On August 19th at 7:35 AM, the 25-story, 600-tonne Zhuque-3 rocket once again escaped Earth's gravity, reaching low Earth orbit. Just over two minutes into the flight, the booster and the second stage carrying a satellite separated successfully. Less than ten minutes after launch, the booster executed two retro-burns and landed softly at a recovery site in Minqin County, Gansu, roughly 390 kilometers from the launch pad. This marked the second time in China's space history that a booster was successfully recovered after orbital insertion, following the Long March 12B, making its manufacturer, LandSpace, the first private Chinese company to recover a rocket—a resounding payoff for 11 years of investment and anticipation in China's commercial space sector.

The rocket is comparable in size to SpaceX's workhorse Falcon 9, yet its overall technical approach is similar to Starship, utilizing a stainless steel fuselage and liquid oxygen-methane engines. This design could potentially enable launch costs lower than those of the Falcon 9. The Zhuque-3 employs a landing strategy with legs that slightly compromises payload capacity. This choice, however, allows for faster turnaround times for refurbishment and reflight, as engineers only need to maintain the rocket itself, avoiding the need to repair ground-based recovery infrastructure.

Commercial space is one of the hottest technology sectors today. China currently has over 20 rocket companies and more than 80 rocket models in development. When LandSpace was founded in June 2015, SpaceX's Falcon 9 had yet to achieve a successful recovery and reuse, and China's aerospace industry was just opening to private capital. As one investor put it, the entire industry had "nothing but opportunity."

Now, with over 1,200 employees and a valuation exceeding 75 billion yuan, LandSpace is the most successful company in China's commercial space sector. Its rise is due to early and unwavering commitment, along with a highly self-reliant development model. Founder Zhang Changwu recognized early on that critical subsystems "cannot be bought." Starting in 2017, LandSpace began developing its own engines and was among the first to build its own test stands, launch sites, and rocket factories. This approach has allowed the company to master most key technologies and iterate quickly internally, independent of external supply chains.

Dong Kai, Deputy Chief Designer of the Zhuque-3, told us that when he first met Zhang Changwu, he asked, "Does LandSpace want to build a reusable rocket?" Zhang said he had always wanted to, but he wasn't sure if the company had the resources to pull it off. At the time, the Zhuque-2's maiden flight had just failed. To save resources for the next rocket, LandSpace had to pause non-essential infrastructure expansion. The only major project that remained fully funded was the reusable rocket program. Dong told Zhang the test vehicle might cost 50 million yuan. Zhang replied, "Can we do something this big for just 50 million? No problem, full support."

The Zhuque-3 was officially greenlit in 2023 and completed its first flight at the end of last year. After more than 490 seconds of flight, the rocket had reached orbit, and the first stage entered its recovery sequence. Dong Kai was in the command center. Seeing the center engine ignite for the landing burn and the rocket begin its slow descent, he hadn't even begun to hope for a first-attempt success when a fireball erupted—the rocket hit the ground. Some colleagues, experiencing this for the first time, were visibly upset. But Dong called out over the comms: "Bring up the big red screen, let's celebrate." One of the two core missions had succeeded, the other had failed. "The results exceeded expectations," he explained.

Rather than fixating on the recovery success itself, Dong places more value on LandSpace's engineering accumulation. When you have enough information, success becomes a natural outcome. His job is to make the rocket as simple, efficient, and reliable as possible, allowing the Zhuque-3 to fly and practice more often within shorter timeframes and at lower costs.

We met Dong Kai at LandSpace's Beijing headquarters in mid-July, shortly after he returned from ground tests in Gansu. A graduate in aircraft design from Harbin Institute of Technology, he initially expected a career in the military-industrial system, "hidden in anonymity." In 2009, he joined the overall design department of the rocket research institute under CASC, where he received comprehensive engineering training and worked on the Long March 5, China's most powerful rocket at the time. He joined LandSpace in 2021 and has been involved in the Zhuque-3's design and development throughout.

