The investment logic publicly expressed by Sun Yuchen is: find a track with a clear direction, lay out plans on both ends simultaneously, and do not bet on the execution power of a single company.

Written by: Lin Wanwan

Source: Zhang Sheng Beatz

Most people in China know Sun Yuchen from his bizarre stories.

Spending 30 million to invite Buffett to dinner, only to bail using a kidney stone at the last moment; spending 6.2 million dollars to buy a tape of a banana, and eating it in front of everyone at the press conference; investing 75 million dollars to become the largest financier of the Trump family's crypto project, taking a seat at the White House dinner; flying past the Kármán line at the age of 35, declaring himself the youngest Chinese commercial astronaut.

The negatives are numerous, including being sued by the SEC in 2023 for market manipulation, with allegations including over 600,000 wash trades to boost TRX prices, and hiring celebrities for promotions without disclosing compensation, currently involved in mutual lawsuits with the Trump family's project WLFI.

These stories have spread so widely that they almost overshadowed one serious matter. This man has hardly missed betting on any trend in the secondary capital market over the past decade.

Starting to purchase BTC in 2013, by 2016, he recommended people born in the 90s not to buy houses, but to buy:

Bitcoin, Nvidia, Tesla, Tencent.

Ten years later, as of May 2026, Tesla has total returns of about 2683%, and Nvidia has total returns of nearly 24000%.

If you had listened to Brother Sun back then, and invested 10,000 yuan in Nvidia, today it would be 2.4 million; 10,000 yuan in Tesla would also be 278,000. An audience member who had bet 200,000 yuan on each item from that 2016 list would see just the Nvidia bet turn into about 48 million, and the Tesla bet turn into about 5.4 million, totaling just over 53 million.

This man is still shooting forward today. On November 6, 2025, Sun Yuchen tossed out a statement:

"Short-term lack of chips, long-term lack of energy, permanently lack of storage."

The reaction of the capital market to this statement only began to heat up in 2026. SanDisk (SNDK), spun off from Western Digital, soared from around 35 dollars to 1439 dollars in one year, achieving a peak increase of nearly 50 times.

HBM memory is fully booked by the three major original manufacturers: Samsung, SK Hynix, and Micron, with all capacities sold out by early 2026, and orders already scheduled until 2027-2028.

While everyone else is still chasing after the storage concept, at the beginning of 2026, Sun Yuchen shifted his focus in a video.

This video originally discussed the outlook for 2026. In addition to writing health into the new year's list, which is rather wellness-oriented, he also dedicated a segment to the young people to focus on: embodied intelligence, drones, spatial computing, space exploration.

The author has collected Sun Yuchen's public statements on these four directions over the past two years, and when pieced together, each path has also produced early leaders of its capital.

Who is the next storage stock?

The first track Sun Yuchen pointed to is embodied intelligence.

The concept of robots has been discussed by humanity for at least a hundred years. The term "robot" was coined by Czech playwright Karel Čapek in 1920; industrial robotic arms have been in use since the 1980s, and Honda's ASIMO could climb stairs more than twenty years ago. However, the real bottleneck has always been in the brain.

In the past two years, the entire industry has turned to the VLA model: Vision - Language - Action. In layman's terms, previous robots worked based on code, while today's robots are starting to work by seeing the world.

Yushutech is set to ship over 5,500 humanoid robots in 2025, making them the world leader, and submitted its IPO application to the Science and Technology Innovation Board in March 2026. Galaxy General raised 300 million dollars (approximately 2.1 billion yuan) in a new round of financing in December 2025, bringing total financing to about 800 million dollars, with a valuation of 3 billion dollars (about 21.1 billion yuan), setting records for both single-round and cumulative financing in the field of embodied intelligence.

Sun Yuchen said he is unlikely to personally build robots, but he has a keen sense for the narrative and flow of funds in this direction. He once stated in an interview with Bloomberg, "In a market where 99% of people do not know what a wallet is, the cost of education must be included in the business model."

This statement, which explained stablecoins in 2018, remains applicable in 2026. 99% of Chinese people have not yet used an embodied intelligent robot, but as long as that robot can cook, move boxes, and assist the elderly, the remaining 1% represents the next opportunity.

His second identified track is drones.

