Dai Zheng's Zhuque-3 Technology News: From Failed Brake to Historic Landing

Dai Zheng watched Zhuque-3 miss its landing in December 2025, then saw its successor stand upright eight months later. The second flight changed this technology news story from an interview about failure into evidence that LandSpace could correct a specific flight-control problem.

Zhuque-3 Y2 lifted off on August 19, 2026, at 7:35 a.m. China Standard Time. Its upper stage reached orbit, while the first stage returned under engine power and landed vertically. It was China’s first recovery of an orbital-class booster on land using deployable landing legs.

That outcome gives new meaning to Dai’s widely shared interview after the first flight. He described the original failure as a braking problem during the final three kilometers. LandSpace had reached orbit, but the booster lost control during its last powered descent.

The new landing does not place LandSpace beside SpaceX as an operational equal. Falcon 9 has accumulated years of launch, landing, refurbishment, and reflight data. However, Zhuque-3 has now crossed the line separating recovery plans from recovered orbital hardware.

The Interview Predicted the Next Zhuque-3 Technology News

Dai Zheng framed the first landing failure as a bounded engineering problem, and the second flight tested that diagnosis.

Zhuque-3 made its debut on December 3, 2025, from the Dongfeng Commercial Space Innovation Pilot Zone near Jiuquan. Its second stage entered the planned orbit, confirming that the new launch vehicle could perform its primary transportation task.

The first stage followed a different path. It separated about 130 seconds after liftoff and began its return toward a dedicated landing zone in Minqin County. The square landing area measured 60 meters on each side, according to the subsequent CCTV interview.

The booster reached the vicinity of the pad, but it did not complete a controlled touchdown. Video showed a high-speed impact and explosion during the landing burn. China’s national space agency later stated that the mission had not achieved its planned first-stage recovery objective.

Dai resisted treating the entire mission as a failure. Customers first need a rocket to deliver payloads into orbit, he argued. Booster recovery serves a different objective, reducing hardware consumption and supporting lower-cost, higher-frequency operations.

His distinction was technically defensible, although commercially incomplete. Orbital insertion validated the rocket’s expendable function. It did not establish that LandSpace could recover, inspect, or reuse its most expensive stage.

Dai described the landing attempt as experimental rather than a mission that required first-flight recovery. From an engineering perspective, he said, collecting real flight data already represented substantial progress.

His most memorable explanation concerned the final descent. He compared the problem to failing to apply the last brake correctly during the final three kilometers. The statement reduced a dramatic explosion to a precise development task.

That framing mattered because reusable rockets depend on staged risk reduction. Engineers must first control ascent, separation, atmospheric return, aerodynamic steering, engine restart, terminal guidance, and touchdown. A failure near the end can still provide data from almost every earlier phase.

The interview was first published in December 2025, not August 2026. Its renewed visibility followed the successful Y2 landing, which explains why the conversation returned to Bilibili’s hot-search rankings.

The underlying event behind the current trend is therefore not a newly recorded interview. It is the August 19 flight, which supplied a result that viewers could compare directly with Dai’s earlier diagnosis.

LandSpace says the Y2 mission achieved both orbital insertion and first-stage recovery. Established reporting independently confirmed the booster’s upright landing and the mission’s significance for China’s commercial launch sector.

The timeline turns the interview into a rare before-and-after engineering record. Dai identified the incomplete final braking phase after Y1. LandSpace then returned with another vehicle and completed the landing on its next flight.

That does not reveal every change made between vehicles. The company has not published a comprehensive failure report or a detailed list of modifications. It does show that its engineers had enough information to produce a materially different outcome.

What Changed Between Failure and Touchdown

Zhuque-3 Y2 completed the same mission sequence that defeated Y1, including the powered landing and stable touchdown.

The second Zhuque-3 launched from the same commercial test zone in northwest China. Approximately 137 seconds after liftoff, its stages separated. The upper stage continued toward orbit while the first stage began its controlled return.

During descent, the booster used aerodynamic and propulsion systems to manage its position, orientation, and velocity. Grid fins, which are steerable lattice-shaped control surfaces, helped guide it through the atmosphere.

The engines then performed the final braking maneuver. Landing legs deployed before contact, and the first stage settled vertically onto the prepared surface.

