fictionpage Chapter 78 The Wall Upstream
Chapter 78 The Wall Upstream
Late October 2019.
The hype surrounding "Why Flight Controllers Aren't We Buying?" has finally died down. Industry media's next focus has shifted to topics related to 5G infrastructure.
Wu Zheng officially started working at Hongyuan on Monday. He is the Chief Flight Control Architect. A printed copy of his appointment letter sits on his desk in his new office.
Along with Wu Zheng came six engineers from the old Tianying team who had passed Hongyuan's normal interview process. Seven out of twelve people came—Zhou Ming had mentioned this result in an email to Su Chen. He originally thought three or four people would stay, but he didn't expect Meituan to leave so early after completing the UI development test.
The flight control team, operating at full capacity, swelled from twenty-five to thirty-two people.
Su Chen was not indulging in celebration. After returning to his office following the first weekly flight control architecture meeting, he immediately archived all the documents related to the Skyhawk incident.
He placed a blank sheet of paper on his desk and began to write a line.
The flight control system was in place. After Wu Zheng arrived, a group of talents were injected into the core architecture of the flight control system.
The algorithms are also in place. Liu Yu's obstacle avoidance and Zhang Lei's inertial navigation fusion are both industry-leading.
Su Chen paused at the sensor layer section.
He discovered that there was a question he had been avoiding sitting on his desk.
All of Hongyuan's products, including IMU inertial measurement units, barometers, magnetometers, and vision sensors, are sourced from external suppliers.
There are four main suppliers: Bosch's automotive-grade IMU, Intersil's barometer, Murata's magnetometer, and Sony's CMOS image sensor. These are four semiconductor and component giants, three of which are foreign, and one is listed in Hong Kong but actually headquartered overseas.
This isn't just a problem for Hongyuan. DJI is in the same boat. The entire domestic consumer and industrial drone industry basically uses IMUs supplied by these four companies.
It hasn't been a problem for the past three years. Prices have been stable, performance has been consistent, and inventory has been ample. Hongyuan Sensors' gross profit margin has been steadily increasing with the expansion of its procurement volume.
But Su Chen has memories of his past life in his mind.
He knew what those imported components would look like ten years from now. He knew that Bosch would once include drone industry customers on its list of customers subject to proactive price reductions during a certain adjustment. He knew that Intersilicon had suddenly tightened its after-sales technical support in a particular year. He knew what kind of behavior Murata would exhibit during periods of international tension.
He knew even better that the truly alarming capabilities of these four suppliers were not just "batch order" and "extension of delivery time"—they could stop supplying you overnight.
He got up and walked to the window.
Technological friction between China and the US began in July 2019. Huawei was added to the Entity List in May. ZTE's orders were delayed. A design bureau in East China was suddenly named last month for using US software.
The wave of remote communication has already reached the doorstep of the drone industry. It's just a matter of time before a solution is found.
Back at his desk, Su Chen drew a circle on a piece of white paper. Inside the circle were the words "sensor." Outside the circle was a large character: wall.
This wall isn't going to collapse right now. But if it does, all of Hongyuan's production lines will be crushed. F4's production capacity, the delivery of the SDK alliance, the finalization of formal industrial-grade contracts—everything depends on whether these imported chips can arrive on time.
He took out an industry research report from the beginning of the month from his drawer.
Analysis of the supply chain of core components for unmanned aerial vehicles (UAVs).
The report was sent to him by Lin Zhiqiang. The purchasing department of Nanwang Smart Energy conducted a supply chain risk assessment for the establishment of files in Ge'en'ai District, which clearly mentioned that: given that all four main suppliers currently rely on foreign countries, it is recommended that Hongyuan address the "single point of failure" in advance.
Southern Power Grid saw it. There's no reason Su Chen couldn't see it.
He simply hadn't made a decision after seeing it. Because making a decision is difficult. He could either develop it himself, find a domestic alternative, or invest in it. None of these three paths were cheap.
Self-developed sensors are a core technology, completely different from flight control. IMU's MEMS (Micro-Electro-Mechanical Systems) process requires more than just algorithms; it demands production lines, manufacturing capabilities, and data processing. The necessary funding is a product-level competitive advantage, not just windfall gains.
Looking for domestic alternatives—we've already checked. One in Jinan, one in Wuxi, and two in Shanghai. None of these four domestic manufacturers can match Bosch's levels in magnetic stability, temperature drift stability, and batch consistency. Replacing them with these would mean the F4's flight quality would drop back to F3 levels.
Of the four domestic companies that invested, two are willing to disclose their investment information. However, Qian Hongyuan, who invested, did not. The 30 million yuan raised in the recently completed Series B round is already sufficient for expanding the industrial-grade team and adapting to various scenarios.
Su Chen drew another circle on the white paper. Inside it, he wrote "PCBA OEM".
This was the fourth option he had considered.
Hongyuan itself does not manufacture flight control modules. The PCBA printed circuit boards and mounting hardware for the modules are outsourced to a medium-sized contract manufacturer in Shenzhen, Likou. Contract manufacturers have been struggling in the past two years—their premium consumer electronics contract business is shrinking.
If Hongyuan Holdings were to become a contract manufacturer, there would be two main advantages. First, it would keep the assembly of sensor modules firmly in its own hands—meaning that if a supplier needed to be switched in the future, it wouldn't have to wait for the contract manufacturer to reschedule production. Second, with the core modules on its own production line, the feedback speed of verification data would be more than twice as fast.
This doesn't solve the fundamental problem of the sensor wall. But it can give us a more proactive stance when the sensors are first deployed.
Su Chen put the blank paper aside. He made a phone call to Zhou Ming.
"Old Zhou, can you help me find out something? Are there any contract manufacturers in Shenzhen and the surrounding areas that make drone-grade PCBAs, with a scale of 300 to 800 employees, preferably with SMT production lines and one or two automated testing lines? I need to see a list tonight."
What sieve diameter?
"We would consider either acquiring a controlling stake or establishing a deep strategic partnership."
Zhou Ming paused for a moment.
"Are you really going to take control of one company?"
"Make a list first. We'll decide after I've reviewed the list."
Su Chen hung up the phone.
He wrote a date in the upper left corner of the white paper—July 23, 2019. Then, in the lower right corner, he wrote: "The wall hasn't collapsed yet. But I need to start preparing the ladder."