Catching Itself: Zhuque-3 and the Industrial Chain Behind China's Private Reusable Rockets

On December 3, 2025, Zhuque-3 (ZQ-3) Y1 lifted off from Jiuquan; its second stage successfully reached orbit, but the first stage burned abnormally in the final dozen-odd seconds before touchdown and failed to stand upright—this was China's first attempt to recover the first stage of an orbital-class launch vehicle. Half a year later, in the summer of 2026, the Y2 vehicle, having completed its static-fire test, is now heading for that same goal: to achieve China's first private-sector recovery of an orbital-class first stage. Just as it awaits launch, the national team's Long March 10B (CZ-10B) has already used a net at sea to pull off the world's first net-capture recovery.

Why does a rocket that has to "catch itself" move all of Chinese manufacturing? This report takes Zhuque-3 as its main thread—running from the engine, materials, control, and reuse economics, to the demand side of constellation deployment, a comparison of China's two recovery routes, the progress spectrum of the private-sector tier, and the coordinates of the international race, and then to a supply chain that is now opening up to Chinese factories—in an attempt to answer one judgment: reusability is, on its surface, a question of aerospace technology, but at its core it is a question of manufacturing—whoever can turn a rocket from a handcrafted work of art into a factory product holds the ticket to the low-orbit era. The full text runs to roughly 30,000 characters; all data has been verified across multiple sources, current as of July 2026.

1. Y2 on the Pad: A Launch Aimed at "Catching Itself"

On June 29, 2026, at the test stand of the Dongfeng Commercial Aerospace Innovation Test Zone in Jiuquan, Gansu, a 66.1-meter-tall rocket was firmly anchored to the ground as nine engines ignited simultaneously, the exhaust plume lighting up half the Gobi. This was not a launch but a static-fire test—the rocket stays put while the engines run at full load, verifying every ground system from propellant loading, ignition sequencing, and nine-engine cluster coordination to servo gimbaling. People's Daily Online confirmed that all systems worked normally and that ground validation was fully completed. For LandSpace, this Zhuque-3 vehicle, code-named "Y2," has now completed its final major ground exam before flight.

What truly draws attention is not how high it will fly, but where it will return to.

A Maneuver Harder Than Reaching Orbit

Half a year earlier, Zhuque-3's first vehicle, "Y1," had already proven it could deliver a payload to orbit. At noon on December 3, 2025, Y1 lifted off from Jiuquan; stage separation, second-stage engine ignition, payload fairing separation, second-stage shutdown, an extended coast, and second-stage engine reignition—LandSpace's official release recorded this string of maneuvers verbatim as "all completed as planned," with the second stage entering its designated orbit at 200 kilometers altitude and 55 degrees inclination. By the traditional standard of "putting something into space," Y1 was a success.

But Y1 also aimed to do something no one in the history of Chinese spaceflight had ever pulled off: to have the first stage—which had completed its mission and separated from the second stage—fly itself back to the ground and stand steadily upright. This step failed—the rocket suffered an anomaly after igniting for the landing phase, and its debris fell at the edge of the recovery pad. So Y2 took up the baton, its core objective reducible to a single sentence: to strive to achieve China's first successful recovery of the first stage of an orbital-class launch vehicle. Xinhua's English edition, in reporting the static-fire test, made the weight of this mission clear.

To understand how hard this is, one must first distinguish two maneuvers. Putting a rocket into orbit is about letting it leave: the engines accelerate it to first cosmic velocity and fling it into space, after which no one tends to the fate of the airframe—it is left to plunge into the atmosphere and burn up. Bringing a first stage back is about having it—after burning off most of its fuel, re-entering the dense atmosphere at several times the speed of sound, and enduring aerodynamic heating and enormous dynamic pressure—reignite its engines to bleed the velocity from hundreds of meters per second down to nearly zero, while using grid fins and a reaction control system (RCS) to hold its attitude steady, and finally touch down gently in a near-vertical posture at one or two meters per second—this is about having it come back. Humanity has done the former for seventy years; the latter was not achieved until SpaceX's Falcon 9 pulled it off for the first time in late 2015.

The difficulty is that what recovery demands of a rocket is almost the opposite of what reaching orbit demands. Reaching orbit calls for "piling on velocity in one go, as fiercely as possible"; recovery calls for "bleeding velocity back off precisely, controllably, and repeatably." A rocket built only to reach orbit can treat its first stage as a single-use consumable, with structure pared to the extreme minimum weight and engines run at maximum ferocity, discarded once used; but the moment it must fly back and be reused, it has to carry dead weight that is only useful for "coming home"—landing legs, grid fins, a reaction control system—the engines must be able to throttle deeply and reignite during a high-speed descent, and the airframe structure must withstand repeated thermal cycles and landing impacts. The entire rocket's design logic, manufacturing method, and cost structure have to be rewritten for the two words "reuse." What Y2 is charging at is precisely this watershed from "use once and discard" to "use and reuse."

After the Static Fire: An Undecided Window

Passing the static-fire test usually means launch is not far off. But as of the time of writing in mid-July 2026, Y2's launch window remains up in the air. LandSpace's initial line was "another recovery test in the first half of the year, with a first recovery-and-reuse flight to be attempted in Q4 as circumstances allow," as Zhuque-3 chief designer Zhang Xiaodong put it at an industry conference in April; Sina Tech recorded this timeline. The window has since slipped repeatedly: from the year's-start "first half of the year," it drifted to mid-May, then to early July, and further reports said the recovery-verification window had been pushed to August.

The vehicle's own preparation has not been slow. After Y1's late-2025 setback, LandSpace completed its technical root-cause resolution and iteration in just over four months; the announcement on Space Day stated that Y2 had entered pre-delivery preparation, after which it was transported nearly 4,000 kilometers by road from the rocket manufacturing base in Jiaxing, Zhejiang, to Jiuquan. What is truly holding things up is the intensive verification of recovery itself. Following the supersonic re-entry aerodynamic glide that Y1 had already run through, Y2 must add the final two phases that Y1 failed to verify: high-precision landing guidance, and landing-leg deployment and soft touchdown. Any unclean set of ground data is reason enough to push the window back further.

Our judgment is that, rather than rushing a date, LandSpace cares more about whether it can complete the full "catch itself" maneuver this time. Behind this caution lies a reality: private-sector spaceflight cannot afford a string of failures that "blow up for the whole world to see." Footage of Y1's debris catching fire on landing has circulated on social media for half a year, and if Y2 stumbles again on that final step, it would be real pressure on the company's valuation, financing, and IPO process. As for the various precise launch dates circulating outside, most come from estimates by third-party launch-tracking sites; LandSpace itself has not fixed a date, and this report holds uniformly to "not yet launched as of July 2026," without chasing a date that could change at any moment.

Two Chinese Routes, Showing Their Hands in the Same Summer

Intriguingly, less than two weeks around Y2's static fire, another Chinese rocket had already pulled off "recovery"—only by an entirely different route. On July 10, 2026, the Long March 10B (CZ-10B) made its maiden flight from the Hainan (Wenchang) Commercial Spacecraft Launch Site; as the satellite reached orbit, the first stage was steadily caught by a recovery vessel named "Linghangzhe" (Navigator) using an outstretched cable net. Xinhuanet called this China's first controlled recovery of a launch vehicle's first stage, and the world's first net-capture recovery.

So in this summer of 2026, China's reusable rockets present two faces at once. One is the national team's Long March 10B: it does not have the rocket stand upright by itself but simply does away with landing legs, fits a few hooks on the airframe, and offloads the hard problem of "recovery" from the rocket onto a net at sea and a 25,000-ton recovery vessel. The other is the private-sector's Zhuque-3: it faithfully follows the path SpaceX blazed, growing grid fins, landing legs, and a reaction control system on the rocket, and using the reverse thrust of its own nine engines to fly back to the ground and stand vertically.

Each route has its trade-offs—net-capture recovery has looser landing-accuracy requirements and saves the dead weight of landing legs, at the cost of building a dedicated recovery vessel and depending on sea-state windows; vertical recovery demands extremely high guidance precision and sacrifices some payload capacity to carry the fuel and hardware needed for recovery, with the upside of not depending on any offshore facility and, in theory, the fastest turnaround. We will compare these two routes specifically in Chapter 7 below. But for now it is worth remembering that they point to the same goal—turning the first stage from something "thrown away" into something "kept." China has not bet everything on one route; instead the national team and the private sector each take one, handing in their answers together in the same summer.

Who LandSpace Is: From a High-Stakes Bet to "First Commercial Aerospace Stock"

Standing behind Y2 is LandSpace, full name LandSpace Technology Corporation Ltd., founded in June 2015 and registered in the Yizhuang Economic and Technological Development Zone in Beijing, its founder Zhang Changwu a finance man with a Tsinghua background. In a field long dominated by the national team, for a private company to build its own engines, build its own rockets, and take on vertical recovery—which even many major spacefaring nations have not achieved—was itself a high-stakes bet.

LandSpace has won one crucial bet. On July 12, 2023, its previous-generation rocket Zhuque-2 lifted off from Jiuquan and successfully reached orbit, becoming the world's first liquid-oxygen/methane (methalox) launch vehicle to do so. Xinhua spelled out the value of this "world first" in its report: before this, both the U.S. company Relativity's Terran 1 and SpaceX's Starship—two methalox rockets—had failed in their bids for orbit, and LandSpace beat them to it. This victory not only proved the methalox technical route viable, but also pushed LandSpace into the top tier of China's private spaceflight—Zhuque-3's choice of methalox stands precisely on the path Zhuque-2 blazed, and we will discuss in detail in Chapter 3 why this route suits reuse.

Today's LandSpace is no longer just a "laboratory company." It has its headquarters in Beijing, does R&D in Xi'an and Shanghai, has built an engine manufacturing base in Huzhou, Zhejiang, and a rocket assembly-and-test base in Jiaxing, Zhejiang, and is laying out a new reusable-airframe base under construction in Wuxi, Jiangsu—a manufacturing map covering R&D, propulsion, and final assembly is now spread out. In 2025, LandSpace launched its STAR Market IPO, going for "first commercial aerospace stock" (valuation and other capital-side figures are treated here only as industry background and not elaborated). It is precisely for this reason that the significance of Y2's flight has long exceeded that of a single rocket: it is the answer sheet handed by a company that has staked its very existence on "reusability" to the market, to investors, and to the entire industry.

Why a Rocket Is a Manufacturing Question

Lift the gaze from the launch pad, and one finds a judgment easily obscured by the spectacle of fire: reusability is, on its surface, a question of aerospace technology, but at its core it is a question of manufacturing.

The reason a rocket is expensive is not that fuel is expensive—the propellant cost of a single Zhuque-3 launch is a rounding error; what is expensive is the airframe that gets burned up after one use, especially the first stage, which is stuffed with nine engines and accounts for about 70% of the whole rocket's manufacturing cost. Cailianpress (CLS), citing industry figures, points out that the first-stage body itself accounts for more than 70% of total cost—which is why rocket companies the world over are scrambling to recover it. But the moment you decide to recover it and reuse it repeatedly, the nature of the problem changes: from "how to build a finely crafted rocket that pays for itself in a single flight" to "how to build a batch of cheap, rugged, refurbishable rockets that can withstand repeated punishment." The former is aerospace engineering; the latter is manufacturing—it tests materials, welding, engine mass production, reliability, cost control, and supply-chain management. Whoever can turn a rocket from a handcrafted work of art into a factory product holds the entry ticket to the low-orbit era.

What best illustrates this shift on Zhuque-3 is its airframe material. LandSpace did not stick with the aluminum alloy customary in aerospace but chose stainless steel, welding its propellant tanks with a self-developed high-performance laser-welding production line. SpaceNews, quoting founder Zhang Changwu, calls this China's first stainless-steel liquid rocket. Stainless steel is heavier and more "unrefined," with nearly three times the density of aluminum alloy—from a pure aerospace standpoint it looks like a step backward; but it is cheaper—the manufacturing cost of a set of large-diameter thin-walled stainless-steel tanks is said to be about 80% lower than aluminum alloy—and it is more heat-resistant and better suited to repeated welding and refurbishment—a textbook case of "manufacturing thinking overriding aerospace thinking." The same logic is written into its engines: the first stage carries nine Tianque (TQ)-12A engines side by side, spreading unit cost thin through quantity and volume, rather than pursuing the ultimate performance of a single engine as traditional high-thrust rockets do. At LandSpace's engine base in Huzhou, Zhejiang, the 100th methalox engine rolled off the line in April 2025—treating engines as "industrial products" to be mass-produced is precisely the other side of the reuse logic.

This thread of "manufacturing thinking" will run through every chapter of this report to come: from the trade-offs of the methalox propellant route, to the craft of the stainless-steel airframe and landing legs, to the algorithms of engine guidance and control, to the ledger of reuse-driven cost reduction, and on to a complete supply chain now opening up to Chinese manufacturing. But before discussing engines, materials, and ledgers, we must first return to that noon in December 2025 and see clearly what exactly happened to Y1 in its final few kilometers and final dozen-odd seconds—because every change on Y2 grew out of that failure.

2. Y1 Recap: Those Few Kilometers of Orbital Success and Recovery Failure

The cruelest thing about a rocket is that it can do ninety-nine percent of the maneuvers right and still lose the whole game in the final dozen-odd seconds. Zhuque-3 Y1 is just such an example. On December 3, 2025, it accomplished nearly everything a reusable rocket ought to do—falling short only at the very last step: standing steadily upright.

To truly grasp why Y2 was changed the way it was, and what homework LandSpace has been catching up on these past six months, one must first take Y1's flight apart phase by phase—seeing clearly where it succeeded, where it did not, and exactly which link the "did not" got stuck on.

The Shot at High Noon

The launch time was 12:00 sharp Beijing time on December 3, 2025, at the Dongfeng Commercial Aerospace Innovation Test Zone of the Jiuquan Satellite Launch Center, pad code LC-96B. Xinhua's report gave these basic coordinates. The vehicle that flew was the baseline version of Zhuque-3—66.1 meters long overall, 4.5 meters in diameter, with a 5.2-meter-diameter payload fairing, a takeoff mass of about 560 tons, and takeoff thrust of 7,542 kilonewtons; the first stage carries nine Tianque (TQ)-12A methalox engines in parallel, and the second stage carries one TQ-15A vacuum-optimized engine.