In Dong's account, there is little room for serendipity or "overtaking on curves" in the Zhuque-3 project. They simply followed their plan, step by step. The successful recovery is a romantic and symbolic milestone, but LandSpace is more focused on accumulating technical and engineering capability. The team will soon begin inspecting the recovered rocket, aiming to send it back into space as quickly as possible.

Recovery is just the first step; next comes disassembly to assess the health of every component.

What improvements were made after the last failed recovery attempt?

Dong: First, we sacrificed some payload weight to strengthen the thermal protection structure. Second, we added gas cylinders in the aft compartment to purge methane before ignition. The previous failure occurred during the landing burn. The plan was to ignite the center engine first, then up to four others as needed to slow the descent, eventually throttling down to one engine before deploying the landing legs. The anomaly happened in the aft compartment right after the center engine ignited, but before the other four were lit.

How long was the preparation between the first and second launches, and what did that involve?

Dong: The launch was on December 3rd. We had almost all telemetry and optical data by the next day. There wasn't much time to celebrate or dwell on the failure. We entered the formal review and fault tree analysis process on December 5th. The aerospace industry has a strict methodology: "presume guilt, require evidence." For example, if the aft compartment exploded, there must have been a fire, combustion, and propellant. So, where did the propellant come from? Where was the leak? This line of reasoning breaks down into a series of underlying events. Each eliminated risk must be supported by objective evidence; we can't rule out possibilities based on opinion. It's like forensic detective work. After eliminating all impossibilities, whatever remains—no matter how improbable—must be the truth. There's only one truth, just like in Detective Conan.

How long did that take?

Dong: About two to three months. For us, a rocket is a significant investment. If we missed the real culprit, the tuition would have been wasted. Once we identified the fault, we moved to fixes. The engineering changes were relatively straightforward; the rest was solving the problem in terms of cost, time, and materials. Ultimately, these are solvable engineering problems. We don't manufacture hardship for ourselves or show off. We choose the most practical path.

A key commercial space logic is iterating based on data from each test. Did the first flight's data collection meet expectations?

Dong: Yes. The key is whether the actual flight data matches our simulations and analysis. Sometimes a successful flight just means the design margin was large enough to cover the discrepancies, not that our predictions were accurate. We saw some deviations and corrected our ground models accordingly. For example, we reduced the number of engines used for the landing burn from five to three, a change driven directly by the data we collected. There's a saying in aerospace: "Entities must not be multiplied beyond necessity." We initially planned for five engines because another model had experienced insufficient aerodynamic braking and needed multiple engines to compensate. After the first flight, we realized the risk wasn't that high, and three engines could handle the landing. The extra two engines would have just added unnecessary failure points. Every solution is a trade-off between benefit and cost.

Now that recovery is successful, what's next?

Dong: We're aiming to refly the recovered rocket as soon as possible. After it returns, we'll disassemble it meticulously to identify any damage from the flight. If every part has a "health bar," our future work is to make each one inspectable, measurable, and repairable, then get it back on the launch pad. The main improvements for the Zhuque-3 are increasing payload capacity, enhancing reliability, and shortening the turnaround time. My internal goal is to get the Zhuque-3 past the threshold of 10 launches a year first.

What's the target turnaround time for reuse?

Dong: Currently, considering our stainless steel production method and engine manufacturing capacity, we can build a Zhuque-3 in a few tens of days on average. We need to make the recovery-to-reflight cycle shorter than the manufacturing time. If it takes four to five months to refurbish a rocket, it would be faster to just build a new one. In games, time-accelerating or time-rewinding magic is a rare item. If I build 10 rockets and reduce the turnaround to two months, that's 10 times 6, or 60 launches a year. If I can get it down to one month, my launch capacity doubles again.

Compared to the Falcon 9, what stage is the Zhuque-3 at now?

Dong: On a scale of 1 to 9, LandSpace is just crossing from Level 6 (laboratory validation) into Level 7 (real-world flight testing). The higher you go, the harder it gets. Level 8 is marked by achieving reflight. Level 9, the highest, is high-confidence, mass-scale launches. It's like in the novel *A Record of a Mortal's Journey to Immortality*. If the Falcon 9 is a "Spirit Severing" cultivator, then the Zhuque-3 is currently at the "Core Formation" stage. Once we achieve reuse, we'll have reached "Nascent Soul."