According to mission coverage, the flight carried the Honghu 03 satellite and completed orbital insertion. The booster remained standing after touchdown.

This combination matters. A low-altitude test vehicle can demonstrate vertical takeoff and landing without experiencing an orbital booster’s speed, heating, structural loads, and long return path. Zhuque-3 Y2 completed the recovery sequence as part of a real orbital mission.

LandSpace had conducted smaller demonstrations before attempting the full profile. A Zhuque-3 test vehicle completed a low-altitude vertical flight in January 2024. That test examined engine throttling, guidance, control, and vertical recovery at modest speed.

A second test in September 2024 climbed to roughly 10 kilometers before returning. The vertical landing test expanded the flight envelope and provided data for the eventual orbital vehicle.

Those demonstrations lowered development risk, but they did not make the Y1 result automatic. An orbital first stage must manage conditions that a short test article never encounters. Its engines, tanks, control system, and structure must work across ascent and descent.

Zhuque-3 uses liquid methane and liquid oxygen, commonly shortened to methalox. Methane burns more cleanly than rocket-grade kerosene, which can reduce carbon deposits inside engines during repeated operation.

That characteristic supports reuse, but fuel choice does not guarantee rapid refurbishment. Engines still experience severe heat, pressure, vibration, and changing aerodynamic forces. Ground teams must inspect recovered hardware before determining whether it can fly again.

The vehicle also uses a stainless-steel structure. Steel tolerates high temperatures and can simplify some manufacturing processes, although it carries different mass and fabrication tradeoffs from aluminum alloys.

LandSpace’s mature Zhuque-3 design targets substantial payload capacity. The reported figure reaches 18.3 metric tons to low Earth orbit when the booster is recovered downrange. An expendable configuration can carry more because it reserves no propellant for return.

Payload figures describe design capability, not demonstrated routine performance. Y2’s importance comes from completing the recovery path, not from proving every planned payload mode.

The contrast with Y1 is still sharp. The first vehicle reached the landing area but arrived without enough controlled deceleration. The second performed the terminal maneuver and transferred its weight onto the landing legs.

That is a mechanism-level change, not merely a public-relations reversal. Guidance software had to command the right trajectory. Engines had to restart and throttle correctly. Navigation sensors had to estimate motion accurately enough for a narrow landing target.

The result supports Dai’s claim that Y1 had already completed much of the recovery sequence. However, the touchdown also demonstrates why the final phase cannot be dismissed as a minor detail. A booster is not recoverable until it stops without destroying itself.

One Landing Is Not Yet Reuse

LandSpace has proved recovery, but commercial reuse begins when the same booster flies again with predictable cost and schedule.

The August landing produced intact hardware that engineers can examine. That access is one of the flight’s most valuable outcomes, because telemetry cannot expose every physical effect.

Teams can inspect engine components, propellant systems, welds, thermal protection, landing legs, and aerodynamic surfaces. They can compare measured wear with preflight models and decide which parts require servicing.

A recovered stage therefore becomes both an asset and a test specimen. If its condition matches expectations, LandSpace gains confidence in its design assumptions. Unexpected damage would identify another development cycle before reflight.

The next decisive question is whether Y2 can fly again. Recovery and reuse are related, but they are not interchangeable.

A booster can survive landing yet demand months of inspection and extensive part replacement. In that case, recovery may provide engineering value without producing an attractive commercial service.

Useful reusability requires repeatability. The operator must land boosters across different missions, process them efficiently, and return them to flight without compromising reliability.

Turnaround time also matters. A reusable stage sitting in a factory does not increase launch capacity. The operational system must coordinate payload integration, range access, propellant supply, maintenance, and vehicle scheduling.

This is where Falcon 9 remains the central benchmark. SpaceX first landed a Falcon 9 orbital booster in December 2015. It then had to demonstrate reflight, routine recovery, multiple landing modes, and steadily rising use counts.

By 2026, Falcon 9 represented an industrial system rather than a single reusable vehicle. Its advantage included experienced teams, established refurbishment procedures, multiple launch sites, drone ships, and a large internal satellite customer.

LandSpace has now reproduced one highly visible part of that model. It has not reproduced its flight cadence or accumulated operational history.

