Confronting DJI head-on with drones, but victory is not the goal
Consumer drones have nearly all failed. Knowing this, they still entered the market—not to grab share, but because making drones could upgrade the company's technology, manufacturing, and supply-chain capabilities in one move, helping it survive the next cycle.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
A father's tears weighs heavier than institutional punishment
In college, Liu Jingkang hacked into the school email system to retrieve exam papers and prove his technical prowess, then publicized it to boast. He was hit with a disciplinary record. But what really revealed the weight of his actions wasn't the punishment—it was seeing his father, whom he'd never seen cry, weep in front of the teacher and apologize. That moment showed him the actual damage to the school's reputation and made him feel, for the first time, that a parent's feedback carries more weight than any external judgment.
— Liu JingkangHis deepest regret: the junior classmates he talked into dropping out
Liu admits that when his first startup failed, he wasn't really a loser himself—his family had resources, and he could take what he'd learned into a good job at a major company. What genuinely haunts him is the few dozen junior classmates he convinced to join. They gave up offers from big tech companies worth multiples of what he was offering, and they forfeited structured mentorship—they wagered their most golden professional years, right after graduation, and the company failed anyway.
— Liu JingkangMeta's billions on VR helped the industry and hurt it equally
Luo Yonghao makes a counterintuitive observation: Meta loses money every year selling headsets. While that's elevated VR experience by generations on the surface, the fact that it's backed by a tech giant willing to burn unlimited cash means real hardware companies won't enter to compete. VR ends up with only one player exploring, burning money alone. Liu agrees and adds that Meta also mistimed its strategy: it painted grand metaverse visions without solidifying satisfaction in specific use cases. Users felt let down, and the category spiraled into low engagement.
— Luo Yonghao / Liu JingkangAR glasses need a decade to become truly consumer-ready
On his second venture building AR glasses, after surveying 60–70 supply-chain factories Liu was initially optimistic that five years would do. The first engineering prototype revealed gaps far wider than expected: temples had to be light enough not to look ridiculous, battery had to last all day, the device had to stay always-on, and it had to sit on skin without generating heat or sweat. Even scaling these parameters by hundreds or thousands of percent falls short. His revised conclusion: this probably takes 10 to 20 years, not five.
— Liu JingkangRushing hardware to market compounds the damage that follows
The modular-lens camera was rushed to market before CES. The team knew more test time would surface reliability issues, but they shipped early anyway. A pandemic manufacturing halt compounded things, and the product arrived with connection instability and fogging problems. Liu's lesson: don't rush hardware. Software bugs can be patched over-the-air; hardware defects cannot. Once you sell it, even a full recall cannot undo the damage in users' minds.
— Liu JingkangChallenging GoPro isn't about winning—it's about growing the whole market
The company never aimed to kill GoPro. Liu's view: three product lines—spherical action camera, thumb camera, wide-angle action camera—each draw their own circle. Each overlaps with GoPro's in some scenarios and can substitute in those moments. But the three circles combined are much larger than GoPro's single circle, because each use case expands the category itself. Revenue surpassing GoPro was the natural result of expanding the pie, not the starting goal.
— Liu JingkangPatent warfare, stripped bare, resembles organized crime
When an NPE demanded an outrageous settlement, Liu's company had no experience and thought they couldn't escape. Then a major supplier approached them: if they kept orders up next year, the supplier could ‘internally apply’ and use the power of their patent consortium to have a word with the opposing lawyers. Weeks later, the plaintiff withdrew. Liu describes the feeling as pure organized crime—patent consortiums collect membership fees in exchange for ‘the boss having your back.’ It's a real, legal, compliant side of modern enterprise operations.
