作者：罗翔

中国这些年最喜欢讲“弯道超车”。但如果把产业发展的路径扒开来看，会发现一个尴尬事实：很多所谓的自主崛起，并不是凭空从零开始，而是在全球产业链转移、国际订单和西方企业严苛标准中，一点一点“喂”出来的。为什么中国电子产业发展得这么快？一个重要原因，就是长期承接惠普、戴尔、苹果等跨国企业的制造订单。外企把产品要求、质量标准和验收规则带进来，中国供应商才真正知道：一块铜箔、一颗螺丝、一个连接器不是“能做出来”就算合格，而是必须满足尺寸、公差、可靠性、寿命和一致性等一整套指标。这其实是一种非常现实的工业教育。企业不是看几份图纸就突然拥有工业能力的。真正的能力，是在一次次订单、失败、返工和退货中磨出来的。客户告诉你哪里不合格，供应商再改；客户提出更高要求，供应商再升级。久而久之，一条成熟的产业链就这样形成了。汽车工业同样如此。几十年来的国际合作带来了零部件体系、生产管理、质量控制和全球供应链标准。中国企业从“照着造”，逐渐发展到“自己造”，再到“自己设计”。今天中国汽车产业拥有庞大的供应链体系，与这些长期积累当然脱不开关系。航空工业也是类似逻辑。中国企业参与波音、空客等国际航空产业链，在机翼、机身结构件、起落架等领域接受国际标准训练。真正重要的，不只是学会制造一个零件，而是学会什么叫航空级质量控制、材料一致性、结构可靠性和适航标准。所以，后来中国能够推出自己的民用飞机，并不意味着过去几十年的国际产业链经验消失了，恰恰相反，这些积累正是后来发展的基础。至于运-10，能造出一架飞机，与能不能建立完整的民航工业体系，是两回事。飞机能飞起来，不等于拥有成熟的民航产业。问题到了航天领域，就变得更加有意思了。因为航天恰恰是一个很难依靠西方产业转移完成工业启蒙的领域。苹果可以给你订单，波音可以给你零部件制造订单，但火箭呢？核心技术、关键材料、发动机、控制系统、飞行软件和测试体系，并不存在一个成熟的西方企业把完整订单体系搬过来，告诉你：“按照这个标准生产。”于是，一个现实问题出现了：即使你知道可回收火箭应该怎么飞，也未必知道组成它的成千上万个零部件，怎样稳定、可靠、批量地制造出来。这就是“知道技术路线”和“真正拥有产业能力”之间的鸿沟。马斯克和SpaceX已经用长期试验验证了可重复使用火箭的路线：垂直起飞、发动机深度节流、姿态控制、返回制导、垂直着陆和重复使用。既然路线已经有人走通，后来者照着走，为什么不能明显更快？因为工程不是电子游戏，更不是把别人的攻略抄下来就能通关。设计图不会自动生产发动机，参数不会自动培养供应商，一次试验成功也不会自动变成稳定量产。真正困难的地方，往往藏在不起眼的细节里：高温合金能否稳定生产？密封件能否经受极端温度和压力？阀门能否保持数千次动作的一致性？隔热材料批次之间会不会出现差异？焊缝有没有微小缺陷？传感器在剧烈振动下还能不能正常工作？这些问题单独看都不起眼，但叠加起来，就是火箭能不能安全返回的生死线。所以，真正决定可回收火箭研发进度的的不只是总设计师手里的图纸，更是整个供应链背后的工业基础。所谓后发优势也只有在产业基础足够成熟时才真正存在。如果别人已经把题目做完，而你连每一种材料、每一个零件、每一套测试设备都需要重新摸索，那么你拿到的不是答案，只是别人写在黑板上的一句话：“这道题可以这样解。”剩下的步骤，仍然要自己摔无数次跤。更讽刺的是我们常把“产业链完整”等同于“工业能力全面领先”。可真正的工业能力，不是工厂多、论文多、宣传片宏伟，而是面对苛刻要求时能不能稳定制造；零件出问题时能不能找到原因；连续生产时能不能保持一致；系统失败后能不能把失败变成下一代产品的经验。别人用了二十年烧钱、爆炸、失败、复盘，才把路线跑通。后来者可以少走弯路，却不能跳过所有试错，更不能跳过供应链成熟、人才积累、测试体系和质量文化的形成。于是熟悉的口号又响起来：“弯道超车，遥遥领先！”只是火箭升空之前，工业能力也得先升空。宣传口号可以一夜之间遥遥领先，供应链却只能用十年、二十年，一颗螺丝、一只阀门、一块材料地把“领先”拧出来。火箭有一个很讨厌的特点：它不看宣传稿，也不看热搜，更不会因为“全球领先”四个字就改变物理规律。发动机不会因为喊“自主可控”就自动提高寿命，阀门不会因为“遥遥领先”就自动完成可靠性验证。那些藏在火箭内部、没人拍照、没人宣传的零件，也不会因为一场盛大的发射仪式就突然拥有几十年的工业积累。它们只认一个东西：你到底能不能把它稳定地造出来。因为弯道可以画在PPT里，领先可以写在标题里，唯独火箭落地以后，还得看看它还能不能再飞一次。当一个国家越来越擅长宣布“遥遥领先”，却仍然需要用十年、二十年去补足一颗螺丝、一个阀门、一种材料背后所需的工业积累时，我们究竟是在实现真正的弯道超车，还是只是在不断制造一种“已经领先”的感觉？