Here we must first clear up a detail easily mixed up. Outsiders often describe Zhuque-3 as "the 76.6-meter big rocket," but those are actually the specs of the enhanced ZQ-3E, which has yet to make its maiden flight—76.6 meters, about 655 tons, 9,000 kilonewtons of thrust, with much higher payload capacity as well. Y1 and Y2 both fly the 66.1-meter baseline version. The two are not the same rocket, and wherever specific numbers are involved, this report holds uniformly to the officially published baseline specs, with LandSpace's official release as the primary basis. Confusing these two variants would graft the payload capacity of a rocket still on the drawing board onto one that has already flown—an error common in commercial-spaceflight reporting, worth nailing down from the outset.

First Half: A Textbook Orbital Insertion

Look first at the parts done right. After Y1 lifted off, stage separation proceeded normally, the second-stage engine ignited, the payload fairing separated, the second stage shut down and entered an extended coast, and then the second-stage engine reignited—for this whole sequence, LandSpace's release used the phrasing "all completed as planned." The second stage ultimately delivered the payload into a 200-kilometer × 200-kilometer near-Earth orbit at 55 degrees inclination; SpaceNews's report added the orbital parameters and the detail of "about a 1,400-second long coast plus reignition."

Do not underestimate those two words, "reignition." Having an engine shut down, coast for a stretch, and reignite in the vacuum, weightlessness, and deep cold of space is a hurdle many rockets cannot clear—propellant drifts about under weightlessness and cannot be stably supplied, and the slightest deviation in ignition timing ruins everything. The TQ-15A vacuum engine has a three-start capability, and Y1's in-orbit reignition amounted to verifying one of the second stage's key capabilities. As for the second stage itself, after completing its mission it re-entered the atmosphere around January 30, 2026, and fell into the South Pacific; china-in-space recorded its final resting place.

One point must be stated honestly: the payload. On this flight of Y1, the official release did not name what satellite was carried; the analytical line of the technical recap holds that it carried no actual satellite but rather a mass simulator, and performed no satellite-vehicle separation—this is analytical inference, not official confirmation, and this report labels it as such and does not treat it as settled. For a test flight whose primary goal was recovery verification, not risking a real satellite is a reasonable choice.

Second Half: Grid Fins Held Steady, the Final Seventeen Seconds Did Not

The real focus is the first stage. After separating from the second stage, this behemoth—stuffed with nine engines and accounting for about 70% of the whole rocket's cost—had to complete a "homecoming" journey alone. By LandSpace's design, this journey has three phases: first the re-entry burn, in which the engines reignite to pull the first stage from a high-speed descent back onto a controllable trajectory; then the supersonic re-entry aerodynamic glide, in which the airframe passes through the dense atmosphere at several times the speed of sound, using grid fins and a cold-gas reaction control system to adjust attitude and withstand maximum dynamic pressure; and finally the landing burn, in which the engines reignite to bleed velocity to near zero and, together with the landing legs, complete a soft landing.

The first two phases, Y1 walked beautifully. Sina Finance, citing LandSpace's recap, shows that the first stage successfully verified the structural thermal protection, overall aerodynamic configuration, and attitude control of the supersonic re-entry aerodynamic glide phase, as well as "the composite control strategy of cold-gas reaction control system plus grid fins," performing well under maximum dynamic pressure. CGTN also disclosed these details. In other words, the "gate of hell" stretch that most severely tests aerodynamics and thermal protection—the first stage returning from space at high speed and crossing the atmosphere—it made it through, with attitude controllable throughout.

The problem came in the final phase. The official characterization is written with great restraint: after igniting for the landing phase, the first stage suffered an anomaly, failed to achieve a soft landing on the recovery pad, and its debris landed at the edge of the recovery pad, the recovery test failing; the word Xinhua used was "abnormal combustion." English-language space media dug up a finer timeline: the landing-phase engines worked in the sequence "first ignite 1, then ignite 4, then switch back to 1," but failed about 17 seconds before touchdown; both SpaceNews and china-in-space recorded this. The first stage ultimately struck the ground at high speed, its impact point about 40 meters from the center of the landing zone, in the vicinity of Minqin County, Wuwei, Gansu, about 390 kilometers from Jiuquan.

As for the physical cause of the "abnormal combustion," there is various speculation going around—for instance, that during the first stage's high-speed descent, the engine nozzles faced a strong airflow that made landing ignition difficult or even "blew it out." But these are all street-level conjectures; as of the time of writing, LandSpace officially has still not given a root-cause conclusion, saying only that the specific cause is under analysis and investigation. We respect this boundary: that the landing ignition in the first stage's final sprint did not hold steady is a fact; as to whether it was combustion instability, thrust-response lag, or something else, what the officials have not said, this report will not say for them.

The Value of Failure: What Was Verified, and What Was Not

Placing Y1's flight within a "verification" framework yields a conclusion more useful than the binary of "success/failure."

What was verified is a long string of technologies never before flown in Chinese spaceflight: lifting-body re-entry aerodynamic control, composite guidance of grid fins plus cold-gas attitude control, pyrotechnic-free stage separation, flight-grade laser-welded stainless-steel tanks, cluster operation of nine methalox engines, and most crucially—the first stage re-entering at high speed from orbital-class velocity, crossing the maximum-dynamic-pressure region, and keeping its attitude controllable throughout. china-in-space laid out these "China firsts." These are no small matters; they are the foundation of the whole edifice of vertical recovery, and Y1 laid that foundation solid.

What was not verified is only the final two blocks: high-precision landing guidance (having the rocket drift precisely to the dead center of that landing point a few tens of meters on a side) and landing-leg deployment and soft touchdown (absorbing the impact of that final ground contact steadily). Y1's first stage fell precisely at the threshold of these two. So a more accurate way to put it is not "Y1's recovery failed," but "Y1 verified every link in the recovery chain except the last one, falling short only at the goal-line kick." It is precisely for this reason that this test flight was taken seriously internationally—CNN's headline called it a "historic first orbital test," even though the first stage ultimately exploded. The authoritative Chinese line characterized it as "China's first attempt to recover the first stage of an orbital-launch launch vehicle," while Zhuque-3 also became China's first launch vehicle designed for reusability to successfully reach orbit.

From Y1 to Y2: Just Over Four Months of Technical Root-Cause Resolution

After the failure, LandSpace moved fast. From Y1's setback in December 2025 to Y2 entering pre-delivery preparation in April 2026, only just over four months elapsed—during which LandSpace completed what outsiders call "technical root-cause resolution": tracing the cause of failure to the bottom layer by layer, iterating the relevant designs and processes, and re-verifying. LandSpace has not disclosed in detail what specific changes were made to Y2, saying only that it is "optimizing the landing procedure"; nor will we make concrete on its behalf the changes it has not announced.

But one clear logic can be seen from public information: Y1 already proved "it can come back and hold attitude steady," and what Y2 must make up is "landing accurately and standing firmly." This also explains why LandSpace has been especially cautious with Y2, willing to push the launch window back again and again—because what must be verified is precisely the landing phase, the very link Y1 stumbled on, and any anomaly in the ground data is reason enough to wait a bit longer. With one high-speed-touchdown failure, Y1 precisely localized the problem to the final dozen-odd seconds; what Y2 must do is walk through those final dozen-odd seconds too.

Worth mentioning is that the reason Y1 could carry the recovery chain this far back has to do with its conservative "margin-leaving" settings. According to analysis by financial media, Y1 shortened the airframe relative to the design ceiling, lowered the first stage's separation altitude, and switched to the more mature TQ-12A engine—using a "downgraded" rocket to run through the recovery procedure first and press risk to the minimum. This "small quick steps—run the procedure first, then max out the performance" approach is itself an engineering philosophy that sets private spaceflight apart from the national team. And the most central link of this philosophy is the engine—why Zhuque-3 dares to run nine engines in a cluster, why it chose methalox, and exactly where in the world Tianque as an engine stands, are the questions the next chapter will answer.

3. The Methalox Route: The Tianque Engine and the Propellant-Route Debate

If a reusable rocket has a heart, it is its engine. Zhuque-3's first stage carries nine Tianque (TQ)-12A engines side by side, and its second stage one TQ-15A; the whole rocket's thrust, its reuse potential, and even which fuel it chose all hinge on this "Tianque" family. To understand why Zhuque-3 grew into what it is today, one must first answer a more fundamental question: why does it burn methane?

The answer to this question draws out a quiet route debate that has played out in global aerospace propulsion over the past two decades.

Why Methane: A Route Debate About Fuel

For liquid-rocket fuel, the mainstream comes down to three routes: liquid-hydrogen/oxygen (hydrolox), liquid-oxygen/kerosene (kerolox), and liquid-oxygen/methane (methalox). Each has its own temperament.

Hydrolox has the highest specific impulse and is the most "energetic," the choice for the upper stages of big rockets like the Long March 5; but liquid hydrogen has extremely low density and needs extreme deep-cryogenic storage at minus 253 degrees, its tanks having to be built large and delicate, with very high cost and maintenance—not worthwhile for a reusable first stage that must be put through repeated punishment. Kerolox goes to the other extreme: kerosene has high density, is storable at room temperature, and is technically mature, and Falcon 9's Merlin engine burns kerosene; but kerosene "cokes and deposits carbon" at high temperatures—after combustion it forms a layer of coke on the engine's inner walls and plumbing, like the bottom of an old greasy wok; flying a rocket once is fine, but repeated use requires repeated cleaning or even parts replacement, adding major trouble to reuse.

Methalox lands right in the middle "sweet spot." LandSpace's technical head Yuan Yu, in an interview with Cailianpress (CLS), put methane's advantages very plainly: methalox "has no coking or carbon-deposit problem," is "low-cost and easy to obtain," and is "simple to service and maintain before reuse." Each of these lines points straight at reuse's pain points—methane burns clean, the engine's inner walls do not coke, and after one flight a wipe-down suffices for another; methane is widely sourced and cheap, and the fuel cost of a single launch is nearly negligible; methane can also regeneratively cool the engine without leaving residue. Even better, methane's boiling point is around minus 162 degrees, fairly close to liquid oxygen's minus 183 degrees, so the two propellants can share a "common-bulkhead" tank, making the airframe more compact. On density and specific impulse, methane sits between hydrogen and kerosene—a compromise optimum.

Methane has one more reason aimed at the more distant future: it can be produced in situ on Mars. This has no direct bearing on today's Zhuque-3, but it explains why the world's most aggressive companies have all, as if by agreement, bet on methane—SpaceX's Raptor, Blue Origin's BE-4, Europe's Prometheus, without exception. Yuan Yu likewise used "reusable engines newly developed internationally since 2010 generally adopt methalox" to attest to the directional soundness of this route. What LandSpace bets on with methane is not momentary performance, but the long-term economics of reuse.

The Tianque Lineage: From TQ-11 to TQ-15

Tianque (TQ) is LandSpace's self-developed family of methalox engines. Within this family, two mainstays and one "veteran" bear directly on Zhuque-3.

The mainstay is the Tianque-12 (TQ-12). This is a methalox engine using a gas-generator cycle, with sea-level thrust of about 658 kilonewtons (officially often called "the 67-ton class," with another figure of 670 kilonewtons), vacuum thrust of about 785 kilonewtons, and a chamber pressure of about 10.1 megapascals, completing its full-system test in May 2019. English Wikipedia records its identity as "China's first liquid rocket engine developed with private capital"—before this, China's liquid rocket engines came uniformly from the national team. Zhuque-3's first stage uses its improved variant, the TQ-12A.

The second stage uses the family's newest member—the Tianque-15A (TQ-15A). According to LandSpace and Jiemian News, this is the domestic vacuum-optimized methalox engine with the greatest thrust, with vacuum thrust of about 836 kilonewtons, and deep thrust-modulation capability from 55% to 110% plus a three-start capability. Its significance goes beyond high thrust: in the past, Zhuque-2's second stage relied on a main engine plus a small "vernier engine" named TQ-11 working together, whereas the TQ-15A does away with the vernier engine outright, controlling via post-pump gimbaling of the main engine, cutting the second stage's weight by about 400 kilograms—for a rocket, every kilogram shed from the second stage means a bit more payload capacity. This "veteran," the TQ-11 (a roughly 8-ton-class vernier engine), thus exits the main stage; it once served as the attitude-control mainstay of Zhuque-2's second stage.

The family also holds a technology reserve crucial to reuse: the pintle injector. The injector is the key component in an engine that atomizes and mixes fuel and oxidizer; the advantage of a pintle injector is that it is inherently suited to deep thrust modulation—and deep thrust modulation is precisely what is required to "bleed velocity precisely to near zero" during vertical recovery. LandSpace has already completed the domestic first methalox pintle injector to pass full-system hot-fire evaluation. But caution is needed here: public materials do not clearly state that the production TQ-12A on Zhuque-3 uses a pintle injector; there is no authoritative source text for the correspondence between the two, and this report states only that "LandSpace has mastered this technology," without overstepping to say "this engine is a pintle engine."

Nine in Parallel: Trading Quantity for Cost

The most intuitive feature of Zhuque-3's first stage is that it bundles nine Tianque-12A engines side by side. This is not for show, but a textbook manufacturing choice.

Traditional high-thrust rockets tend to use a few "strongman" engines—the greater the thrust per engine, the fewer needed, and the simpler the plumbing and control. But this path has two problems: first, developing a single super-high-thrust engine has a long cycle and high risk; second, once a single engine has a problem, the whole rocket is done. Paralleling many small engines follows the opposite philosophy: a single Tianque-12A's thrust is not top-tier, but with nine bundled together the first stage's total sea-level thrust reaches about 7,542 kilonewtons, enough to lift the 560-ton rocket. More crucially, this approach brings two benefits for the reuse era. The first is redundancy—in theory, if a few engines misbehave, the rest may still salvage the mission through thrust reallocation, a safety cushion for both crewed flight and reuse. The second is volume—nine engines per rocket means that building a few rockets a year amounts to demand for dozens to over a hundred engines, turning the engine from a "finely crafted single piece" into a "batch product on a production line," with unit cost spread thin as volume rises.