Any regrets from the design to launch process of the Zhuque-3?

Dong: Not many regarding the product design itself. The first flight result was a minor imperfection, not a regret. My only real regret is that the first flight was a bit later than planned.

What is the main bottleneck for LandSpace's launch capacity now?

Dong: We haven't hit our limits in terms of launch pads or production capacity. LandSpace can produce about 20 rockets a year. This is also partly related to the strictness of the current launch licensing approval process. Early exploration tends to be more cautious. Even Musk has complained about the FAA, and they treat him quite well now.

What would make the policy environment more relaxed?

Dong: Some things are "believed because they are seen." So, the bottleneck is still the same: the Zhuque-3 hasn't launched enough times. Our reliability and data accumulation aren't comprehensive enough to give everyone complete confidence. We just need more practice. The more we fly and the higher our success rate, the faster things will naturally accelerate.

LandSpace has its own "Skunk Works" for developing core technologies.

What was the logic behind choosing stainless steel for the Zhuque-3's fuselage?

Dong: To solve manufacturing challenges. I worked on the Long March 5, where scaling the fuselage from 3.35 meters to 5 meters was incredibly difficult. We wondered if there was a simpler solution, and stainless steel was the answer. First, welding it is much easier than thin-walled aluminum alloys. Thin-walled aluminum pressure vessels are almost exclusively used in aerospace; no other industry has that technology, so the supply chain is very limited. With stainless steel, the supply chain is vast. Second, it's cheap—tens of thousands of yuan per ton. Third, it has excellent heat resistance, eliminating the need for a thermal protection layer. This isn't just about cost and weight. If we have a heat shield, we'd have to maintain it after every recovery, which is a huge workload and requires complex repair processes. Fourth, it offers great scalability; stainless steel can be made very thin.

SpaceX announced they would use stainless steel in 2018. If it's so great, why didn't anyone use it before?

Dong: We considered it earlier. Back then, the aluminum alloy supply chain was too limited. Because its use is so narrow, supply and production were almost entirely concentrated within the state aerospace system. We could have built our own, but that would mean retracing the same steps others had already taken. Time-wise, it wasn't the optimal solution. Also, the early environment wasn't as favorable as it is now. We had to first prove we were capable, so some early decisions were forced. Plus, there's a perception that stainless steel is too heavy. Its strength is higher than aluminum, but its density is also higher, making it heavier overall. But in reality, based on material strength, a rocket wall only needs to be a few millimeters thick. The problem is that producing something that thin is difficult; welding causes deformation. On a 20-meter-long stainless steel cylinder, the front end might be perfectly round when you start welding, but by the time you reach the back, it's become an ellipse due to cumulative deformation. After enough welding, you can't meet the tolerance requirements for mating with other sections. Welding itself is simple. Any tank manufacturer in the Yangtze River Delta or Jing-Jin-Ji region could do it. But they can't weld it this thin, nor can they maintain the required geometric tolerances at this scale.

How did you solve the welding problem?

Dong: We focused on process development, using laser welding, which has the lowest heat input. The thin-wall welding issue is specific to the 4.5-meter diameter. If you scale up to 9-10 meters, like Starship, the strength requirements are higher, the wall needs to be over 2 millimeters thick, and the welding difficulty decreases. LandSpace has a department called Tianma Laboratory, similar to Lockheed Martin's Skunk Works, which is tasked with solving these cutting-edge stainless steel process challenges. We now use automated laser welding. But we didn't start with machines. We accumulated experience step by step, first observing how much 2-millimeter material deforms and where, then compensating in the robot's path planning. We also studied the material itself, looking not just at deformation but also at its intrinsic properties, like temperature during welding. Through extensive time and data accumulation, we progressed from 2 millimeters to even thinner, eventually achieving very thin walls with minimal error and high production yield. Only then did we freeze the data and parameters. Every millimeter saved on the fuselage directly translates to improved rocket performance.