编辑：周志刚  校对：王滨 翻译：周敏

The Eastern Power Overtakes on the Bends and Is Far Ahead, So Why Can’t It Manage Reusable Rockets?

Author: Luo Xiang

Abstract: The Communist Party of China has long publicized that Chinese industry is far ahead, but in reality, mature technologies in civilian industry mainly rely on transfers from Western enterprises. China lacks core technologies in the aerospace industry field.

In recent years, China’s favorite slogan has been “overtaking on the bends.” But if you examine the process of industrial development, you will discover an embarrassing fact: many so-called independent rises did not start out of thin air from zero, but were “fed” bit by bit amid global supply chain transfers, international orders, and the rigorous standards of Western enterprises.

Why did China’s electronics industry develop so fast? An important reason is the long-term undertaking of manufacturing orders from multinational enterprises such as HP, Dell, and Apple. Foreign companies brought in product requirements, quality standards, and acceptance rules, and only then did Chinese suppliers truly learn: a piece of copper foil, a screw, or a connector is not considered qualified merely because it “can be made,” but must satisfy a whole set of metrics including dimensions, tolerances, reliability, lifespan, and consistency. This is actually a very realistic industrial education.

Enterprises do not suddenly possess industrial capabilities just by looking at a few blueprints. True capability is ground out through order after order, failure after failure, rework, and returned goods. The customer tells you where it is unqualified, and the supplier modifies it; the customer raises higher requirements, and the supplier upgrades again. Over time, a mature supply chain is formed in this manner.

The automotive industry follows the same logic. Decades of international cooperation brought component systems, production management, quality control, and global supply chain standards. Chinese enterprises gradually developed from “manufacturing by copying” to “manufacturing by themselves,” and then to “designing by themselves.” That China’s automotive industry today possesses a huge supply chain system is, of course, inseparable from these long-term accumulations.

The aviation industry follows a similar logic. Chinese enterprises participated in international aviation supply chains such as Boeing and Airbus, receiving international standard training in areas like wings, fuselage structural components, and landing gear. What is truly important is not just learning how to manufacture a part, but learning what is meant by aviation-grade quality control, material consistency, structural reliability, and airworthiness standards.

Therefore, that China was later able to launch its own civilian aircraft does not mean the international supply chain experience of the past few decades vanished; quite the contrary, these accumulations were precisely the foundation for later development. As for the Y-10, being able to build one airplane and being able to establish a complete civil aviation industrial system are two different things. An airplane being able to fly up does not equal possessing a mature civil aviation industry.

