicle’s first stage using landing legs, as well as the country’s first land recovery of the first stage of an orbital-class launch vehicle.
From the launch vehicle’s project approval in August 2023, to the Yao-1 maiden flight in December 2025, and now to Yao-2 standing upright after landing, a private company has completed the entire journey from drawings to recovery in less than three years. It failed only once—and that failure showed the team exactly how to take the final step.
The head of commercial launch services at the European Space Agency reportedly described Zhuque-3 as “the first non-U.S. rocket to match or even surpass Falcon 9 on the major performance metrics.”
Looking back at Musk’s 2011 post makes the moment all the more striking: the endgame of the race, he said, would be all about China.
Fourteen years later, that post has received an answer. China has not been absent from this endgame. It is now participating deeply, on its own terms, at its own pace, and along its own technological paths.
From “catching with a net” to “standing on legs,” Chinese rockets have answered the same question in two different ways: not only can we send rockets up—we can bring them back.
While others are still deciding which path to take, China has put both approaches into practice. China can learn from roads others have already traveled, while also forging paths of its own.
Net-based recovery is China’s original answer; landing-leg recovery is its implementation of a proven international approach. With both technologies in hand, China’s reusable-rocket capabilities may prove more substantial than many expect.
From catching up to running in parallel, China’s reusable launch vehicles are now accelerating along two different paths. On the Gobi, the rocket standing firmly on its landing pad has already sent a message to the world: China has entered this endgame over space—and it has stood up.
Editor: Grace