LandSpace's engine manufacturing base in the Nantaihu New Area of Huzhou, Zhejiang, was built precisely for this "mass-production" logic. According to China Aerospace News, this base, whose groundbreaking was in 2018, has the capacity to produce over a hundred methalox engines a year, and in April 2025 the 100th methalox engine rolled off the line here. An engine factory with "annual output of a hundred units" is itself the most concrete footnote to the phrase "building rockets as factory products." During the static fire, the 45 seconds in which the nine engines ignited in batches, worked stably, and completed servo gimbaling tested precisely whether this cluster coordination can be relied upon.

A Generational Gap That Must Be Stated Honestly

At this point, it is necessary to pour a bucket of cold water. Although Tianque-12 is the benchmark of China's private methalox engines, placing it on the world's coordinate system reveals a technical generational gap that cannot be avoided—and this gap lies not in thrust magnitude but in "cycle type."

An engine's "cycle type," put plainly, is how it drives the turbopump and how it wrings the propellant's energy to the utmost. Tianque-12 uses a gas-generator cycle (also called an open cycle): a small stream of propellant is burned separately to drive the turbopump, and after doing its work this gas is dumped out the side, meaning a small portion of propellant does not fully participate in the main combustion, taking a discount on efficiency. This is a mature, reliable, easy-to-master scheme, but it is a relatively basic tier among cycle types.

By contrast, SpaceX's Raptor uses a full-flow staged-combustion cycle—the most complex and most efficient tier among cycle types, wringing nearly every bit of propellant into the main combustion chamber, achieving sea-level thrust of about 2,254 kilonewtons; Blue Origin's BE-4 uses an oxygen-rich staged-combustion cycle, with thrust of about 2,400 kilonewtons, likewise a "closed," advanced cycle. Tianque-12's gas-generator open cycle is one to two tiers below these two. This means that on the engine's "internal skill" alone, there remains a real gap between Tianque and the world's most top-tier reusable engines.

How should this gap be viewed? Our judgment is: it is real, but there is no need to negate Tianque's value because of it. For Zhuque-3 today, using a mature, reliable engine that can be mass-produced, deeply throttled, and repeatedly ignited to first accomplish "vertical recovery" matters more than pursuing the ultimate cycle efficiency of a single engine. Upgrading the engine cycle is homework that can be caught up on step by step in later models; whereas once the system-level capability of "recovering the rocket and reusing it" is up and running, its value is far greater. Honestly acknowledging the generational gap is precisely how to see clearly the true position of China's private spaceflight—already seated at the table, but with a way to go before reaching the very top.

How an Engine Gets "Batch-Printed"

Tianque's ability to take the "mass-production" path also owes to the spread of one manufacturing process—metal additive manufacturing, commonly known as metal 3D printing.

The rocket engine is one of the most geometrically complex machines: the turbopump housing, the gas-generator body, the thrust chamber, the nozzle laced with dense regenerative cooling channels—many are irregular parts that traditional machining struggles to make, or that require many welds to piece together. Metal 3D printing prints these complex parts as "one-piece forms," saving a great deal of welding and assembly, shortening the cycle, and even building fine structures like cooling channels directly into the part. For private rockets chasing volume and cost, this is an almost tailor-made process.

On Tianque's manufacturing chain, one clear name can be found: BLT (Bright Laser Technologies) of Xi'an, Shaanxi. According to this company's public case studies, it undertook the metal 3D printing of the gas-generator body and combustion chamber of the Tianque engine; 3D Science Valley recorded this collaboration. BLT also named it in its 2025 half-year report, stating that it had helped LandSpace's Zhuque-3 reusable rocket with its first large-scale vertical-takeoff-and-landing flight test, driving key components from engineering validation toward mass production—this is one of the few suppliers found so far named at the company-announcement level in connection with Zhuque-3 itself. Beyond metal 3D printing, Tianque's manufacturing has also broken through a series of processes including large-nozzle laser welding, high-efficiency cryogenic pumps, turbopump fluid dynamic-pressure sealing, post-pump gimbaling, and propellant subcooling (these are mostly summaries at the official technical level, with specific material grades not disclosed). We will elaborate on this chain of "mass-producing engines with advanced manufacturing processes" in more detail in Chapter 11, as a concrete entry point for manufacturing to break into commercial spaceflight.

Engines solve the problem of "whether there is power and whether it can be reignited repeatedly." But a rocket meant to be recovered repeatedly needs more than a good heart—it also needs a "skeleton" that can withstand repeated punishment. Why Zhuque-3 uses stainless steel, and what the grid fins and landing legs are about, are the subjects of the next chapter.

4. Building the Rocket Like a Factory Product: The Stainless-Steel Airframe and the Recovery Trio

In December 2023, when LandSpace unveiled Zhuque-3 (ZQ-3), what most surprised the industry was neither its payload capacity nor its nine engines, but a seemingly "backward" decision: building the airframe out of stainless steel. Tencent News headlined it at the time simply as "China's first stainless-steel rocket."

In an industry that has long held "weight reduction" as its supreme article of faith, choosing stainless steel was almost heresy. But it is precisely this choice that best explains the underlying logic of Zhuque-3—it is not trying to build a piece of aerospace art, but a batch of factory products that can withstand being used over and over again.

A "Counterintuitive" Choice: Building a Rocket Out of Stainless Steel

For aerospace airframes, the traditional first choice is aluminum alloy, and even more prized is aluminum-lithium alloy—low density, high strength, able to save every gram of dead weight. The rocket engineer's instinct is to fight tooth and nail against weight. From that instinct, stainless steel is simply a disaster: its density is roughly three times that of aluminum alloy, so the same propellant tank built from steel would be far heavier.

So why did LandSpace still choose it? The answer is hidden in the word "reuse." When you plan to use a first stage dozens of times over and over, the criteria for evaluating a material change—you no longer ask only "how light is it," but rather "how cheap, how heat-resistant, and how easy to repair is it." On all three of these questions, stainless steel's answers are more attractive than aluminum alloy's. This is not an isolated case unique to LandSpace: across the ocean, SpaceX chose stainless steel for its next-generation Starship for exactly the same reasons. It should be noted that the two companies use different steels; LandSpace has officially confirmed only "high-strength stainless steel plus laser welding," and has not publicly disclosed the specific grade (a certain grade circulating in rumor comes from a single unofficial source, which this article does not credit), so one cannot directly transplant Starship's steel processes onto Zhuque-3—but on the judgment that "steel is better suited than aluminum for a reusable rocket," the two companies have arrived at the same place.

Behind this judgment is a concession from aerospace thinking to manufacturing thinking. Aluminum alloy is the "performance-first" choice; stainless steel is the "cost-and-ruggedness-first" choice. When a rocket goes from "flying once" to "flying dozens of times," the balance naturally tips from the former toward the latter.

The Stainless-Steel Ledger: Heavier, Yet Cheaper and More Heat-Resistant

The advantages of stainless steel can be broken down into three accounts.

The first is the cost account. According to LandSpace's public figures, the manufacturing cost of a large-diameter, thin-walled stainless-steel propellant tank is about 80% lower than that of aluminum alloy, and the manufacturing cycle can also be shortened by about 40%. SpaceNews cites Zhang Changwu as saying that stainless steel gives the whole rocket an 80% to 90% cost-reduction potential compared with an expendable rocket. For a reusable rocket that must win by being "cheap," an 80% cost reduction is an overwhelming temptation—it is enough to offset the slight payload loss caused by steel being heavier than aluminum.

The second is the heat-resistance account. When the first stage re-enters the atmosphere from space at high speed, its surface must endure severe aerodynamic heating; aluminum alloy begins to soften and lose strength at just a few hundred degrees, whereas the upper temperature limit of stainless steel is much higher. This means a first stage built from stainless steel relies less on thermal-protection layers during re-entry, and the structure itself better withstands repeated thermal shock—and "withstanding repeated thermal shock" is precisely a hard requirement for reuse. SpaceNews makes this point as well: heat resistance is the key to stainless steel being advantageous for multiple returns and reuse.

The third is the maintainability account. A first stage that has flown once must be inspected, refurbished, and possibly re-welded, and stainless steel is "easier to handle" than aluminum alloy in this respect—its welding properties are more forgiving, and repeated welding and repair are more relaxed. If you view a rocket as a piece of "equipment" that requires regular maintenance and repeated duty, how easy the material is to repair directly determines the cost and cycle of refurbishment.

It is worth mentioning that Zhuque-3 also made changes to the propellant-tank layout—it places the fuel tank on top and the oxidizer tank on the bottom. This new configuration is favorable for large-attitude maneuvers in the weightless coast phase and also for settling and feeding cryogenic propellant. Every choice of material and configuration answers the same question: how to make this rocket fly back, and be able to fly a second time.

Making the Steel-Skinned Cylinder Fit Perfectly: The Laser-Welding Hurdle

Choosing stainless steel is one thing; being able to weld it well is another. A rocket propellant tank is a large, thin steel-skinned cylinder that must hold hundreds of tons of cryogenic propellant at over minus one hundred degrees, and the quality of the welds directly determines success or failure—a single cold-lap weld could crack open in flight.

LandSpace's answer is laser welding. According to LandSpace's official site, the company has developed in-house a complete production line of high-performance laser-welding process equipment, dedicated to welding this kind of high-strength stainless-steel thin-walled tank. A common misunderstanding must be clarified here: many reports take for granted that large-diameter tanks are all "friction stir welded (FSW)," but Zhuque-3's public statements uniformly say laser welding; the two are different processes and must not be conflated. The advantages of laser welding are a small heat-affected zone, high weld quality, and ease of automation and mass production—yet another process choice pointing toward "batch production."

Connecting the material, the welding, and the configuration, one finds that Zhuque-3's airframe is itself a "manufacturing manifesto": it does not pursue the ultimate in any single performance metric, but pursues an entire package that can be made cheaply, stably, and batch by batch, and can also withstand repeated recovery and refurbishment. This line of thinking and its nine parallel engines and its batch-production engine factory are different facets of the same logic.

The Homecoming Trio: Grid Fins, Cold-Gas Attitude Control, and Landing Legs

A rocket meant for vertical recovery grows three things—needed only for "coming home"—that an expendable rocket does not have: grid fins, a reaction control system, and landing legs.

Grid fins are the small lattice-shaped wings on the upper part of the airframe, used specifically during re-entry and descent to control attitude and adjust the landing point. Zhuque-3's first stage is fitted with four deployable-retractable grid fins in a "P-shaped partially swept" configuration, plus two strake wings; Chinese Wikipedia records this aerodynamic layout, and there is even an academic paper published specifically studying the aerodynamic characteristics of Zhuque-3's grid fins. The material and actuation method of the grid fins have not been publicly disclosed, and this article does not speculate.

The reaction control system (RCS) is an attitude fine-tuning device used at high altitude where the atmosphere is thin and grid fins are ineffective. Zhuque-3 uses a cold-gas RCS—generating tiny thrust to adjust attitude by jetting high-pressure cold gas. The "composite control strategy of cold-gas RCS plus grid fins" that Y1 validated is what lets these two work in relay at different altitudes, always keeping the airframe's attitude pinned down. Public information mentions only the cold-gas RCS, with no mention of gas-jet attitude control, and this article states it accordingly.

Landing legs are used at the very last moment of touchdown. There is an easily confused detail worth nailing down here: the orbital-class Zhuque-3 has four landing legs, whereas the test vehicle that performed the 10-km-class VTVL flight test in 2024 had three legs—LandSpace's official records confirm this. The deployment mechanism and cushioning method of the landing legs are likewise not detailed officially. Y1 fell precisely at the doorstep of this final step—"deploying the landing legs and softly touching down"—it never got the chance in flight to validate the landing legs at all, and this is one of the lessons Y2 must make up.

Who Makes These "Recovery Parts": An Almost Blank Supply Chain

A phenomenon that deserves the special attention of researchers is this: for the three most core "recovery parts"—grid fins, landing legs, and thrust-vector actuation—there is almost no publicly verifiable external specialist supplier to be found.

The reason is that leading rocket companies generally emphasize in-house development of these recovery mechanisms. Take Deep Blue Aerospace, which follows the same vertical-recovery path: its Nebula-1 was the first to adopt an all-carbon-fiber landing cushioning mechanism developed for an orbital-class rocket, which The Paper calls China's first landing cushioning device to enter engineering application—everything from deployment and locking to shock absorption is developed in-house, with no external whole-part supplier disclosed. The specific manufacturer of Zhuque-3's landing legs is likewise not findable through public channels. As for the rumor circulating in the market that "a certain listed company exclusively supplies grid fins/landing legs," most of it comes from stock forums and financial self-media, lacks first-hand confirmation from the rocket companies, and this article does not credit any of it. A line must be drawn especially clearly here: the "net," "cable," and "hook" suppliers named on the offshore net-capture recovery path belong to an entirely different technical route, and must never be mixed into the supply chain for vertical landing legs of the Zhuque-3 type—such mix-ups have already taught hard lessons in past supply-chain reporting.

What can be documented are mostly peripheral actuation and structural parts. On the actuation-servo side, Beijing Hangyu Servo Technology masters valve-controlled hydraulic, high-power electromechanical, and other categories of servo mechanisms, with disclosed customers including Orienspace, CAS Space, and Space Transportation, among others (but LandSpace is not named); this is information at the level of the company's own account. On the structural-parts side, media report that Xi'an Jiaye Aviation, a subsidiary of Shenjian Co., supplies aft-section structural parts for Zhuque-3—engine mounts, cabin-section shells, connecting frames, and other components that bear launch thrust and recovery impact loads—and has signed a long-term framework agreement with LandSpace; Sina has related reporting, but this concerns structural parts, not the "recovery mechanism proper," and the first-hand source leans toward stock media, so it is advisable to cross-check against exchange announcements. We record this "supply-chain blank" itself as a conclusion—it both illustrates how high the technical barrier of recovery mechanisms is, and is precisely a door of industrial opportunity being pushed open, a point that Chapter 11 will return to in detail.