The distinction prevents this technology news from turning into a premature parity claim. Zhuque-3’s landing is significant because it creates a credible path toward reuse, not because it closes the entire gap.

The comparison also includes important design differences. Falcon 9 burns liquid oxygen and rocket-grade kerosene. Zhuque-3 uses methane with liquid oxygen.

Methane can support cleaner engine operation, especially when designers intend repeated ignition and reuse. However, SpaceX has shown that kerosene-fueled boosters can also achieve high flight counts.

Propellant choice therefore establishes a maintenance hypothesis. Reflight data will determine whether LandSpace translates that hypothesis into lower servicing requirements.

Zhuque-3 also competes within China, not only against SpaceX. State-owned and commercial developers are pursuing reusable launchers as large satellite constellations create demand for more orbital capacity.

The Long March 12A reached orbit on its debut in December 2025 but failed during its recovery attempt. Other private companies, including Space Pioneer, Deep Blue Aerospace, Galactic Energy, iSpace, Orienspace, and CAS Space, have reusable projects under development.

China completed its first orbital-stage recovery in July 2026 using the Long March 10B. That booster descended toward a ship and entered a net-based recovery system.

Zhuque-3’s August flight followed a different architecture. It landed under its own power on solid ground using deployable legs, making it the country’s first orbital-class landing of that kind.

The two methods address the same economic problem through different systems. Net capture can reduce the mass assigned to landing gear. Legged landing gives the booster an integrated touchdown mechanism but carries that hardware throughout flight.

Neither method wins on spectacle alone. Their value depends on recovery reliability, refurbishment needs, payload penalties, and the number of missions each system can support.

LandSpace now holds an early lead among Chinese private companies because it has placed a recoverable orbital booster back on the ground. That lead can shrink quickly if competitors achieve higher cadence or faster reuse.

The Pressure Moves From Engineering to Operations

Zhuque-3’s next challenge is proving that a successful landing can expand launch supply rather than remain an isolated milestone.

China’s commercial launch market is being shaped by demand for large low-Earth-orbit constellations. These networks require repeated deployment missions, replacement satellites, and continuing capacity additions.

An expendable rocket discards engines, tanks, avionics, and structure during every launch. Manufacturing must replace the entire first stage before the next mission.

A reusable system changes that production equation. Factories can concentrate on upper stages, new fleet growth, and replacement hardware while recovered boosters return to service.

That benefit appears only when the recovery rate is high enough. Failed landings consume vehicles, while lengthy refurbishment limits how often each stage can contribute.

Zhuque-3’s successful touchdown pressures LandSpace to publish results beyond launch-day video. Customers will want evidence about inspection findings, scheduled reflight, payload performance, and mission reliability.

It also pressures competitors. A planned reusable rocket is now being measured against a vehicle that has reached orbit twice and landed once.

The state-owned launch sector faces a similar shift. Traditional programs prioritize high mission assurance, but constellation deployment rewards both reliability and frequency. Reusability adds rapid iteration and fleet management to that equation.

Dai discussed this cultural tension after the first flight. He pointed to SpaceX’s willingness to push vehicles toward their limits and use failures to identify boundaries.

That approach can accelerate learning when failures occur in controlled tests. It becomes more complicated when commercial payloads, public safety, or customer schedules are involved.

LandSpace cannot treat every launch as an open-ended experiment. It must separate development risk from the service reliability expected by satellite operators.

Y2 suggests the company can learn quickly from failure. The larger test is whether it can preserve that speed while standardizing hardware and operating procedures.

Financial pressure accompanies the technical work. Recoverable rockets require landing infrastructure, additional software, reserve propellant, specialized teams, and postflight inspection facilities.

These investments can reduce cost across many missions. They can also increase near-term expense before a sufficient launch rate develops.

The business case therefore depends on demand as much as engineering. A reusable rocket needs enough missions to spread its development and infrastructure costs across repeated flights.

Large constellation programs can provide that demand. Yet their deployment schedules, satellite readiness, and government approvals influence how quickly launch providers can build cadence.

Zhuque-3 must also compete for payloads while its reusable system matures. Customers could select proven expendable rockets if schedule certainty matters more than future cost reductions.