So the welding improvements were developed in-house, without suppliers?

Dong: Yes, the entire automated welding production for the rocket is solved by LandSpace. We don't reinvent the wheel for things suppliers do well, but outsourcing rocket welding gives us no advantage. Initially, we wanted to validate stainless steel's strength, so we hired welders from the Yangtze River Delta, giving them 2-millimeter plates to test with any method, even manual TIG welding. That only solves the "yes/no" problem. To truly optimize performance, we have to do it ourselves. Also, managing quality control with an outsourced supply chain is difficult. A failure of the fuselage pressure vessel would be catastrophic. We have strict requirements for maintaining the shape and controlling tolerances on the main rocket body.

When you approached welders at steel plants in the Yangtze River Delta to weld a rocket, what was their reaction?

Dong: They were fine with the diameter and volume. You have to trust the ingenuity of the private supply chain, and they are extremely cost-conscious. But when we specified the maximum thickness, the tolerance range, and the requirements for geometric tolerances, alignment, parallelism, and roundness, they said they'd never seen such demands. Normally, clients worry about them cutting corners and want things welded thicker. Asking them to weld thinner was outside their expertise. Other steel-using industries rarely have these requirements, so there's no incentive for them to master such a difficult process for a niche application like rockets.

Is the stainless steel you use special, or is it standard?

Dong: It's standard, off-the-shelf material, costing tens of thousands of yuan per ton. Even though we use small quantities, we treat the rocket as an industrial product, not a luxury item. There's no need for a highly customized, exotic material.

After switching to stainless steel, what other parts of the rocket had to be redesigned?

Dong: There weren't many major hardware changes. We did rethink some of the systems. Since we're using stainless steel, we might as well push it to its limits. For instance, in the return guidance strategy, the stronger fuselage allows us to choose a more aggressive trajectory with higher dynamic pressure. If using stainless steel forced us to add weight elsewhere, it would indicate a fundamental flaw in the design. We don't have "unlimited ammunition." Every technical choice is about solving a problem, not creating one just to prove we can overcome it.

How does being a commercial company with finite resources change your approach?

Dong: Commercial companies are very cautious when choosing a rocket's diameter because different diameters require vastly different infrastructure investments. The state system doesn't have to think about this as much; they can find the optimal data, whether it's 4.37 meters or 7.5 meters, and build the ground support to match. Commercial companies want to minimize unnecessary investment. We only choose diameters that the national team has already used, like 4.2 meters or 3.35 meters, so we can use existing ground support and tooling interfaces. Again, entities must not be multiplied beyond necessity.

How did you manage to build the entire rocket in 28 months?

Dong: We had a head start. The Tianque-12A engine used on the Zhuque-3 was developed during the Zhuque-2 era. We also began validating the manufacturing processes for the Zhuque-3 with a test vehicle starting in late 2022. This accumulation of experience and lessons learned prevented us from repeating mistakes. There are no shortcuts in engineering, only ways to avoid detours. We never emphasized 996 work hours. Instead, we clearly communicated the project milestones and ensured a free flow of information across all departments. In an organization with a strong sense of honor, when everyone else is pushing to meet deadlines, how can you tolerate your own progress lagging behind? That's how we spent those 28 months.

Managing an aerospace project

How do you break down a project as large and heavy as a skyscraper that needs to fly to space and return safely?

Dong: The project is structured around major ground tests. Tests like modal analysis and engine static fire tests have specific requirements for each system, defining what state they need to be in and by when. It's like an army converging for a battle; each unit can only advance ahead of schedule, never fall behind.

So each system has autonomy, as long as they achieve their goals on time?

Dong: A supreme commander doesn't need to direct every platoon. As an army group commander, I just tell a division where to be and when. The division commander has the autonomy to figure out the details. This is a hallmark of LandSpace's management style: full delegation.

Is that also how it's done in the state system?

Dong: I don't think it's a difference between the state and private systems, but rather between different rocket programs. If a program faces intense external competition and has a hard deadline, then delegation is essential. The traditional approach does tend to be more micromanaged, even down to which document should be completed when. This method places a very high demand on the planning department. We recognize that we don't have the capacity to manage at that granular level, so we release some autonomy.