From "Precision Assembly" to "Rugged and Durable"

Gathering the threads of this chapter, Zhuque-3's airframe conveys a temperament very different from traditional aerospace. A traditional launch vehicle is like a precision instrument, pursuing the ultimate in lightness, in accuracy, with every part finely carved, retiring after a single flight. Zhuque-3 is more like a piece of engineering machinery—using steel that is heavier but cheaper, using laser welding that can be automated, using standardized parts that can be batch-produced, pressing costs to the lowest and pushing reliability and maintainability to the top, all for the sake of repeated duty.

It even uses a "pyrotechnic-free" scheme for stage separation—relying not on one-time explosive bolts, but on a reusable mechanism to separate, a detail again aimed at reuse. This shift from "precision assembly" to "rugged and durable" is not a technological regression, but a switch of the objective function: when the standard for measuring success or failure changes from "how well it flies" to "how cheaply it flies once, and how many times it can fly," the entire way of building the rocket must be rewritten.

But no matter how ruggedly it is built or how fully it is equipped, whether the rocket can finally stand firmly still depends on its "brain"—that guidance-and-control system which, in the final few kilometers, commands the engines, grid fins, and landing legs to work in concert. Y1 fell precisely at this final "last kick." In the next chapter, we take apart this brain that lets the rocket "fly itself back."

5. Letting It Fly Itself Back: GNC and Vertical Landing Control

The most counterintuitive thing about rocket recovery is that it is not a "hardware problem" but, first and foremost, an "algorithm problem." Grid fins, landing legs, and a deep-throttling engine are all necessary conditions; but molding them into a whole that can let the rocket fly itself back and stand firmly still depends on a "brain" called GNC—guidance, navigation and control. Y1 validated the hardware and validated the aerodynamics; the fall it took in the end was taken precisely at the last kick of this brain in the landing phase.

A Control Problem With Only One Chance

First, understand why this problem is hard. A first stage returns from over a hundred kilometers up, at several times the speed of sound, to finally stand vertically on a landing point a few tens of meters square; the whole process lasts only a few minutes, and there is no second chance—it cannot go around, circle, and try again like an airplane. Speed, attitude, position, and remaining fuel are all changing at breakneck pace, and if any single link is miscalculated or the reaction is half a beat slow, the result is a pile of wreckage.

LandSpace's official phrasing of Y1's GNC capability is quite restrained in wording yet very weighty: this was the country's first flight validation of "high-precision return navigation, guidance and control" for an orbital-class reusable rocket, breaking through onboard trajectory optimization as well as attitude stabilization and high-precision control under complex constraints, wide-range state variation, and a strongly uncertain environment, as LandSpace's official site records. Translating these mouthful terms into plain words: the rocket must, under the extreme condition of plunging at high speed while being repeatedly disturbed by the atmosphere, compute in real time "where should I fly now, how much should I open the engine, how much should the fins deflect," and it must compute this both quickly and accurately. This capability is the most essential dividing line between vertical recovery and "putting something into space"—the latter's trajectory is relatively fixed, while the former requires the rocket to continuously correct itself along the return journey.

Onboard Planning: The Rocket Computes How to Fly by Itself

Traditional rocket flight mostly follows a "preset trajectory"—the ground computes in advance how it should fly, and the rocket executes accordingly. But recovery cannot do this. Along the return journey, the high-altitude wind, the fluctuations in atmospheric density, and the airframe's actual state all deviate from the preset; if it rigidly sticks to the preset trajectory, the deviation will snowball, and the landing point will end up thousands of miles off.

Zhuque-3 uses "onboard trajectory optimization"—in plain terms, the rocket recomputes the optimal return trajectory in real time, over and over, along the flight, adjusting how it flies next based on its current position, velocity, attitude, and remaining fuel at any moment. This is like an expert driver on an icy downhill road who, rather than memorizing a fixed set of moves, corrects steering and braking at every instant according to how the wheels are slipping. Y1 validated this capability in the supersonic re-entry aerodynamic gliding phase: it returned in a lifting-body manner, relying on the coupling of aerodynamics and control to keep its attitude controllable under maximum dynamic pressure at all times, with landing-point deviation held to the meter level—this ability to "cross the atmosphere at high speed while still holding steady and computing accurately" is one of Y1's most solid gains.

Deep Throttling: Making a Single Engine "Pinch" the Speed to Zero

By the final landing phase, the problem becomes a pure mechanics puzzle: how to make a rocket plunging at high speed have its speed reach almost exactly zero at the instant of touchdown?

This relies on the engine's "deep throttling" and "precise ignition timing." So-called deep throttling means the engine can continuously adjust thrust over a very wide range—too much thrust, and the rocket gets pushed back up and bounces; too little, and it cannot brake. Zhuque-3 takes the SpaceX-style route: in the final descent segment it ignites a single engine to decelerate (colloquially known in the industry as a "hover-slam" landing, though it does not actually hover), relying on precise control of this engine's thrust and ignition timing to press the speed down to near zero the instant before touchdown. It must be specially clarified here: Zhuque-3 does not "hover"; public materials do not support the "hover" claim, and this article does not use that word; it takes the route of no margin, a one-shot precise braking stop. According to china-in-space, Y1's landing-phase engine ignition sequence was "light 1 engine first, then light 4, then switch back to 1"—using multiple engines first for coarse deceleration, then narrowing to a single engine for fine adjustment, a clear logic.

The problem arose precisely here. This landing burn failed about 17 seconds before touchdown, officially characterized as "abnormal combustion." The landing phase is the least fault-tolerant segment of the entire recovery: altitude and speed are already very low, leaving the system only a dozen-odd seconds to correct errors, and if the engine's thrust response or combustion stability goes even slightly awry, there is no time to save it. Y1 got every other link in the recovery chain to work, failing only to clear this final barrier of "the last dozen-odd seconds, with only one chance."

A Textbook-Grade Ten-Kilometer Test

To prove this GNC is not merely on paper, one must look at a beautiful test LandSpace conducted before Y1. At noon on September 11, 2024, in Jiuquan, a Zhuque-3 test vehicle named VTVL-1 flew to a peak of 10,002 meters above the ground, then returned vertically and landed successfully; both Xinhua and LandSpace's official site recorded it.

The data from this test can be called a report card of GNC capability. The test vehicle completed the country's first mid-air re-ignition at an altitude of 4.64 km and a speed of 0.8 Mach, with the engine achieving thrust modulation from 45% to 111%; under transonic high dynamic pressure, the grid fins, cold-gas attitude control, and engine jointly completed guidance and control, and it also engineered real-time high-altitude wind correction. What best illustrates the precision is the set of numbers at touchdown: the landing point was only 1.7 meters from the center of the landing pad, the vertical velocity was 1.65 meters per second, and the attitude-angle deviation was only 0.3 degrees. This test vehicle was 3.35 meters in diameter, 18.3 meters long, with a liftoff mass of about 68 tons, using an 80-ton-class Tianque-12 (TQ-12) improved engine. Using a "scaled-down" rocket, it ran through the complete chain from high-altitude re-ignition to precise vertical landing—this test proved that LandSpace's GNC passes muster at the ten-kilometer scale.

From Ten Kilometers to Orbital Class: How the Difficulty Jumps Sharply

So the question arises: since the ten-kilometer-class test was done so beautifully, why did Y1's orbital-class recovery still fail?

The answer is that from ten kilometers to orbital class, the difficulty does not increase linearly but jumps sharply. The ten-kilometer-class test vehicle flew up vertically by itself and then landed vertically, with far more moderate speed and altitude; the orbital-class first stage is accelerated to near orbital velocity, flies up to over a hundred kilometers, and then slams back into the atmosphere at several times the speed of sound—the aerodynamic heating, dynamic pressure, velocity range, and attitude change it experiences are all several orders of magnitude more violent than the test vehicle's. Correspondingly, the landing-phase engine must ignite in a far harsher incoming-flow environment, and the challenge to combustion stability is also greater. Y1 got through the most perilous segment of "high-speed re-entry crossing the atmosphere," which precisely shows that its aerodynamics and the main body of its GNC hold up; what stopped it was the reliability of the specific link of the landing burn at the orbital-class scale.

This also explains why LandSpace is being so cautious on Y2. The main framework of the GNC has already been proven feasible by Y1; what Y2 must hone is the stability of this final link, the landing phase, under the demanding conditions of the orbital class—this kind of "last mile" reliability often requires repeated honing through tests that hew ever closer to real operating conditions, and cannot be rushed. From this angle, the repeated slipping of Y2's launch window is less a matter of falling behind schedule than the reverence that the matter of "only one chance" deserves.

Up to here, the three blocks of hardware, materials, and control have all been taken apart. But behind all these technical choices actually presses a colder question: after paying such a great price to recover the first stage, is it worth it after all? Can reuse really bring launch costs down? By how much? That account is the subject of the next chapter.

6. Reuse Economics: The Ledger Where the First Stage ≈ 70% of the Whole Rocket

The previous five chapters all discussed "how to recover the first stage"—from the engine and materials to control. But a colder question has hung overhead all along: after all this effort to recover it, is it worth it? If recovering and refurbishing it costs more than building a new one, then all the grid fins, landing legs, and deep throttling become expensive performance art.

To answer this question, one must first see clearly where a rocket's money actually goes.

A Ledger of Seventy Percent

A rocket's cost structure is rather different from most people's intuition. Fuel is worth almost nothing—the liquid-oxygen/methane (methalox) propellant for a single Zhuque-3 launch costs only a fraction of the whole rocket. What is truly expensive is that airframe that "burns up after one use," and within the airframe the most expensive part is the first stage: it is packed with nine engines, the largest propellant tanks, and the most complex structure. Cailian Press cites industry figures noting that the first-stage body itself accounts for over 70% of the whole rocket's manufacturing cost; PEdaily also has the statement that "hardware such as engines and propellant tanks accounts for over 70% of total cost."

This "seventy percent" is the key to understanding the entire reuse logic. If every launch burns up a first stage that accounts for 70% of the cost, it is like throwing away the car every time you drive to your destination and buying a new one for the next trip—no amount of wealth can withstand such waste. What reuse aims to do is to drive this "car" back, fill it up, give it an inspection, and then drive a second trip, a third trip. In theory, as long as the first stage can be used over and over, the airframe cost amortized per launch will keep dropping with the number of reuses. Cailian Press gives the empirical figure that advantages show within 5 reuses, and used more than 10 times, the cost per launch can drop by about 80%.

Reuse Is No "Free Lunch": Refurbishment Costs Money Too

However, reuse is not simply a matter of picking the first stage back up and being done. After it lands, it must be inspected, refurbished, and possibly have parts replaced—all of which cost money and time. The economics of a reusable rocket is essentially a subtraction problem: the manufacturing cost of that one saved first stage, minus the cost of recovery and refurbishment, minus the revenue corresponding to the payload capacity lost from carrying the recovery hardware (landing legs, grid fins, extra reserved fuel)—what remains is the net gain.

This is why seemingly dull material details like "methane doesn't coke, stainless steel welds well" get written directly into the economic account: what they determine is precisely "how much refurbishment costs." An engine that burns kerosene and whose inner walls are caked with coke is slow and expensive to refurbish; an engine that burns methane and whose inner walls are clean can fly again after a wipe. Every "manufacturing choice" Zhuque-3 makes in materials and fuel ultimately cashes out as real money in the refurbishment-cost column. Only by understanding this can one grasp why LandSpace would rather carry the payload burden of "stainless steel being heavier" and still choose it—it is betting on the total ledger over the full life cycle, not on the performance of a single launch.

Falcon 9 Lays the Ledger Open for the Whole World to See

As for how much reuse can actually save, the most persuasive sample is not in China but in SpaceX's Falcon 9—it has, over ten years, laid this ledger open for the whole world to see.

First, unit cost. The price to send one kilogram of payload to low Earth orbit was as high as about USD 54,500 in the Space Shuttle era, while Falcon 9, through reuse, has pressed it down to around USD 1,500—a reduction of over 90% (such unit costs are mostly aggregated figures, cited only as an order-of-magnitude reference). Next, the declining curve of refurbishment cost: according to industry-institution estimates, the refurbishment cost of a Falcon 9 first stage fell over five years from about USD 13 million all the way to about USD 1 million, and the turnaround time was also compressed to about three weeks, with a single refurbishment taking as little as nine days at the fastest. Behind these numbers is a clear learning curve—reuse is not a one-time saving, but a continuous decline in cost as the number of uses grows and the process becomes well-honed.

The latest milestone occurred on July 9, 2026: a Falcon 9 first stage numbered B1067 completed its 36th flight, becoming the rocket first stage with the most re-flights in human history; the related report recorded the moment. As of mid-July 2026, SpaceX had cumulatively and successfully recovered first stages nearly six hundred times. What does one airframe flying 36 times mean? It means amortizing the manufacturing cost of a first stage over 36 missions, with the airframe cost apportioned to each launch being only a few tens of times less than building a new rocket. This is the far shore that Zhuque-3, and all Chinese reusable rockets, want to reach.

Zhuque-3's Cost-Reduction Target: From Fifty Thousand to Twenty Thousand

So what kind of ledger does Zhuque-3 itself compute?

According to LandSpace's public and prospectus figures, Zhuque-3 (enhanced) has three tiers of payload capacity: about 21.3 tons to low Earth orbit when fully expendable, about 18.3 tons when the first stage is recovered at sea within the flight corridor, and about 12.5 tons when the first stage returns to the launch site for recovery. Its first stage is designed for no fewer than 20 reuses; 5 reuses can reduce cost by about 45%, and by 20 uses the incremental cost is already near the margin—that is, the more it flies, the cheaper it gets, giving it an 80% to 90% cost-reduction potential overall compared with an expendable rocket.

Coming down to the most intuitive price, LandSpace's ultimate target for Zhuque-3 is to press the cost of reaching orbit per kilogram down to RMB 20,000. As a reference, current Chinese commercial launch prices are roughly RMB 50,000 to 100,000 per kilogram, and higher for special orbits—bringing it down to RMB 20,000 is almost halving and halving again. Whether this target can be achieved depends on whether the number of reuses can really be stacked up and whether refurbishment cost can really be pressed down, and the premise of all this is that Y2 first turns "being able to recover" from a design into reality. Cost reduction is not shouted out; it is flown out through one successful recovery after another, one re-flight after another.