When it comes to the field of aerospace, the problem becomes even more interesting. Because aerospace happens to be a field where it is very difficult to complete industrial enlightenment by relying on Western industrial transfer. Apple can give you orders, Boeing can give you parts, but what about rockets? Core technologies, key materials, engines, control systems, flight software, and testing systems—there does not exist a mature Western enterprise that moves a complete order system over and tells you: “Produce according to this standard.”

Thus, a realistic question emerges: even if you know how a reusable rocket should fly, you might not know how the tens of thousands of components that make it up can be manufactured stably, reliably, and in mass quantities. This is the chasm between “knowing the technical route” and “truly possessing industrial capability.”

Musk and SpaceX have already verified the route of reusable rockets through long-term testing: vertical takeoff, deep engine throttling, attitude control, return guidance, vertical landing, and reuse. Since someone has already walked the route through, why can’t latecomers follow it and be noticeably faster?

Because engineering is not a video game, let alone clearing a level by copying someone else’s walkthrough. Blueprints do not automatically produce engines, parameters do not automatically cultivate suppliers, and a single successful test does not automatically turn into stable mass production.

Where it is truly difficult is often hidden in inconspicuous details: Can high-temperature alloys be produced stably? Can seals withstand extreme temperatures and pressures? Can valves maintain consistency over thousands of actions? Will differences appear between batches of insulation materials? Are there tiny defects in the weld seams? Can sensors still function normally under violent vibration? These questions look inconspicuous individually, but stacked together, they form the line of life and death regarding whether a rocket can return safely.

Therefore, what truly determines the speed of reusable rockets is not merely the blueprints in the hands of the chief designer, but the industrial foundation behind the entire supply chain. The so-called latecomer advantage truly exists only when the industrial foundation is sufficiently mature.

If others have already finished the exam questions, while you still need to explore anew every single material, every single part, and every single set of test equipment, then what you have obtained is not the answer, but merely a sentence written on the blackboard by others: “This question can be solved this way.” The remaining steps still require you to stumble and fall countless times yourself.

Even more ironically, we often equate a “complete supply chain” with “comprehensively leading industrial capabilities.” But true industrial capability is not about having many factories, many papers, or magnificent promotional videos, but whether you can manufacture stably when facing rigorous requirements; whether you can find the cause when a part encounters a problem; whether you can maintain consistency during continuous production; and whether you can turn failure into experience for the next generation of products after a system failure.

Others spent twenty years funding development, enduring explosions and failures, and reviewing what went wrong, and conducting post-mortems before running the route through. Latecomers can take fewer detours, but cannot skip all trial and error, nor can they skip the maturation of the supply chain, the accumulation of talent, and the formation of testing systems and quality culture. So the familiar slogan resounds again: “Overtake on the bends, far ahead!”

It’s just that before the rocket ascends into the sky, industrial capability must ascend first. Promotional slogans can become far ahead overnight, but the supply chain can only build that lead, one screw, one valve, and one material at a time with one screw, one valve, and one piece of material over ten or twenty years.

Rockets have a very annoying characteristic: they do not read publicity drafts, nor do they look at trending searches, and even less will they change physical laws because of the four words “globally leading.” Engines will not automatically increase their lifespan because you shout “independently controllable,” and valves will not automatically complete reliability verification because of “far ahead.” Those parts hidden inside the rocket that no one photographs and no one publicizes will not suddenly possess decades of industrial accumulation because of a grand launch ceremony either. They recognize only one thing: whether you can actually manufacture it stably.

Because bends can be drawn in PPTs, and being ahead can be written in titles, but only after the rocket lands must one see whether it can fly once more.

When a country becomes increasingly adept at announcing “far ahead,” yet still needs to use ten or twenty years to make up the industrial accumulation of a screw, a valve, and a material, are we achieving a true overtake on the bends, or are we merely continuously manufacturing a feeling of “already leading”?

Editor: Zhou Zhigang    Proofreader: Wang Bin    Translator: Zhou Min