You mentioned the ideal state is autonomy. That's a high bar, especially in a field where success isn't guaranteed.

Dong: This culture of high honor isn't achieved through lectures or preaching. It relies on two things. First, systems. LandSpace advocates for "pragmatic innovation, decisive execution." We encourage experimentation and don't penalize people for trying and failing. Rewards and punishments are clear. Second, providing opportunities for real-world challenges. In any organization, the truly effective core is the top 20-30%. Their primary motivation often isn't just salary or bonuses, but a sense of achievement and self-fulfillment. LandSpace provides a platform for constant, real-world practice, which builds that sense of accomplishment. Think about a brilliant engineer: what's their biggest frustration? It's spending days working on PowerPoint presentations instead of doing what's truly important. Wasted effort breeds fatigue. We give people a stage. Honor is built through one battle after another. I believe this logic applies to most knowledge-based work.

Rocket engineers often come from physics or materials science. But now many AI companies are also recruiting the brightest minds. How do you attract top talent?

Dong: We have a different definition of talent. LandSpace wants "moldable talent." In a highly innovative industry like commercial aerospace, I don't think past education is the most important thing. It just means you can help us avoid a few detours. We place more value on growth potential and learning ability. The prerequisite is still shared values. People who choose this industry often have a sense of patriotism. We offer compensation that ensures they don't have to make a difficult choice between passion and poverty, like "I chose aerospace, so I must endure hardship."

You mentioned that many in this field have a deep love for it. How did your passion for aerospace develop?

Dong: My first rocket launch was the Long March 5 in 2016. At that moment, nothing else mattered. It was more than enough. The moment a rocket lifts off is an immense emotional shock and awe for anyone who builds it. Many people love it from childhood. For me, it was in 2003, when Shenzhou-5 launched and Yang Liwei went to space, that I suddenly felt this profession was truly noble. When I graduated and started working in 2009, it was the 60th anniversary of the PRC. The training and the songs we learned were steeped in patriotism. Entering this industry means making sacrifices due to confidentiality requirements. It's not as mysterious as weapons development, but there are still rules. The song we learned as new recruits was "The Motherland Will Not Forget." It was different from what we learned in middle school; the feeling was different. When you're surrounded by that atmosphere, you're unconsciously inspired and you genuinely feel your work is noble. I still feel that way. I inexplicably feel a sense of historical mission. From entering the institute to leaving it and moving into commercial aerospace, I've been driven by that original aspiration.

What was the most important thing you gained from your time at CASC?

Dong: Engineering discipline. I was lucky. Right after graduating, I joined the Long March 5 and the new generation rocket program. New projects are great for development. I still deeply admire the commanders I met back then. Like in the movie *Decisive Engagement: The Liajin-Shenyang Campaign*, when the commander says, "Liu Yalou, take a note, make the following deployment," you think, "This is a top-tier commander." I've worked with people like that, and now I unconsciously try to emulate their approach. This is the 50-year legacy of Chinese aerospace. Sometimes I see news reports treating the national team as some sort of "competitor" to be compared and contrasted with private companies. I think that's a big misunderstanding that greatly underestimates the decades of accumulated expertise and organizational capability of the Chinese space program.

What aspects of the national team's capabilities are underestimated?

Dong: First, it must always control the lower limit of what can go wrong. This is different from commercial space, which takes risks to pursue an upper limit. It's like why you go to a top-tier hospital for a serious illness instead of a street clinic. A clinic might advertise that they once cured a terminal case—they're chasing the upper limit. You go to a major hospital because its lower limit is stable and reliable. Aerospace decision-making works on this principle. Second, "all martial arts originate from Shaolin." The real "Whampoa Academy" for China's commercial space is the national team. Most of our early employees come from there. If you consider the national team as part of China's commercial space sector, it holds the flagship position and bears the burden of a lot of trial and error. We are able to avoid these detours today because they have already paved the way.

