摘要：无人系统，是国家神经系统，不是工业爆款，在中国，无人机正在成为“出口明星”“流量热点”“科技名片”。全球每10架消费级无人机，就有7架是我们造的；一整年卖出去800多万架，全世界最多。但我要说：这一切只是表象。无人机不是做出来卖的，是打出来的。你可以靠价格赢下

Chen Long: Flying More Doesn’t Mean Stronger — China’s Drone Strategy Needs a Strategic Cognition Reset

无人系统，是国家神经系统，不是工业爆款，在中国，无人机正在成为“出口明星”“流量热点”“科技名片”。全球每10架消费级无人机，就有7架是我们造的；一整年卖出去800多万架，全世界最多。但我要说：这一切只是表象。无人机不是做出来卖的，是打出来的。你可以靠价格赢下市场，却不能靠组装赢下一场战争。

Unmanned systems are the neural network of a nation, not an industrial fad. In China, drones have become “export stars,” “online traffic magnets,” and “symbols of technology.” Seven out of every ten consumer drones globally are made by us; over 8 million units are sold annually—the highest in the world. But I must say: this is all just surface. Drones are not built to be sold, but to be fought with. You can win a market with pricing, but you can’t win a war with assembly.

01

中国无人机的问题

不在“造得慢”，而在“认知落后”

I. The Problem with Chinese Drones Is Not Slow Production, but Lagging Strategic Cognition

我们确实“飞得多”。但“战”与“算”，我们远远落后。绝大多数无人机停留在图像采集、航拍工具的层面；缺乏真正作战所需的AI决策能力、战场网络能力、干扰对抗能力；关键芯片、核心飞控系统、信号链路，受制于人，一旦断供、断网，就地瘫痪。

We do fly a lot. But in terms of “combat” and “computation,” we are far behind. Most drones remain at the level of image collection or aerial photography tools; they lack the AI decision-making, battlefield networking, and jamming resistance capabilities required for real combat. Key chips, core flight control systems, and communication links are reliant on others—once supply is cut or networks fail, the systems become paralyzed.

美军的“死神”“全球鹰”，不是飞机，而是战争节点。它们被深度集成在C4ISR系统中，是“看得见”“传得回”“打得准”的作战链路一部分。而我们的主流产品，哪怕外形再酷、飞行再稳，很多仍然只是“会飞的摄像头”——与现代战争要求的“自主智能节点”相距甚远。更严重的问题在于：整个行业还沉迷在“出货量”“航拍民用化”“性价比”这些工业思维中，尚未完成战略思维的升级。

The U.S. military’s “Reaper” and “Global Hawk” aren’t just aircraft—they are warfare nodes. They’re deeply integrated into the C4ISR system, forming a combat chain that can “see clearly,” “relay information,” and “strike accurately.” Our mainstream products, no matter how cool they look or how stable they fly, are still often just “flying cameras”—far from the “autonomous intelligent nodes” that modern warfare demands. Worse still, the entire industry remains fixated on concepts like “shipment volume,” “civilian use of aerial photography,” and “cost performance”—industrial thinking that has yet to evolve into strategic thinking.

02

GPS曾是死穴，现在我们有北斗

—— 但还不够

II. GPS Was Once Our Fatal Weakness—Now We Have BeiDou, But That’s Not Enough

1996年台海危机，美国人为关闭GPS信号，我们的导弹飞控失灵、空中指引全失，几乎造成系统级瘫痪。这件事震动国防战略体系，也催生了北斗计划。今天，北斗系统已全球组网，全面具备替代GPS的能力。它不仅能提供高精度定位、授时服务，更具备抗干扰、自主控制、安全可用的战略属性。但问题在于：我们很多无人机依然默认绑定GPS模块，甚至大量操作系统、SDK都封装在安卓体系内，无法真正实现信号独立、控制主权。脱离GPS只是第一步，未来更重要的能力，是多源融合导航：将北斗、视觉、惯导、地磁等多模态信号融合；抗干扰与欺骗识别能力：真正实现“作战级鲁棒性”；在没有卫星信号下的高维认知飞行：像昆虫一样具备自主路径重构能力。

During the 1996 Taiwan Strait crisis, the U.S. deliberately cut off GPS signals, causing our missile flight control to fail and aerial guidance to collapse—nearly triggering a system-wide paralysis. This event shocked our national defense strategy and led to the launch of the BeiDou program. Today, the BeiDou system is globally networked and fully capable of replacing GPS. It not only offers high-precision positioning and timing services but also provides strategic features such as anti-jamming, autonomous control, and secure availability. Yet the problem remains: many of our drones are still bound by default to GPS modules. Even worse, a large number of operating systems and SDKs are built on Android frameworks, making true signal independence and control sovereignty impossible. Decoupling from GPS is just the first step. The more critical future capabilities are: multi-source navigation fusion—integrating BeiDou, vision, inertial guidance, geomagnetism, and other modalities; anti-jamming and spoofing detection—achieving true combat-level robustness; and high-dimensional cognitive flight without satellite signals—drones that, like insects, can autonomously reconstruct their flight paths.