— Liu JingkangBuilding drones means betting on latent, unstated demand
An internal executive once asked: if DJI committed not to make spherical cameras, could the company commit not to make drones? After discussion, the answer was still yes to drones. Liu breaks the logic into three layers. First, the mission—capturing and sharing life—requires a freely-moving camera. Second, demand-wise, DJI offers ‘a better horse,’ but what users really need is to get from A to B; drones' tactile control and free perspective solve that more fundamentally. Third and deepest: building drones forces the company to upgrade its entire tech, manufacturing, supply chain, and channel capabilities in one shot. That's the investment that keeps an enterprise ready for the next cycle.
— Liu JingkangIn their own words · checked verbatim
But seeing my father—because I'd never seen him cry—that was actually much bigger feedback. That's when I truly understood I'd really gone too far.
但其实看到我父亲 因为从来没见过他哭 那个事情 其实是更大的负反馈 就那个时候 就真的知道自己 这次真的是做过了
Liu Jingkang1:04:42
I think the only thing I truly feel sorry about is those junior classmates who followed me, and I genuinely fooled them.
我觉得真的唯一对不起的事情,就是那些家务的这些学弟学妹,那个就真的蠢忽悠了。
Liu Jingkang1:16:55
I think Meta's contribution to the VR industry—I'd say it's half good, half bad.
我觉得Facebook 就或Meta 对VR行业的贡献 我觉得只能说 功过都有 一半一半
Liu Jingkang1:57:41
First, when you wear this thing, it can't look like a mad scientist or mentally ill—or it'll never become a consumer product.
首先这东西戴着 不能像个科学怪人 像个精神病 要不然就永远不会成为消费类产品
Liu Jingkang2:08:47
So the lesson for the company at that time was: don't rush your product to the market. Especially not with hardware.
所以那个时候就大概对公司的教训 就是说就不要Rush your product to the market 对 尤其是硬件
Liu Jingkang2:14:49
This is GoPro. We have three different products that can encircle it. But looking at the big picture, there's a part that replaces it, but those three circles together are much bigger than GoPro's circle.
这是GoPro 我们有三个不同的产品 这是它可以围攻它 但是本上整个图看上去 是有一部分替换了它 但是三个圈加上它加起来 要比它圆的圈要大很多
Liu Jingkang2:40:05
It turns out that running a modern enterprise legally and compliantly, some things look just like organized crime. Some patent consortiums—you just pay membership dues. Once you pay the membership fee, you have the boss's backing.
原来合法合规的运营现代化的企业 有些事也跟黑帮差不多 他还种就是有些专利联盟你就交 你就交入会费 你交了入会费之后的话 你就有老大罩着你
Liu Jingkang2:52:18
If the only thing in the market right now were carriages, and you asked a customer what kind of better carriage they need, they'd say, 'Oh, I hope my horse doesn't need to sleep or eat and can just keep running all the time.'
假如我们现在市场上只有马车 你问客户为什么 你需要个什么更好的马车 那客户就会说 哎呦希望我的马 就不睡觉 不吃饭 也可以一直跑
Liu Jingkang3:13:34
Figures
| Liu's high school IQ test | approximately 140 | 20:16 |
| First startup seed capital | 150,000 yuan personal + 150,000 yuan borrowed from family | 1:19:02 |
| IDG Series A funding | 4 million RMB | 1:19:02 |
| GoPro founder IPO year cash-out | $750 million | 2:42:08 |
| Mi Band launch price vs. market price at the time | 79 yuan vs. 300–400 yuan | 1:48:31 |
| Huawei team identified potential reduction | 70–80 screws | 1:46:24 |
Glossary
- 337 Investigation
- The U.S. International Trade Commission's expedited procedure to determine whether imported products infringe patents; can directly ban product imports from sale.
- NPE
- Non-practicing entity; an organization that acquires patents but does not manufacture products, and files infringement lawsuits for profit.
- PMF
- Product-market fit; the state in which a product meets real market demand and has been validated by users.
- DFM
- Design for manufacturability; product design optimized for ease of mass production and cost reduction.
How to listen
Hardware and consumer electronics founders, supply-chain directors, and anyone trying to understand how companies survive cycles by building new capabilities.
The first 40 minutes of childhood background can be skipped.