The "Payload Tax" That Recovery Must Pay

Reuse saves money, but it is not without cost. The most direct cost is payload capacity—to recover the first stage, the rocket must leave behind a portion of fuel that could have been used to deliver payload in order to decelerate, and must also carry the dead weight of "homecoming-only" items like landing legs and grid fins up into space. This loss can be called the "payload tax."

How heavy this tax is depends on which recovery method is used. According to figures from JZYJ in an interview, when the first stage returns to the launch site for vertical recovery, because it must fly back and deceleration consumes the most fuel, the payload loss is about 40%; if instead it is recovered on an offshore platform along the trajectory without returning to the site, the loss can drop to about 23%; East Money recorded this comparison—which is also why over 80% of Falcon 9 missions choose offshore drone-ship recovery. Zhuque-3's own three payload tiers also bear out this account: its 12.5 tons for return-to-site recovery, compared with 21.3 tons when expendable, is a payload loss of about 41%, which fits exactly with the figure of "about 40% loss for land vertical recovery."

Once this "payload tax" is understood, one can understand why there will be a contrast of "two Chinese routes" in Chapter 7—one important motive for the national team's Long March 10B (CZ-10B) choosing offshore net-capture recovery is precisely to "eliminate the landing legs and save the dead weight," pressing this payload tax as low as possible. Every choice of recovery method is essentially about finding a balance point between "saving cost" and "preserving payload."

Why Private Companies Cannot Afford to Lose This Cost-Reduction Battle

Finally, one must ask: why is it private companies, rather than the deep-pocketed national team, that charge hardest down the most difficult path of vertical recovery?

The answer is hidden in the business model. The national team shoulders national missions, and costs can be coordinated by the state; but private rocket companies, to survive in the market, have only one way out—to drive launch prices down and push launch frequency up, making money by "hauling more and running faster." Under this logic, reuse is not a bonus but a lifeline—if a first stage accounting for 70% of the cost is discarded every time, a private company simply cannot bring prices down to a competitive level. At the same time, as Chapter 9 will discuss, the enormous launch demand from China's LEO constellation deployment is rushing in; whoever can use reusable rockets to provide both cheap and high-frequency payload capacity will take this slice of the cake; conversely, whoever is slow to achieve reuse will be squeezed out by the double pressure of cost and frequency.

So for a company like LandSpace, that "catch itself" moment on Y2 has never been merely a technical validation, but a hard-fought battle over whether the company can close its commercial loop. The ledger of reuse, computed to the end, is a survival problem. And the other side of this survival problem—why it is precisely now, and just how large the payload-capacity gap on the demand side is that forces reuse—we leave to Chapter 9 to calculate in detail; the next chapter first returns to that suspense: facing the same reuse problem, China's national team gives an answer utterly different from Zhuque-3's.

7. Two Chinese Approaches: Net-Capture Recovery vs. Vertical Recovery, Side by Side

In the summer of 2026, China's space sector did something intriguing: in less than a month, using two completely different methods, it each pushed the hard problem of "recovering a rocket's first stage" forward by a large step. On July 10, the state team's Long March 10B made its maiden flight in Hainan, catching the first stage with a net at sea; meanwhile at Jiuquan, the private-sector Zhuque-3 Y2 had just completed a static-fire test and was preparing to have the first stage stand on its own via vertical retro-propulsion. The same goal, two diametrically opposed roads—this is not a coincidence, but a key to understanding the landscape of China's reusable rockets.

The Same Summer, Two Ways to "Catch"

Let us first place the two events side by side. At 12:15 on July 10, 2026, the Long March 10B made its maiden flight from the Hainan Commercial Spacecraft Launch Site; as the satellite entered orbit, the first stage was steadily caught by a recovery ship. Xinhua called this China's first controlled recovery of a launch vehicle's first stage, and the world's first net-capture recovery—achieved successfully on the maiden flight, no less. The rocket was led overall by the China Academy of Launch Vehicle Technology (CALT), with a 5-meter diameter and a two-stage configuration: the first stage burns liquid-oxygen/kerosene (kerolox) (inherited from the Long March 10A's first stage), the second stage burns liquid-oxygen/methane (methalox), with a liftoff thrust of about 890 tonnes, a liftoff mass of about 760 tonnes, and an overall length of about 63 meters; in reusable configuration its low-Earth-orbit (LEO) payload capacity is about 16 tonnes.

By contrast, Zhuque-3 Y1 had, half a year earlier, also carried out China's first technology demonstration of orbital-class first-stage recovery—only the final step failed; Y2 is now going for "the first successful recovery by a Chinese private company." One of these rockets comes from the state team, the other from the private sector; one nailed the recovery that same month, the other is still sprinting toward it. But their real dividing line is not "who succeeds first," but rather that they chose completely different recovery mechanisms.

Net-Capture Recovery: Offloading the Hard Part from the Rocket to the Ship

The Long March 10B's logic can be summed up in one sentence: since making the rocket stand on its own is too hard, why not simply not make it stand on its own, and instead have the ground catch it.

Its first stage has no landing legs; in their place are several lightweight hooks. During recovery, a dedicated recovery ship sails to the sea area of the first stage's designated landing point, and atop a mast erected on the ship a "well"-shaped (井) cable net is spread open; when the first stage returns, the hooks on the rocket body catch the cable net, the net mouth cinches closed, and the first stage is captured and secured. Media such as Sina described this mechanism. The most ingenious aspect of this approach is that it completely bypasses the hardest maneuver of all—"vertical precision landing." The rocket does not need to precisely pinch its speed down to zero at the instant before touchdown, does not need to deploy landing legs, and does not need that extremely low-margin landing burn; it only needs to fly roughly above the net and let the hooks catch on. The requirement for landing-point accuracy is thereby loosened considerably.

In other words, net-capture recovery "offloads" the complexity of recovery from the rocket onto the ship and the net. For China this is a move that plays to its strengths and avoids its weaknesses—China has the world's strongest marine engineering equipment and shipbuilding capability, and building a large ship that can steadily spread a net at sea to catch a rocket is precisely a strength of Chinese manufacturing.

The Lingxing (Pilot) Vessel: A Ship Born to Catch Rockets

The core equipment of this approach is a recovery ship named "Lingxing" (Pilot). Delivered on November 30, 2025, it is about 144 meters long and about 50 meters wide, with a full-load displacement of about 25,000 tonnes, and is fitted with a hydraulic damping system to counteract the sway of the waves so that the net stays as stable as possible on the undulating sea surface. A 25,000-tonne ship, built specially to "catch a rocket returning from space"—this image itself illustrates the logic of net-capture recovery: using the "heaviness" of marine engineering equipment to resolve the "difficulty" of rocket landing.

This also incidentally explains the cost of this approach. It requires a costly dedicated recovery ship, a window in which sea conditions permit operations, and the rocket's landing point to fall roughly within a sea area the ship can reach. In exchange, it saves the dead weight of landing legs, as well as the extremely difficult control of vertical precision landing—trading "looseness" in payload capacity and landing accuracy for "dependence" on the ship and sea conditions. According to public figures, the Long March 10B's LEO payload capacity in reusable configuration is about 16 tonnes, which also confirms how removing the landing legs and simplifying the onboard structure helps preserve payload capacity.

Vertical Recovery: Pushing the Hard Part Back onto the Rocket Itself

Zhuque-3 takes the exact opposite road: without relying on any external facility, it has the rocket fly itself back and stand upright.

On this road, all the complexity is pushed onto the rocket itself. It must carry landing legs, grid fins, and a reaction control system (RCS) into the air; it must reserve fuel for the return deceleration (the "payload tax" discussed in Chapter 6); it must have engines capable of deep throttling and of precise ignition during high-speed descent; and it must have that GNC "brain" compute speed, attitude, and landing point all accurately within the final dozen or so seconds—Y1 fell precisely at this last step. Its advantages are equally clear: it needs no dedicated recovery ship and does not depend on sea conditions, so in theory it offers the fastest turnaround and the strongest scalability; moreover it is benchmarked against SpaceX's Falcon 9—a mature paradigm validated over nearly a decade and flown nearly six hundred times—so the path is clear and the ceiling is known.

Zhuque-3 (enhanced variant), in downrange-recovery configuration, has a LEO payload capacity of about 18.3 tonnes, a bit higher even than the Long March 10B's 16 tonnes—this does not mean vertical recovery is more payload-efficient; the two differ in configuration, scale, and recovery method, and cannot be simply compared by size. What it shows is that both roads can, under the premise of "sacrificing part of the payload for recovery," keep payload capacity at a usable level.

The Trade-offs Behind a Comparison Table

Placing the key differences of the two approaches side by side makes the trade-offs clearer:

Dimension	Long March 10B (net-capture recovery)	Zhuque-3 (vertical recovery)
Affiliation	State team (CASC / CALT)	Private sector (LandSpace)
Recovery mechanism	First-stage hooks; sea-based recovery ship spreads a net to capture	First stage autonomously retro-propels and lands vertically to stand
Where the complexity lands	Ground/ship (marine engineering equipment)	The rocket itself (control + structure)
Landing legs	Removed, replaced by hooks	Four landing legs retained
Landing-accuracy requirement	Relatively loose	Extremely high (meter-level)
Main dependencies	Dedicated recovery ship, sea-condition window	Its own GNC, engine deep throttling
Benchmark	World-first, no foreign precedent	SpaceX Falcon 9 mature paradigm
LEO payload in reusable configuration	About 16 tonnes	About 18.3 tonnes (downrange recovery)

The most thought-provoking row in this table is "Benchmark." The Long March 10B's net-capture recovery is a world first, with no foreign precedent to follow—an original creation of China's space sector; Zhuque-3's vertical recovery, by contrast, follows the mature road that SpaceX blazed, winning on a clear path. One pursues "novelty," the other "steadiness"—each has its own rationale.

Why China Bets on Both Roads at Once

A natural question arises: since SpaceX has already proven vertical recovery viable, why does China not concentrate its efforts on that one road, and instead divert energy into net-capture recovery, which no one has ever done?

Our judgment is that this is precisely a smart combination bet. Each of these two roads plays to a different strength of China's. Net-capture recovery plays to China's marine engineering and shipbuilding strengths—transferring the difficulty of recovery onto the large ships, large structures, and sea-based operations that China is best at; vertical recovery plays to China's efforts to catch up to the world frontier in engines, control, and precision manufacturing—it forces China's private space sector to gnaw through the hardest bone of all. The state team, using an original net-capture scheme, provides a self-controllable recovery path for high-value, high-reliability national missions; the private echelon, using the vertical scheme, competes for the low-cost, high-frequency commercial launch market. The two roads are not in a competitive relationship but each occupy an ecological niche, together propelling China into the first tier of "orbital-class controlled recovery."

Of course, on the vertical-recovery road, LandSpace is not alone. Behind it stands an entire echelon of Chinese private rockets—i-Space, Space Pioneer, Galactic Energy, Deep Blue Aerospace, JZYJ… Some follow close behind, some blaze their own trail, some have just stumbled and are now troubleshooting back to zero. Just how far this echelon has run is what the next chapter will take stock of.

8. The Private Echelon: A Genealogy of Recovery Progress Beyond LandSpace

Zhuque-3 is running out in front, but it is not running alone. Behind it, China's private rockets have already lined up into a visible echelon—some biting close and not letting go, some blazing their own trail, some having just stumbled and now climbing back up. Taking stock of this echelon's progress one by one makes one thing clear: 2026 really is a year in which China's private reusable rockets converge on the finish line all at once, but beneath the excitement, differentiation has already begun.

It should first be noted that these companies' maiden-flight dates and test progress change extremely fast; all timing figures in this chapter are as of July 2026, and strictly distinguish three states—"already succeeded," "sprinting," and "already failed, awaiting troubleshooting to zero"—without treating plans as facts.

The Close Pursuers: i-Space and Space Pioneer

Among the closest pursuers of Zhuque-3, i-Space and Space Pioneer are two names impossible to skip over, but their situations are very different.

i-Space's Hyperbola-3 takes the same road as Zhuque-3—methalox, vertical recovery. This rocket has a diameter of 4.2 meters, a LEO payload capacity of about 14 tonnes in expendable configuration and about 8.5 tonnes in recovery configuration; its goal is to achieve "orbital insertion plus sea recovery" all at once on the maiden flight, with the maiden flight expected around the end of 2026 according to Sina. Its confidence comes from its self-developed Jiaodian (Focus) series of methalox engines, of which the Jiaodian-2 has already entered batch production. i-Space was one of the earlier companies in China's private space sector to achieve orbital insertion with a solid rocket, and now, on the road of liquid reusability, it is one of Zhuque-3's most direct competitors.

Space Pioneer's situation, by contrast, has been far rockier. Its Tianlong-3 is benchmarked against the Falcon 9, taking a partially reusable kerolox road, with a LEO payload capacity exceeding 20 tonnes, and was once held in high hopes. But its road has been full of bumps: on June 30, 2024, when the Tianlong-3 first stage was performing a nine-engine parallel static-fire test, the rocket body unexpectedly detached from the test stand, rose, and then crashed; Guancha recorded this accident, caused by a weak aft-structure design. Afterward, in September 2025, it completed a 30-second static-fire test on a sea platform, clawing back a round; but on April 3, 2026, the Tianlong-3 Y1 maiden flight failed and did not reach orbit, with Caixin reporting that it is undergoing troubleshooting to zero. Space Pioneer's story reminds people that on the reusability road, taking a fall is the norm; whether one can get back up is what matters.

Steady and Solid: Galactic Energy and Orienspace

If i-Space and Space Pioneer are "racing the clock," Galactic Energy and Orienspace are more "steady and solid."