03

AI的关键不是会写诗

而是让机器能战场上推理决策

III. The Key to AI Is Not Writing Poetry—It's Battlefield Reasoning and Decision-Making

我们正在迎来一场算法革命。以DeepSeek为代表的大模型，不只是语言工具，而是未来无人系统的“作战大脑雏形”。传统无人机靠预设程序飞行，面对突发事件只能等待“人”下指令；但AI大模型具备：情境理解、任务自主重构、多目标评估与推理、自主博弈与学习能力、也就是说，未来不是“人告诉AI打哪里”，而是“AI告诉人应该怎么打”。我们应该立即推动如下战略转向：将大模型从“桌面推理”推进到“边缘部署”，部署在飞控芯片、地面站、前线节点；从“辅助导航”走向“全域战术指挥辅助”，不仅识别目标，更识别战场态势与风险优先级；构建“千机一脑”能力，让大模型同时协调数百架无人系统，各司其职，协同演化。

We are entering an algorithmic revolution. Large models like DeepSeek are not just language tools—they are prototypes of future combat brains for unmanned systems. Traditional drones fly based on preset programs and can only wait for human commands in emergencies. But AI large models have the ability for contextual understanding, autonomous mission reconfiguration, multi-objective evaluation and reasoning, autonomous gaming, and learning. In other words, the future is not about “humans telling AI where to strike,” but “AI telling humans how to fight.” We must immediately promote the following strategic shifts: move large models from “desktop reasoning” to “edge deployment,” placing them in flight control chips, ground stations, and frontline nodes; shift from “assisted navigation” to “global tactical decision support,” where AI not only identifies targets but also evaluates battlefield dynamics and prioritizes risks; and build a “one brain for a thousand drones” architecture, enabling large models to coordinate hundreds of unmanned systems, each with defined roles in adaptive collaboration.

04

未来的战争，不比谁的无人机飞得快

而比谁的系统协同得更快

IV. Future Wars Won’t Be About Who Flies Faster, But Who Coordinates Smarter

无人机不是装备本身，而是作战网络的节点。未来战场的关键不是“某架无人机的性能”，而是：谁能在 一分钟内激活1000架无人智能体，谁能让它们在没有GPS、没有人指令、在强干扰环境下自动构建任务网络，谁能实现“空中战场瞬时协同”，组成真正意义的智能作战云。我们要构建的，是一个跨越平台、跨越场景、可自我演化的空天智能神经系统。

Drones are not stand-alone weapons—they are nodes in a combat network. The key to future battlefields won’t be the performance of a single drone, but rather: who can activate 1,000 autonomous drones in under a minute, who can enable them to build mission networks autonomously in GPS-denied, communication-disrupted environments, and who can achieve real-time aerial battlefield coordination to form a truly intelligent combat cloud. What we need to build is a space-air-ground intelligent neural system that spans platforms, adapts across scenarios, and evolves on its own.

05

建议：

国家顶层要重定位，企业战略要重构

V. Recommendations: National Strategy Must Be Repositioned, Enterprise Strategy Must Be Rebuilt

国家层面：将“智能无人系统”纳入国家战略科技力量核心体系，摆脱制造业附属定位；设立“空地智能协同”专项攻关基金，重点投入北斗算法、飞控OS、AI模型国产化；支持西部高原、沙漠、东南海岸等地设立无人系统靶场，允许“战场演化式测试”；推动无人系统出口模式升级，从“卖产品”走向“输出算法、接口标准与能力体系”。

At the national level: Integrate “intelligent unmanned systems” into the core of national strategic science and technology forces, breaking free from the status of manufacturing sub-sector; establish a special R&D fund for “air-ground intelligent coordination,” focusing investment on BeiDou algorithms, domestic flight control OS, and AI model localization; support the setup of unmanned system testing ranges in western plateaus, deserts, and southeastern coastlines to allow “battlefield-evolution style testing”; promote an upgrade in export models from selling products to exporting algorithms, interface standards, and full capability systems.

企业层面：从制造商转型为“系统平台构建者”；深度绑定算法团队，参与模型训练与战术调优；投入开发“模块化+异构+自组织”型蜂群作战框架；构建从飞控、算力、导航、接口到集群智能的一整套体系链条。

At the enterprise level: Shift from manufacturer to “system platform architect”; form deep partnerships with algorithm teams, participating in model training and tactical optimization; invest in developing modular + heterogeneous + self-organizing swarm combat frameworks; build an integrated chain of flight control, computing power, navigation, interfaces, and swarm intelligence.

结语：中国无人机真正的价值，不在于飞多快、卖多少，而在于能不能构建一整套属于自己的“低空国家神经系统”。未来战争的比拼，将不再是坦克、战机的正面对抗，而是智能节点之间的系统级缠斗。谁掌握算力、网络、协同体系，谁就掌握胜负。我们不能再自我陶醉于航拍爆款、工业代工。我们必须醒来，走向系统自主、智能突围、战场神经网络化的新时代。

Conclusion: The true value of China’s drones does not lie in how fast they fly or how many are sold, but in whether we can build a complete, self-owned “low-altitude national neural system.” Future warfare won’t be about tanks or fighter jets engaging head-on, but about system-level entanglement among intelligent nodes. Whoever controls computing power, networks, and coordination systems will control the outcome. We can no longer indulge in popular aerial photography or industrial subcontracting. We must awaken—and step into a new era of system autonomy, intelligent breakthrough, and battlefield neuro-networking.

——陈龙
优选空天科工集团有限公司 董事长
北京襄阳企业商会 会长

— Chen Long
Chairman, Optimal Aerospace Science and Industry Group Co. Ltd.
President, Beijing-Xiangyang Chamber of Commerce.

来源：优选空天科工OASI