Looking at SpaceX's history, many things they did early on were also things NASA did before, just at a lower cost. But eventually, they shifted roles and took on a leadership responsibility.

Dong: Yes, that's right. On a larger scale, SpaceX is the big brother of world commercial space. Since Musk has proven it can be done, we can too.

Over the past two decades, every part of space technology has changed. What do you see as the most significant changes?

Dong: In 2009, China's aerospace industry was always talking about catching up with the world. Today, if it weren't for SpaceX, China would be number one, and the national team would be ahead of NASA. The United Launch Alliance, which we once thought was unassailable, has also undergone massive changes. Who would have thought Boeing's "space black shop" would have its comeuppance? Second, the development of heavy-lift rockets has exceeded expectations. Back then, it felt like working on a heavy rocket was a lifelong endeavor. The Long March 5, from feasibility study to launch, spanned nearly 40 years. When I first joined that program, a veteran colleague said, "You're very lucky to be part of a new rocket development." It was like a soldier finally getting a chance to fight in a major battle. Looking back, the changes from 2016 to 2018 were dramatic. SpaceX went from the Falcon 9 to the Falcon Heavy in just over two years. For me, that was a bigger shock than seeing the Falcon 9 landing. And Starship is even more so. Third, there are now perhaps 80 or more rockets waiting to launch in China, named or unnamed. In the past, there were only about a dozen in the Long March series. For practitioners, it feels like the spring of commercial aerospace has arrived. When I first started, I thought I might be lucky to work on a heavy rocket after the Long March 5. Now the landscape has changed, and the entire global aerospace industry feels more dynamic. When Musk talked about launching 42,000 satellites, we all thought it was an unrealistic number. That later became a lesson for me: never make hasty judgments. We have always had respect for SpaceX, not because of their grand narratives, but because of their Falcon 1, which failed three times and they still dared to organize a fourth launch. "They can do that? They dare to do that?" After that, it was like a psychological anchor: "We can fail three times and still succeed."

After the Zhuque-3, what is LandSpace's next key technology focus?

Dong: The most immediate priority for the Zhuque-3 is upgrading the engine to increase thrust. The Tianque-12B engine is scheduled for delivery between the end of this year and next year. In parallel, we're developing the full-flow staged combustion engine, the Lanyan, which has undergone numerous test firings, but it will take a while before it's integrated into a complete rocket. We're also in the technology reserve phase for a larger rocket, but the project timeline depends on market demand and how the Zhuque-3 performs. At this stage, we want to increase the Zhuque-3's launch frequency to 30-50 times a year.

That number doesn't sound very aggressive.

Dong: In 2023 and 2024, China's total annual space launch capacity was about 160 tons. What does that mean for the Zhuque-3? In reusable mode, that's just a little over ten launches.

Historically, even the US national team couldn't afford failures. Now it seems commercial companies have more room for error, but rockets are still expensive. How does LandSpace balance more attempts with safety and success rates?

Dong: These aren't necessarily contradictory. First, we absolutely do not allow management or process risks. Second, we have systems in place to ensure human safety—unmanned, automated, and remote operations.

But as a commercial company, you must also control costs.

Dong: Our core motivation for cost control is survival, not just improving a few percentage points of gross margin on a financial statement. We don't think, "If we reduce this cost from 50 million to 30 million, how much will our gross margin increase?" Instead, we ask, "If we save that 20 million, how many more tests can we run? How many more opportunities to learn?" There are certain red lines and bottom lines in design that we will not cross. On top of that, we try to reduce unnecessary, non-essential expenses.

It still sounds very difficult. What is the fundamental advantage of the commercial approach?