Galactic Energy's Pallas-1 takes the kerolox road, with the first stage clustering seven self-developed Cangqiong-50 (CQ-50) engines, a liftoff mass of about 283 tonnes, a LEO payload capacity of about 7 tonnes, and a design reuse of no fewer than 25 times. Its play is "first make orbital insertion solid, then add recovery on top": according to Galactic Energy's official website, the CQ-50 engine has accumulated over 10,000 seconds of test firing, with a thrust-modulation range of 32% to 105%; in November 2025, the Pallas-1 first stage completed a sea-based power test firing, and the Y1 vehicle rolled off the line. The company has not announced an exact maiden-flight date, and plans to carry out orbital-class recovery demonstrations step by step as missions progress—rather than chasing the aggressive goal of recovery on the maiden flight, it advances one solid step at a time.

Orienspace's Gravity-2, by contrast, aims for "big." This is a medium-to-large kerolox recoverable rocket, with the first stage clustering nine already-batch-produced Yuanli-110 engines, an overall height of about 70 meters, a liftoff mass of about 715 tonnes, and a LEO payload capacity of no less than 21.5 tonnes, taking the core-stage-recovery road. According to reports such as CLS, Gravity-2 is expected to make its maiden flight between the third quarter and early fourth quarter of 2026. If successful, it will be one of the highest-payload recoverable rockets in China's private sector. What Galactic Energy and Orienspace have in common is that both take "first ensure orbital insertion, then steadily add recovery" as their rhythm, forming an interesting contrast with Zhuque-3's play of "gnawing through the hard bone of recovery first."

The Radicals: Deep Blue Aerospace and JZYuan

The echelon also has two "radicals," whose shared label is "orbital insertion plus recovery on the maiden flight"—all in one step, leaving themselves no way out.

Deep Blue Aerospace's Nebula-1 has a tandem two-stage configuration, using self-developed Leiting-R1 (Thunder-R1) kerolox engines, with an overall height of over 30 meters and a LEO payload capacity of about 2 tonnes—modest in scale but high in ambition. In September 2025 it completed a high-altitude vertical-recovery flight test of the first stage, achieving 10 of 11 objectives according to Taibo, and plans to achieve orbital insertion plus recovery on the maiden flight in the second quarter of 2026. Deep Blue also has a technical highlight worth noting: it self-developed an all-carbon-fiber landing-cushioning mechanism for Nebula-1, which The Paper called China's first landing-cushioning device to enter engineering application.

JZYuan, by contrast, is more like Zhuque-3's "shadow"—its Yuanxingzhe-1 (Element Traveler-1) is another domestic rocket in the "stainless steel plus methalox plus sea recovery" mold, with a liftoff mass of about 575 tonnes, a payload capacity of about 7 tonnes in recovery configuration, and a design reuse of over 20 times, likewise setting its ultimate goal at 20,000 yuan per kilogram. According to Sina, it completed a 2.5-kilometer-altitude flight recovery with a demonstrator vehicle back in 2024, and did another flight-recovery test in Yizhuang, Beijing in June 2025, aiming for a maiden flight of orbital insertion plus sea recovery by the end of 2026. Interestingly, JZYuan has fused two roads—"stainless steel + methane" and "sea recovery"—together: it learns Zhuque-3's material and fuel while also learning the Long March 10B's sea recovery, taking a hybrid road.

The Odd One Already in the Air: CAS Space's Lijian-2

Amid a din of "planned maiden flights," CAS Space's Lijian-2 is an "odd one" that has already flown. On March 30, 2026, Lijian-2 made a successful maiden flight, sending the Qingzhou cargo spacecraft's prototype into the air, as reported by the Chinese Academy of Sciences. This is its first liquid rocket to adopt a common-bulkhead-core (CBC) configuration, with a core-stage diameter of 3.35 meters, an overall length of 53 meters, a liftoff mass of 625 tonnes, and a LEO payload capacity of about 12 tonnes.

But it must be given an accurate positioning: Lijian-2 is currently an expendable rocket, not yet recoverable. CAS Space's reusable version will not arrive until 2028 at the earliest, and within the year it plans to first use a test vehicle called Lihong-2 to conduct a hundred-kilometer-class recovery test. So the significance of Lijian-2 is that it proves CAS Space possesses the orbital-insertion capability of a large liquid rocket, paving the way for subsequent reusable models—it stands on the step of "able to reach orbit," but has not yet climbed onto the step of "able to recover." Listing it here is precisely to remind us that orbital insertion and recovery are two different thresholds; clearing the former does not mean clearing the latter.

The Shovel Sellers: Jiuzhou Yunjian and the Engine Specialists

In a gold rush, besides the gold miners, there are also the shovel sellers. In the private-rocket industry chain, there is a class of companies that do not build whole rockets themselves but specialize in engines—they are the "shovel sellers" of this track.

Jiuzhou Yunjian (JZYJ) is a representative among them. It specializes in methalox engines, and its Longyun engine features multiple restarts, recoverability, and high thrust; according to related reports, its cumulative hot testing has exceeded 20,000 seconds with 19 units delivered, and it was once used on the Long March 12A; in April 2026 the company relocated to Minhang, Shanghai. The engine is the highest-technical-threshold segment of a reusable rocket, and also the one most likely to form a specialized division of labor—not every whole-rocket company must self-develop its engine from scratch; procuring from a specialist like JZYJ can equally get a rocket built. This "whole-rocket company + engine specialist" division of labor is itself a sign of an industry maturing: when specialized component suppliers begin to appear in an industry, it shows that the industry is moving from a startup phase of "every firm does everything itself" toward an industrialized phase of "socialized division of labor and collaboration."

The Echelon's Real Position: Differentiation Behind the Excitement

Laying out this entire echelon, a sober judgment surfaces: it is true that China's private reusable rockets are "clustering at the finish line" in 2026, but the only one that has actually run to the very brink of "orbital-class recovery" is, at present, Zhuque-3 alone.

The positions of the rest can be ranked as follows: i-Space, Orienspace, and Galactic Energy are targeting maiden flights within the year or a bit later, with recovery to follow; Deep Blue Aerospace and JZYuan have proclaimed the aggressive goal of "orbital insertion plus recovery on the maiden flight," but have not yet delivered; Space Pioneer's Tianlong-3 just failed its maiden flight and is troubleshooting to zero; CAS Space's Lijian-2 has already reached orbit but is not yet recoverable. The excitement is real, and so is the differentiation—some are at the door, some are halfway along, some have just fallen. This precisely shows how high the reusability threshold is: shouting slogans is easy, but bringing a first stage back intact and sending it up to fly a second time is the real work.

This also makes that single flight of Y2 weigh even heavier—it concerns not just LandSpace's own commercial loop; in a sense, it is a "trailblazing" for the entire echelon of China's private reusable rockets. And the fundamental force propelling this whole echelon forward lies not in the technology itself but in an exploding demand side: China's LEO constellation deployment is creating an unprecedented payload gap. How big that gap is, and how it is forcing everyone to pursue reuse, is the theme of the next chapter.

9. The Demand Side: The Constellation Deployment Wave and the Payload Gap

Up to this point, this report has been discussing the "supply side"—how well Zhuque-3 can be built, whether it can be recovered, how much money reuse can save. But no matter how advanced a rocket is, it only has meaning if there is cargo to haul. To understand why China's entire space community collectively rushed toward reusability around 2026, one must turn the lens toward the "demand side": an unprecedented satellite constellation wave is brewing in the sky, and the payload demand it creates is so large that existing rockets simply cannot haul it. The reason reusability has gone from a "bonus point" to a "lifeline" has its most fundamental impetus here.

A Sky Being Filled with Satellites

Let us first establish an intuitive sense of scale. All the satellites launched by humanity over the past sixty-odd years add up to only a bit over ten thousand; and now, just a few LEO constellations that have not even been completed plan satellite counts in the tens of thousands, or even hundreds of thousands. The sky is being "staked out" anew.

The frontrunner in this land grab is America's Starlink. According to public data, as of early July 2026, Starlink had cumulatively launched about 12,414 satellites, of which about 10,722 are operating normally in orbit; in the first half of 2026 alone it deployed about 1,589. By way of comparison, the United States completed about 122 launches in all of 2025, adding about 3,190 satellites—behind this launch density is precisely the Falcon 9's reusability providing support. Starlink has, in a decade, proven one thing: LEO mega-constellations are a viable commercial model, but its precondition is having a rocket fleet that can launch at high frequency and low cost.

China will not hand this sky over to others. LEO orbits and frequency resources are finite, and follow the international rule of "first come, first served"—whoever launches satellites first and occupies the orbits and frequency bands holds priority. This strategic implication makes China's constellation deployment no longer merely a commercial act, but one carrying the urgency of seizing strategic resources. And the only way to seize them is to send satellites into the sky as fast and as many as possible—which transmits the pressure directly onto the rockets.

Three Big Nets: Guowang, Qianfan, and More

China's LEO constellations currently on the table are chiefly three big nets, each more staggering in scale than the last.

The first is "Guowang," that is, the GW constellation led by China SatNet. According to the public scale figures, it plans a total of 12,992 satellites, to be fully deployed by 2035. This is a net led by the state team, carrying the heaviest strategic weight.

The second is "Qianfan," also called the G60 constellation, led by Shanghai. It plans a satellite count of over 15,000 (a figure of 14,000 also coexists), with a deployment cadence going from 648 by the end of 2025, to about 1,296 by 2027, and then to 15,000 by the end of 2030; Shanghai has invested about 6.7 billion yuan for this (this financing figure serves only as industry-scale background). Qianfan is, of the three nets, currently the fastest-deploying and the first to form scale.

The third is "Honghu-3," planned at about 10,000 satellites, which submitted its filing to the International Telecommunication Union (ITU) in May 2024. In addition, at the end of 2025 China newly filed LEO constellations totaling over 200,000 satellites—among them two constellations named CTC each filed for about 96,700. Here we must tap the brakes and make an honest distinction: filing is not deployment. Filing with the ITU is the first step to seizing resources, but between filing and actually launching tens of thousands of satellites into the sky lies an enormous chasm of capital, technology, and payload capacity. This 200,000-satellite filing scale—the media themselves are already questioning its feasibility. So this report, when citing constellation figures, uniformly distinguishes three tiers—"official scale, ITU filing, media conjecture"—and does not treat paper filing numbers as facts about to launch.

The ITU Clock: Why It Has to Be Now

So why does this constellation wave erupt in concentration precisely around 2025 and 2026? The answer is hidden in the ITU's stringent "timetable."

Under ITU rules, after a constellation has filed for its orbits and frequencies, it cannot occupy them indefinitely without use, but must actually launch satellites within prescribed deadlines to "prove" that it is using these resources—specifically, it must launch the first satellite within 7 years, and thereafter deploy 10% within two years, 50% within five years, and complete full deployment within seven years. Failing to meet these milestones, the occupied resources may be voided and taken by someone else.

This "milestone assessment" system amounts to setting a countdown alarm clock on every constellation. Take Guowang as an example: some research reports estimate that, in order to hold onto its frequency resources, it needs to deploy about 1,300 satellites before September 2029. Stacking up the time nodes of the several big nets forms a launch demand over the coming years that is frighteningly dense—this is not a commercial preference to "launch a bit faster," but a hard constraint of "must launch enough by a certain date." Once the ITU clock rings, everyone has to run. It is precisely because of this alarm clock that China's space sector cannot afford a slow, expensive, and sparse expendable rocket fleet—what it needs is the payload capacity to send tens of thousands of satellites into the sky within a few years, and only reusable rockets can provide that.

The Failed "Ten Satellites on One Rocket" Tender: The First Alarm of the Payload Gap

With demand this fierce, can existing payload capacity keep up? A real-world signal gives a negative answer.

According to public reports, Yuanxin Satellite, responsible for launching the Qianfan constellation's satellites, once issued a "ten satellites on one rocket" launch-service tender, which failed because fewer than three qualifying suppliers could be found. This is a highly signal-laden event: it shows that the qualified payload capacity in the market able to take on such high-frequency, batch launch tasks is so scarce that a normal tender cannot even muster enough bidders. The payload gap is no longer a paper calculation, but is really and truly choking constellation deployment.

The reason this alarm matters is that it turns "supply-demand imbalance" from an abstract industry judgment into a concrete commercial fact. Satellites are already queued up waiting to go to the sky, but there are not enough rockets to haul them up—this is precisely the market space left for the likes of Zhuque-3. Whoever can first provide cheap, high-frequency, and reliable payload capacity can take on these endless orders. Conversely, this also explains why capital and entrepreneurs keep pouring into the rocket track wave after wave: because what they see is an enormous market where demand is already clearly laid out, waiting only for payload capacity to be filled in.

Tallying the Overall Account: The Chasm Between Demand and Capacity

Putting demand and supply on the same account makes the chasm clearer.

On the demand side, just the peak deployment of a few major constellations over the coming years requires launching satellites by the thousands per year. Some experts give the figure that, to meet constellation-deployment needs, China needs to reach a level of no fewer than 100 liquid-rocket launches per year and no fewer than 2,000 satellites launched per year. On the supply side, China completed over 90 launches in 2025, of which commercial launches were about 50, accounting for about 54%; the launch target for 2026 is raised to about 140. This growth rate is already fast, but relative to the appetite of constellation deployment, it still falls short—some planning figures mention that supporting the deployment would require raising the annual launch capacity to about 610 launches, of which reusable liquid rockets would need to shoulder about 208.

Placing these numbers side by side, the conclusion is clear: the existing launch system, dominated by expendable rockets, cannot sustain this constellation wave, whether in terms of launch frequency or cost. Expendable rockets—build one, fly one—have capacity limited by the assembly line; their price cannot come down either, and satellite companies cannot afford such expensive payload capacity to launch tens of thousands of satellites. Between demand and supply lies a chasm that expendable rockets can never fill.

Reuse Is the Only Solution

And so all the clues finally converge on the same answer: reusability.

This answer was forced out, not chosen. An expendable rocket has capacity limited by "how fast it can be built"; a reusable rocket, with the first stage flying back and refueling to go up again, is equivalent to multiplying launch frequency with the same set of hardware—this is the only realistic way to fill the frequency chasm. Likewise, only by repeatedly reusing the first stage that accounts for 70% of the cost can the launch price be pushed down to a level that satellite companies can afford and that makes the constellation math work out—this is the only realistic way to fill the cost chasm. To both hurdles of frequency and cost, reusability gives the answer at once.