The Y2 landing improves LandSpace’s position in those negotiations. It provides a visible capability that competitors still need to demonstrate.

However, an intact booster does not reveal the price of the service, the cost of refurbishment, or the reliability of the next launch. LandSpace has not released enough operating data to evaluate those factors independently.

Readers should therefore treat cost-reduction claims as targets. The physical landing makes those targets more credible, but only a reflight can begin validating them.

This distinction matters beyond rockets. Technology companies often describe a successful demonstration as proof of a scalable product. Engineering history shows that scaling introduces supply, maintenance, reliability, and scheduling problems that demonstrations do not encounter.

Zhuque-3 has advanced from demonstration toward engineering application, as LandSpace described the Y2 mission. That wording is useful because it acknowledges the transition without claiming a mature service.

Three Signals Will Decide What the Landing Means

The next booster reflight, another successful recovery, and measurable launch cadence will determine whether Zhuque-3 changes the market.

The first signal is the condition of the recovered Y2 booster. LandSpace should eventually reveal whether major components remained within design limits and whether the stage is approved for another mission.

A scheduled reflight would strengthen the claim that Zhuque-3 is reusable rather than merely recoverable. Flying the same stage would test inspection procedures, component durability, and confidence in postflight analysis.

Failure to refly Y2 would not erase the landing. It would indicate that recovery exposed damage, maintenance demands, or certification concerns that still separate the vehicle from routine reuse.

The second signal is another successful landing. One touchdown can result from a capable system operating under favorable conditions. Repetition begins to establish reliability.

A later flight should demonstrate that the guidance, propulsion, and landing systems work across normal mission variation. Different payload masses and trajectories can change the energy available for recovery.

Another successful landing would also show that Y2’s result was not tied to one specially prepared vehicle. It would support the idea that LandSpace can produce reusable boosters consistently.

A new landing failure would reveal how narrow the system’s current operating margin remains. Reusable rockets can tolerate occasional losses, but frequent failures weaken both economics and customer confidence.

The third signal is launch cadence. LandSpace must move from occasional development flights toward a schedule that serves real deployment demand.

Cadence includes more than counting launches. It reflects manufacturing output, range availability, mission integration, booster processing, and customer readiness.

A rising flight rate with recovered hardware would place pressure on other Chinese launch providers. It would also give LandSpace the data needed to improve reliability and refine maintenance.

A slow schedule would limit that learning loop. Even a well-designed reusable rocket develops slowly if it flies only once or twice each year.

These three signals should be evaluated in order. First inspect and qualify the recovered booster. Then repeat the landing. Finally show that reuse supports a sustained launch tempo.

The August 19 result strengthens Dai’s explanation of the first failure. His “last brake” description now looks less like an attempt to soften a loss and more like a concise account of an incomplete sequence.

Still, the successful second landing does not validate every company claim. LandSpace must disclose more evidence about hardware condition, reuse, and operating efficiency.

That uncertainty is the most important part of the story. The landing settled whether Zhuque-3 could return an orbital booster safely. It did not settle how often, how cheaply, or how reliably the company can repeat the process.

For North American readers following technology news, Zhuque-3 offers a useful warning against treating reusable launch as a permanent one-company category. The underlying techniques are spreading, even though operational experience remains uneven.

SpaceX still holds the strongest reusable launch record. Its advantage rests on accumulated missions and integrated operations, not only an early landing.

LandSpace’s progress matters because competition begins before two companies achieve equal scale. A credible alternative can influence satellite planning, government investment, supplier capacity, and rival development schedules.

Dai Zheng’s interview became popular again because audiences could compare words spoken after failure with hardware standing after flight. The contrast delivers a cleaner narrative than most engineering programs provide.

The next chapter will be less cinematic. Inspection reports, reflight decisions, maintenance work, and launch schedules rarely generate the same attention as a landing video.

Those details will decide whether Zhuque-3 becomes an operational transportation system. They will also determine whether this technology news marks a durable change or one impressive flight.

Watch what LandSpace does with the recovered booster, not only what it says about it. If Y2 returns to flight, later stages land reliably, and cadence rises, China will have more than a reusable rocket demonstration. It will have the beginnings of a reusable launch service.