Dong: Greater freedom in the supply chain. The state system has basic quality control requirements for its suppliers. It uses supplier certainty to manage the uncertainty of new products. The state supply chain is constrained by both economic indicators and other regulations. It's not just a simple transaction. If there's a problem, they need to trace it back to an individual and potentially issue penalties. This model increases cost and lead times, and is less efficient, but it controls the lower limit. With a market-based supply chain, we can access cheaper components. This isn't about squeezing supplier margins. It's about the fact that we're a major industrial nation with many highly competitive industries that have good efficiency and yield. We can be bolder in our trade-offs between cost and reliability. However, greater freedom is both an advantage and a disadvantage. It can be harder to control the lower limit, which places higher demands on our own quality control. We can't just send a supplier a task order, have them say "done," and accept the product. We need our own methods to verify and test that they actually meet the requirements, rather than building reliability on their commercial credibility. In the past, a rocket designer's job ended when they signed off on a document. Now, they are responsible for the final outcome. In this sense, commercial aerospace places higher demands on technical personnel.

How does the organization, personnel, and collaboration differ from your time in the state system?

Dong: On the surface, it might look similar, but the essence is quite different. It all comes down to quality control. The old organization's approach was to have a dedicated person responsible for each thing, breaking the system down into very fine pieces to control the lower limit, ensuring stability even with staff turnover. The cost is increased system complexity. From an engineer's perspective, almost all quality problems occur at the interfaces between handoffs. It's like in war; to break out of an encirclement, you attack the seam between two enemy units. If it's a single unit fighting alone, they know they must hold their ground and won't let the enemy through. The traditional system relies on more elaborate procedures to ensure these handoff points don't fail. This leads to a situation where the state system has dozens of times more people. LandSpace, with its one or two thousand employees, is considered a large company in commercial aerospace. They have tens of thousands, but their efficiency isn't 10 times higher, because a significant portion of their workforce is dedicated to managing the risks that come with increased management complexity. They've found their balance. Commercial aerospace inherited this system but reduces the complexity. The key structures are retained, and the principle of "each unit holds its ground" remains, but the personnel aren't divided as finely, allowing for greater integration. A task that might have gone through five or six handoffs can be handled by a single group, which helps us control the interface risks. Giving individuals more weight can increase risk, but it can also reduce it. This is a balance issue that any organization faces as it develops. Late last year, our COO said that LandSpace will be a startup for a long time, and solving quality issues and reducing management complexity will remain an important task.

As commercial space becomes more visible to the public, has the talent you attract changed significantly?

Dong: Commercial companies tend to hire more experienced professionals, so the density of fresh graduates isn't as high as in the state system. A new graduate might benefit from training in a well-established system, but we also have development mechanisms. We just might not be as comprehensive as the "Shaolin Temple" of the state system. We prefer to put people directly into the field. For experienced hires, we've attracted people from other industries, like automotive. Traditional aerospace engineers might not be as cost-sensitive and assume a part should just be expensive. But someone from the automotive industry will tell us, "In the automotive industry, this part costs a few yuan, not the price you're paying."

Can you give an example?

Dong: Take a common temperature sensor for a rocket. It might cost tens or even over a hundred yuan. Someone from the automotive industry would say, "We use these for a few yuan each. Why is it so expensive?"

Doesn't aerospace require specific certification standards for all parts, which drives up costs?

Dong: We certify based on results. If we can control the outcome of a part, the verification process can be faster. I even believe this is the future trend for commercial aerospace. Some things are called "automotive-grade" on the ground. But if they've been to space and proven to work, I don't see why they can't be called "space-grade."

When did you start to believe that a private company building a reusable rocket was definitely going to succeed?

Dong: We believed China's reusable rocket would definitely succeed. From a technical logic standpoint, LandSpace had already gained experience with the propulsion system from the Zhuque-2, and the Zhuque-3's first launch was successful. The only remaining key technology was the 3.3 kilometers of the recovery landing. With the technical conditions at the time, we couldn't guarantee 100% success, but we couldn't identify any rigid technical obstacles remaining. But you don't need to overthink the final result. When it comes to business costs and outcomes, we defer to our boss, Zhang Changwu. If the boss is willing to take that gamble and has that vision, then we don't need to worry about it. Our job is to do the design well, get as close to the goal as possible within our limited resources, and create more opportunities. Our situation is nowhere near as difficult as Musk's back then. Even Armstrong said it couldn't be done. That's the significance of pioneers. They proved it was possible. Since he could do it, so can we.