So when we look back again at that single act of Y2 "catching itself," its significance is not merely a technical show-off, nor even merely LandSpace's own commercial loop—it is blazing, on behalf of the entire ambition of China's LEO constellations, the road to "cheap and high-frequency" payload capacity. The floodgate on the demand side is already open, waiting only for the supply side to solve the puzzle of reusability. Of course, China is not the only country solving this puzzle. Looking across the globe—from SpaceX to Blue Origin, from Europe to Japan and India—a race over reusability is under way at the same time. Where China's private echelon stands in this race is what the next chapter will compare.

10. International Benchmarking: Beyond SpaceX, How the World Is Pursuing Recovery

Placing Zhuque-3 (ZQ-3) back into the global frame of reference is at once encouraging and cause for sober reflection. Encouraging, because private Chinese aerospace has already squeezed into the first tier of "orbital-class controlled recovery," a tier that has very few players worldwide; sobering, because the player standing at the very front of that tier, SpaceX, has pulled far ahead of everyone else. To see Zhuque-3's true position clearly, we first have to lay out this global map.

A Tiered Race

The global race in reusable rockets is not a neat, single starting line but a clearly stratified set of tiers. The first tier is SpaceX, which has already turned reusability into a daily operation and flown nearly six hundred times; the second tier consists of chasers who have only just touched the threshold of "orbit plus recovery," and at present only Blue Origin barely reaches it; the third tier consists of newcomers still doing technology validation at low altitude and suborbit, with Europe, Japan, and India all among them. China's private tier sits roughly between the second and third tiers—it has already achieved flight validation of orbital-class recovery, but has not yet made it a stable reality. Seeing this stratification clearly means one will neither be blindly optimistic just because one has "entered the first tier," nor overly self-deprecating just because one is "still far from SpaceX."

The Runaway Leader: SpaceX and Its Two Rockets

SpaceX's lead is a runaway one. It holds two cards: the already-deified Falcon 9, and the still-in-testing Starship.

Falcon 9 is a living textbook of reusability. On July 9, 2026, the first stage numbered B1067 completed its 36th flight, setting a new record for human rocket reuse; related reporting documented this milestone. As of mid-July 2026, SpaceX had cumulatively and successfully recovered first stages nearly six hundred times, and in August 2025 it also achieved its 400th autonomous drone ship landing. Its recovery can either fly back to a ground landing zone or touch down on an offshore autonomous drone ship, and it has three drone ships in active service. One rocket flying 36 times, cumulative recoveries approaching six hundred—these numbers piled together represent a level of maturity that the Zhuque-3s of the world look up to: at SpaceX, reusability has long since ceased to be a question of "can it be done," and has become a question of "how many times has it been repeated, quickly and cheaply."

Starship, meanwhile, represents a more radical next generation. According to public information, as of May 2026, Starship had accumulated 12 test flights, with seven succeeding and five failing; the 12th flight on May 22, 2026 used the V3 configuration and Raptor 3 engines for the first time, but the Super Heavy booster crashed abnormally into the sea during its return; the 13th flight had completed a static-fire test on July 10, 2026, with launch as early as July 15. Starship's fortunes swing dramatically, but it had previously used Super Heavy boosters in the V2 configuration to successfully complete "chopstick catch"-style tower recoveries multiple times, and reused two of them—this recovery method, in which the launch tower's mechanical arms directly catch the booster in mid-air, is more radical than Zhuque-3's landing-leg approach. Starship reminds us that the technical ceiling of reusability is still far from reached.

The Chaser: Blue Origin's Highs and Setbacks

Behind SpaceX, the only one to barely reach the threshold of "orbit plus recovery" is Jeff Bezos's Blue Origin, but its 2026 has been turbulent.

Its New Glenn rocket also takes the vertical-recovery route. Its maiden flight in January 2025 (NG-1) reached orbit successfully, but the booster landing failed; in the NG-2 mission on November 13, 2025, the booster successfully landed on an offshore recovery ship for the first time, making Blue Origin the world's second company after SpaceX to achieve "orbit plus first-stage recovery"; NG-3 on April 19, 2026 went a step further, reusing a recovered booster for the first time, as documented in related reporting. But the high point was followed by a heavy blow: on May 28, 2026, New Glenn exploded during a static-fire test, destroying the airframe and also severely damaging its sole launch pad, halting the program, which is striving to fly again only by year's end. Blue Origin's experience here is almost a microcosm of the fate of all reusability players—the highs and the accidents are often separated by just a single ignition, and whether a company can climb back out of an explosion says more about its true mettle than a moment of success.

The Newcomers: Three Branch Lines from Europe, Japan, and India

One tier further back are the newcomers still doing validation at low altitude and suborbit, and they represent three different branch lines of reusability.

Europe takes the path most similar to Zhuque-3's. The Themis demonstrator, jointly developed by ArianeGroup and the European Space Agency, uses a hundred-ton-class liquid-oxygen/methane (methalox) engine, Prometheus—methalox once again. According to SatNews, Prometheus completed tests of four restarts in a single day and thrust modulation from 30% to 110% in 2025, and Themis plans to conduct its first hop test of about 20 meters at Esrange in Sweden in spring 2026. This progress is considerably behind even Zhuque-3's 10-km-class flight test—Europe is clearly a chaser in reusability.

Japan, meanwhile, contributes an "unexpected contestant": Honda. This company, famous for making cars, conducted a launch-and-landing test of a reusable rocket on June 17, 2025 in Taiki Town, Hokkaido; the test rocket rose to a height of 271.4 meters, with a landing deviation of just 37 centimeters and a flight of 56.6 seconds. According to Honda's official website, its goal is to conduct a suborbital test flight in 2029. A carmaker entering the field to do rocket recovery is itself a sign that the technical threshold of reusability is being reached by the manufacturing capabilities of more and more industries.

India, meanwhile, takes a distinctive branch line. The Indian Space Research Organisation (ISRO)'s RLV demonstrator "Pushpak" uses not vertical propulsive landing but a winged, gliding, horizontal landing like the Space Shuttle—in the LEX-03 test in June 2024, it was hoisted by a helicopter to a height of 4.5 kilometers and released, gliding autonomously and landing horizontally on a runway. This path of "gliding back with wings" and Zhuque-3's path of "landing vertically" are two completely different philosophies, and remind us that reusability comes in more than just the SpaceX paradigm.

Placing China Back on This Global Map

Now, let us place China back on this map. The national team's Long March 10B (CZ-10B), using the world's first sea-based net-capture recovery, succeeded on its maiden flight and vaulted into the very front rank of those that "have achieved orbital-class controlled recovery"; the private-sector Zhuque-3 sprints along the vertical propulsive paradigm, with Y1 having done flight validation and Y2 charging toward success. Taken together, China is the country with the deepest lineup and the most complete set of routes in this race, aside from the United States—it has both an original net-capture solution and a vertical solution benchmarked against SpaceX, plus a whole private tier queued up behind.

This position is a subtle one: in terms of "having achieved orbital-class recovery," China has already entered the door of the first tier; in terms of "stably, at high frequency, and at low cost repeating recovery," China still has a considerable way to go from the kind of maturity SpaceX has of "flying 36 times, recovering nearly six hundred times." In other words, China has already proven it "can do it," but has not yet proven it "can do it well, cheaply, and as a daily routine."

The Gap and the Coordinates

Our judgment on this international benchmarking is: there is no need to be complacent just because one has entered the first tier, nor any need for excessive self-deprecation just because one lags behind SpaceX. What is truly valuable is to see the coordinates clearly—China's position in reusability is that of a "chaser in the first tier," ahead of the low-altitude validation of Europe, Japan, and India, behind SpaceX's scaled operations, roughly on the same level as Blue Origin at "just having touched the threshold," and with a more complete set of routes than anyone.

And China's truly unique advantage may lie not in the technical parameters of a single rocket, but in the world's most complete, cheapest, and most resilient manufacturing system standing behind it. SpaceX's strength lies to a large extent in its manufacturing capability of building rockets as factory products; and "making complex products cheaply and in large quantities" is precisely Chinese manufacturing's deepest moat. As the competition in reusability moves from "can it fly back" into the second half of "who can build it in greater quantity and more cheaply," the depth of China's supply chain will become an increasingly critical variable. Just how deep this supply chain is, and where manufacturing enterprises should cut into this emerging commercial-aerospace market, is the map the next chapter will unfold.

11. A Map of Industrial Opportunity: How Manufacturing Can Cut into the Private Rocket Supply Chain

The preceding ten chapters were about the rocket itself. This chapter shifts perspective—standing in the position of a Chinese manufacturing enterprise and asking: what does this wave of reusable rockets have to do with me? Can I—and where should I—cut into this supply chain that is taking shape?

This question is especially apt for the Tianxia Gongchang Industry Research Institute. What we have always cared about is not how high the rocket flies, but how many opportunities for Chinese factories lie hidden in the manufacturing chain that lofts the rocket into the sky. So this chapter conducts its inventory with a discipline of "three tiers of credibility": which claims are named and confirmed by rocket companies officially or via corporate announcements (marked "Confirmed"); which are enterprises' own statements that they serve aerospace but without model-specific backing (marked "Self-reported"); and which are merely inferences by brokerage research reports or financial self-media (marked "Speculative"). Distinguishing these three tiers clearly is the bottom line of industrial research—because on this topic, what is least in short supply is the hype of directly calling "a company that makes some kind of material" a "rocket supplier."

A Supply Chain Opening Up to Manufacturing

First, look at the skeleton of this chain. A reusable methalox rocket slices demand into six major blocks: the superalloys and additive manufacturing of the engine and hot section; the stainless steel and composites of the airframe; the mechanical components used for recovery; onboard electronics and telemetry, tracking and command; propellant and ground fueling; and the industrial parks that carry all of this. Each block corresponds to one or several mature industrial clusters within Chinese manufacturing.

This chain has an enticing feature: it is new enough that its structure is far from solidified. The traditional national-team aerospace supply chain has, for decades, been supplied layer upon layer by institutes within the system, and outsiders find it very hard to break in; whereas private reusable rockets are new territory, where the supply relationships of many segments are still forming, and there are even entirely blank blocks. For Chinese factories with the corresponding manufacturing capabilities, this means a window of opportunity. But precisely because it is new and hot, this chain is full of the noise of concept-chasing—to discern the real opportunities amid the noise is exactly where the value of research lies.

The Engine Hot Section: The Only Entry Point with a Name

The engine is the segment of a reusable rocket with the highest technical threshold, and also the racetrack where domestic substitution is most concentrated, with feasible entry points concentrated in superalloys, precision casting, and metal additive manufacturing.

Within this segment, there is one entry point "with a name" for Zhuque-3 that is worth recording prominently: BLT (Bright Laser Technologies) of Xi'an, Shaanxi. In its 2025 interim report, it named that it had aided LandSpace's Zhuque-3 reusable rocket in its first large-scale vertical takeoff and landing flight test, driving key components from engineering validation toward batch production; it had previously also undertaken the metal 3D printing of the gas-generator body and combustion chamber of the Tianque (TQ) engine, a collaboration documented by 3D Science Valley. Among all the enterprises reviewed in this report, BLT is one of the very few that has corporate-announcement-level naming of "Zhuque-3 itself"—this belongs to the "Confirmed" tier, with the highest gold content. Metal 3D printing has become standard equipment for rocket engines because it can form irregular parts such as turbopump housings, combustion chambers, and nozzles with cooling channels in one piece, eliminating large numbers of weld seams and shortening cycle times—precisely suiting the appetite of reusable rockets for "batch production and low cost."

For the superalloy precision-casting segment, the most solid sample is Yingliu Group in Huoshan County, Lu'an, Anhui. By its public account, it is a continuous supplier of superalloy castings for LandSpace's Zhuque-2 methalox rocket—note that what is named here is Zhuque-2, belonging to the "Confirmed" tier; whereas the circulating claim of "exclusive castings for the Tianque turbopump" comes from an investor article and can only be counted as "Speculative." As for the upstream of superalloy materials, Gaona Aero Material (CISRI-Gaona) of Beijing, Tunan Alloys of Danyang, Jiangsu, and Fushun Special Steel of Fushun, Liaoning all possess aerospace-grade superalloy capabilities, but their relationships to specific rocket models are mostly research-report phrasing or in the aero-engine direction, and talk of supplying Zhuque-3 does not even reach "Confirmed." This kind of restraint at the granular level is precisely what distinguishes a research institute from a stock forum.

Airframe Materials: Three Openings in Carbon Fiber, Aluminum Alloy, and Stainless Steel

The airframe block has three openings, each with its own quality.

Carbon-fiber composites are used in payload fairings, interstages, and engine casings, and the domestic leaders—Sinofibers Technology of Changzhou, Jiangsu; Weihai Guangwei Composites of Weihai, Shandong; and Zhongfu Shenying of Lianyungang, Jiangsu—all possess mass-production capability for aerospace-grade carbon fiber and have all publicly stated that they serve commercial aerospace. But to be honest, their statements about rockets currently all stop at the "Self-reported" tier of "having the capability, already collaborating," and not one has produced model-specific backing as a "Zhuque-3 carbon-fiber supplier." The aluminum-alloy opening is a bit more mature: Southwest Aluminum of Chongqing is a core domestic base for aerospace aluminum materials, and its large-diameter aluminum-alloy integral forged rings are applied on Long March 5—this is the "Confirmed" tier, but it points to the national team's Long March series, not a private rocket. As for Zhuque-3's most distinctive stainless-steel airframe, which steel mill exactly supplies its steel cannot be found from an authoritative source through public channels—this is a real information gap, which this report faithfully marks and does not guess at.

Connecting these three openings, one intriguing phenomenon is: Chinese enterprises capable of making the relevant materials are actually everywhere, but those genuinely named "Confirmed" by rocket companies as supplying a specific model are very few. This gap is caused partly by information opacity, and also shows that the formal supply relationships of this new supply chain are still being established—for factories, this is both a regret not yet nailed down and a space still up for grabs.

The Three Major Recovery Components: An Honest Blank

What best embodies the value of a research institute is the courage to point out a "blank."

For the three most emblematic recovery components of a reusable rocket—grid fins, landing legs, and thrust-vector actuation servos—one can hardly find verifiable external professional whole-component suppliers in public information. The reason is that leading rocket companies generally emphasize in-house development of these core mechanisms: Deep Blue Aerospace, which takes the same vertical-recovery route, developed China's first all-carbon-fiber landing-buffer mechanism to enter engineering application for Nebula-1, handling everything from deployment to shock absorption itself, without disclosing any external whole-component supplier; and the manufacturer of Zhuque-3's landing legs likewise cannot be found. On the actuation-servo side, Beijing Hangyu Servo commands multiple types of servo mechanisms and has disclosed clients including Orienspace and CAS Space, among others, but has not named LandSpace, so it only counts as "Self-reported."

This blank must come with an eye-catching pitfall warning: those "net," "cable," and "hook" suppliers named by stock commentary in the sea-based net-capture recovery route belong to a completely different technical route, and must absolutely not be mixed into the supply chain of vertical landing legs like Zhuque-3's—this kind of mix-up has already, in past supply-chain reporting, produced lessons of being contradicted by corporate announcements. We treat this blank itself as a conclusion: the external supply chain for the three major recovery components has not yet taken shape, which both shows how high its technical threshold is, and is precisely a door of opportunity open to Chinese factories with capabilities in precision mechanisms, servos, and composites. A blank is sometimes more valuable than a list.

Onboard Electronics and Telemetry, Tracking and Command: First Distinguish Capability from Contract

The segment of onboard electronics and telemetry, tracking and command most needs sober judgment. The flight control and avionics of reusable rockets are mainly developed in-house by the rocket companies, and whole assemblies are rarely purchased externally—Zhuque-3 achieved, for the first time on a reusable rocket, an integrated electronics system unifying control, measurement, and health management, belonging to the rocket side's in-house architecture. So the external opportunities in this segment lie more in component-level domestic substitution, ground-station networks, and commercial telemetry, tracking and command services, rather than flight-control assemblies.

This segment is also a disaster area for "concept-chasing," and two corrections must be made. First, the market once widely rumored that a certain listed company participated in Long March 10B's sea-based recovery, but that company formally announced on July 12, 2026, clarifying that its business has no connection to that project and that it did not participate in the development or supply—this reminds us that an unconfirmed "supplier concept" can be punctured by a single announcement at any time. Second, Gaohua Technology of Nanjing, Jiangsu has named supply for Long March 3B (the "Confirmed" tier), but with LandSpace it is only a "Self-reported" "establishment of collaboration," while the circulating talk of "supplying 1,600 sensors to Zhuque-3, exclusively" and the like is media phrasing that does not reach "Confirmed." The truly documentable industrial opportunity is a third-party commercial telemetry, tracking and command operator like Aerospace Yuxing (Space Yuxing) of Haidian, Beijing—its client list includes rocket companies such as Galactic Energy and Space Pioneer (the "Confirmed" tier), providing telemetry, tracking, command, and data-transmission services for the launch phase, early orbit, and return phase. Components, ground stations, telemetry and command services—these are where manufacturing and technology enterprises can steadily cut in.

Propellant and Ground: The Segment Most Like "Selling Water"

If the preceding several segments all require competing on hardcore technology, then the propellant and ground-support segment follows the most classic "selling water" logic—the success of gold-panning is uncertain, but supplying water and shovels to the gold-panners earns steady money.

This segment is precisely where the "Confirmed"-tier evidence is most solid, because it hangs on official tenders in black and white. Hainan International Commercial Aerospace Launch Co., Ltd. publicly tendered for the procurement of 5,844 tons of liquid methane for launch missions from 2026 to 2028, as well as two fixed launch pads—this kind of official tender is itself the hardest evidence on the demand side, showing that liquid-oxygen/methane has become a standard propellant procurement category for launch sites. On the supply side, Hainan Jiufeng Special Gas of Wenchang, Hainan (controlled by listed company Jovo Energy) built a special-fuel-and-gas base beside the launch site, with the capability of producing 48,000 tons of liquid oxygen and 20,000 tons of high-purity liquid methane per year, and the project's approval has formal-announcement backing; Hangyang (Hangzhou Oxygen Plant Group) of Hangzhou, Zhejiang self-reports that it has provided liquid-oxygen propellant and aerospace-grade krypton and xenon to the Zhuque series; CIMC Enric of Hebei self-reports that its in-hand orders for liquid-oxygen and methane cryogenic storage tanks in commercial aerospace have exceeded 100 million yuan; and Aerosun of Nanjing, Jiangsu is a designated plant for aerospace metal hoses. Among these enterprises, the launch-pad hardware itself and the launch-site tender belong to the "Confirmed" tier, the enterprises' supply statements are mostly of the "Self-reported" tier, while those claims by financial self-media of "a certain company winning a several-hundred-million-yuan ground-system contract for Zhuque-3" are, without exception, "Speculative" and not worth crediting.

Once liquid-oxygen/methane became the mainstream fuel of private rockets, the special fuels and gases, cryogenic fueling, and storage-and-transport equipment for a single launch carry considerable value, and demand grows rigidly with launch frequency—this is the segment on this chain with the clearest commercial logic and the best suited for scaled manufacturing enterprises to cut into.

A Map of Opportunity, and Its Three Disciplines

Gathering the six segments into a single map, we then distill three disciplines for manufacturing enterprises.

First: Distinguish the model. Those with corporate-announcement-level naming for "Zhuque-3 itself" are, at present, mainly BLT (additive manufacturing); Yingliu has a casting-supply account for Zhuque-2. The large number of remaining enterprises are transfers from "the Long March series," "other commercial rockets," or supply-chain inference; before writing them into one's own business plan, one must be sure to distinguish exactly which rocket is being supplied. Second: Recognize the blank. The external supply chain for the three major recovery components is basically blank, which is not bad news but a window of opportunity not yet carved up—whoever can produce reliable grid fins, landing legs, and servo mechanisms and be trusted by rocket companies may fill this blank. Third: Beware of concepts. Any supply claim backed only by brokerage research reports, stock forums, or wealth accounts, without a corporate announcement or rocket-company confirmation, must be marked with a question mark—that debunking announcement by Spacety-Yuda (Xingwang Yuda) is the latest cautionary tale.

Behind these three disciplines is, in fact, something more fundamental: for a manufacturing enterprise to cut into commercial aerospace, the first step is not to build connections or chase hot topics, but to first precisely match its own real capabilities to the real demand of this supply chain. This is exactly what the Tianxia Gongchang Industry Research Institute and the industry platform behind it—which gathers 4.8 million operating factories—want to do: to honestly and precisely connect the real manufacturing capabilities scattered across the country's industrial belts with emerging demand like commercial aerospace. Which factory in which county can do aerospace-grade superalloy precision casting, can weld large-diameter stainless-steel propellant tanks, can make precision servo mechanisms—this granular information is the true starting point for a factory to actually cut into the rocket supply chain.

Having reviewed this supply chain, the final question this report must answer also emerges: when the rocket goes from a "weapon of the nation" to a "factory product," what does this really mean for Chinese manufacturing? That is the main thread the conclusion will gather up.

12. Conclusion: From a "Chinese Falcon" to China's Own Reuse Paradigm

This report began with that Y2 vehicle waiting to launch on the test stand at Jiuquan, took a long detour through engines, materials, control, the ledger, demand, benchmarking, and the supply chain, and finally returns to the most plain judgment—the true significance of Zhuque-3 lies not in whether it is a "Chinese Falcon 9," but in the fact that it marks China forming its own reuse paradigm, and that the base color of this paradigm is manufacturing.

From "Weapon of the Nation" to "Factory Product"

The rocket was once a typical symbol of a "weapon of the nation": costly to build, few in number, forged by the nation's full effort, each one a meticulously carved unique piece. What reusability sets out to do is, in essence, to invite the rocket down from this altar and turn it into a "factory product" that can be mass-produced, used repeatedly, and made cheap and rugged.

Every key choice embodied in Zhuque-3 moves in this direction. Choosing stainless steel over aluminum alloy is for cheapness and durability; choosing liquid-oxygen/methane over kerosene is so it does not coke during reuse and is easy to maintain; using nine engines in parallel rather than a few powerhouses is to spread out cost through batch production; using laser welding and metal 3D printing is to be able to automate and mass-produce. Each of these choices, viewed alone, is a technical detail, but strung together they form a complete shift in thinking: from "building the single best rocket" to "building a batch of the most cost-effective rockets." The former is the logic of aerospace engineering, the latter the logic of manufacturing. Whoever completes this shift holds the key to opening the low-Earth-orbit era.

This is also why we said from the very beginning that reusability is on the surface an aerospace proposition but at its core a manufacturing proposition. What it tests is not whether one can build a stunning rocket, but whether one can build the rocket cheaply, ruggedly, in batches coming off the line, and able to withstand repeated recovery and refurbishment—every one of these tests a nation's manufacturing foundation.

Two Routes, One Answer

In this summer of 2026, the answer China gives is especially interesting: it did not bet everything on a single route, but let the national team and the private sector each take one. Long March 10B, using the world's first sea-based net-capture recovery, transfers the difficulty of recovery onto China's strongest offshore-engineering and shipbuilding capabilities; Zhuque-3, following SpaceX's vertical propulsive paradigm, gnaws on the hardest bones of the engine, control, and precision manufacturing.

On the surface these are two diametrically opposed routes, but viewed more deeply they are two sides of the same answer—both routes play to the strengths of Chinese manufacturing. Net-capture recovery plays to the strengths of "large structures, large equipment, large ships," while vertical recovery plays to the strengths of "precision manufacturing, batch production, cost control." China did not simply copy a SpaceX, but, according to its own industrial endowments, simultaneously explored two roads toward reuse, with a whole private tier queued up behind to try still more possibilities. This approach of "combined bets, each playing to its own strength" is itself the most composed posture for a great manufacturing nation facing a new racetrack.

After Y2: An Unfinished Story

It must be honestly said that this story still had no ending at the time of this report's writing. Y2 had not yet launched, private China's orbital-class first-stage recovery had not yet truly succeeded, Zhuque-3's cost target of "from 50,000 down to 20,000" still remained on paper, and the vast majority of those 200,000 declared satellites had not yet gone into space. This is a story in progress, full of uncertainty—Y2 could entirely fail once again in the final dozen or so seconds, just like its predecessor Y1, and just like New Glenn across the ocean, which had just blown up its launch pad in a static-fire test.

But it is precisely this incompleteness that makes it worth recording carefully and tracking over the long term. We judge that, regardless of whether this flight of Y2 succeeds or not, the broad direction of the "manufacturing-ization" of China's reusable rockets is already irreversible—because behind it are the hard constraints of the ITU clock, the rigid demand of constellation deployment, and the commercial logic that private companies must rely on reuse to survive; these underlying forces will not change because of a single success or failure. Y2 will succeed sooner or later, and after success there will be Y3, Y4, there will be the first re-flight of a first stage, the second re-flight. This road will be walked with bumps and stumbles, but it will keep moving forward.

The Manufacturing Ticket to the Low-Earth-Orbit Era

Standing in the position of the Tianxia Gongchang Industry Research Institute, what we most want to leave readers with is not a technical worship of a single rocket, but a judgment about opportunity.

When the rocket goes from a weapon of the nation to a factory product, it is no longer only the business of aerospace institutes, but becomes a supply chain open to the whole of Chinese manufacturing—from a workshop in a county town that does superalloy precision casting, to a factory that can weld large-diameter stainless-steel propellant tanks, to a team that can make precision servo mechanisms, all may find their place on this new chain. This chain is still young, still full of large blanks, still full of noise, but precisely for this reason the window is especially precious. And to precisely connect the real manufacturing capabilities scattered among the country's 4.8 million operating factories with emerging demand like commercial aerospace is exactly the meaning of this industry platform's existence—what it cares about has never been how high the rocket flies, but how many new opportunities for Chinese factories can grow on the manufacturing chain that lofts the rocket into the sky.

That moment of Zhuque-3 Y2 "catching itself" catches not just a single first stage. What it catches is a ticket for Chinese manufacturing to the low-Earth-orbit era. How much this ticket is worth, and how far it can carry Chinese factories, is worth all of us watching over the long term, with patience.

Data Sources & Key References

The facts, data, and timeframes in this report all come from publicly verifiable authoritative sources and have been cross-checked across multiple sources; for claims that are disputed or found only through a single non-authoritative channel, the main text has clearly marked phrasing such as "unconfirmed" and "according to market speculation," and draws no definitive conclusion. Technical parameters are based on LandSpace's official press releases and authoritative media reporting; commercial and industrial data are based on official announcements, tender documents, and mainstream financial media reporting.

• Tianxia Gongchang Industry Platform—a database of China's operating factories and industrial-chain data (gathering some 4.8 million operating factories), the data support and attributed data source for this report's supply-chain granularity and industrial-belt positioning.
• LandSpace official website (LandSpace)—the official technical account for Zhuque-3 Y1/Y2, the Tianque engine, and the 10-km-class vertical takeoff and landing test.
• Xinhua News Agency / Xinhuanet, China Media Group (CCTV.com)—national-level authoritative reporting on the Y1 maiden flight, Long March 10B sea-based net-capture recovery, and more.
• Science and Technology Daily, People's Daily Online / People's Daily—reporting on Y2 progress, static-fire tests, and industry developments.
• China National Space Administration (CNSA)—commercial aerospace launch sites, industry policy, and action plans.
• SpaceNews, NASASpaceflight, AIAA, Space.com—authoritative international aerospace reporting on Zhuque-3, Falcon 9, Starship, New Glenn, and more.
• China-in-Space, Global Times (English edition)—English-source cross-checking of Zhuque-3 recovery details and progress.
• Cailian Press, Jiemian News, Shanghai Securities News, and other mainstream financial media—reuse economics, supply-chain enterprises, and official tender information (for listed companies, only business, technology, capacity, and contract subjects are cited, not share prices, market capitalization, or valuations).

(Note: Zhihu, Baidu Baike, Sohu, and industry self-media are used only as leads and are not cited as authoritative bases for this report.)