While humanoid robots are still ramping up production, drones have already taken the lead in commercial deployment. Drones are naturally suited for tasks that AI excels at, from autonomous navigation to group collaboration and data collection. They do not need to walk; flying is easier than for humanoid robots.

On the battlefield of Ukraine, AI drone swarms have seized over half of the role previously held by tank units, with Ukraine's annual drone production target now climbing into the millions. Over Chinese rural rice paddies, DJI agricultural drones are flying, with one drone replacing ten agricultural workers. Meituan has successfully operationalized drone food delivery in Shenzhen, with order-to-delivery times under 15 minutes.

Drones are ahead of humanoid robots. They represent the first commercially viable form of physical AI.

The third track Sun Yuchen pointed out is spatial computing. This is the least mainstream among the directions he identified.

When Apple releases the Vision Pro in 2024, most people perceive it as an overpriced VR headset. This could be a misunderstanding.

The ambition of the Vision Pro has little to do with VR. It is Apple's first attempt to enable AI to understand space—how big your living room is, how far the table is from you, whether the coffee cup is on the left or right side of the sofa, and whether you can reach it by extending your hand. This task sounds simple but is ten times harder than training ChatGPT. Large language models only need to understand language; spatial computing requires understanding physics.

This just happens to be the common prerequisite for robots, drones, and autonomous driving—they all need a form of spatial intelligence. Nvidia's Cosmos platform, Google's Genie 3 world model, and Tesla's FSD are all working on the same goal: transitioning AI from understanding text to understanding the world.

ChatGPT is sufficient for understanding language, but the next generation of AI must understand the world itself.

In the first three tracks, Sun Yuchen only mentioned them verbally, but when it comes to space, he truly went there physically.

On August 3, 2025, he boarded Blue Origin's "New Shepard" NS-34 capsule and flew past the Kármán line.

After returning to the ground, he expressed an ambition to hope that his company would no longer just be a "cryptocurrency exchange," but a "service provider for space economic infrastructure," allowing blockchain to solve issues like asset rights verification in space, satellite data transactions, and interplanetary payments. It sounds like science fiction. But looking back at how he preached USDT ten years ago, one could argue that people felt the same way about it back then.

Back on Earth, he directed a message to young people more directly: "Space exploration is a shared mission for all humanity. I hope that through this flight, more young people will be inspired to engage in technology and innovation, collectively shaping the interstellar future of humanity."

Listen to Brother Sun's investment logic

Sun Yuchen's publicly expressed investment logic is: find a clearly defined track, lay out plans on both ends simultaneously, and do not bet on the execution power of a single company.

For the robot line, his framework is to separate bets on the body and the brain.

Tesla is betting on the body, announcing at the beginning of 2026 that it will cease production of Model S and Model X, transforming the Fremont factory into the Optimus production line, aiming for an annual output of one million units, with a unit price in mass production ranging between 20,000 and 25,000 dollars; the current version of Optimus has already been performing parts handling and sorting in the Austin and Fremont factories, and the Gen 3 production line will launch in the summer of 2026.

Nvidia is betting on the brain. Jetson Thor integrates server-level AI inference into the robot body, and Isaac GR00T has almost become the industry’s universal foundation; Huang Renxun announced at the GTC that there will be one billion humanoid robots globally by 2035.

Whether Optimus can deliver on schedule is a question for Musk, not for Nvidia; as long as the track runs smoothly, the toll fees will be collected.

For the drone line, the core judgment is the irreversibility of Physical AI in military scenarios.

AeroVironment's Switchblade loitering munition has become a hallmark weapon in Ukraine, with monthly production capacity ramped up from 40 to 500 units, with a target of 1200 units; Kratos's XQ-58 Valkyrie acts as a "loyal wingman" for the F-35, where, when a manned aircraft goes on a mission, the drone runs a flank, with a unit price being less than one-fifth of a fifth-generation fighter jet, increasing by 280% in 2025 and another 72% in 2026.

One makes tanks less feasible, and the other makes manned fighters redundant, with the logic of both ends complementing each other.

For the space line, Sun Yuchen purchased a flight seat with Blue Origin for 28 million dollars in 2021, donating that amount to the Blue Origin STEM charity fund, distributed to 19 non-profit organizations. On August 3, 2025, he completed suborbital flight on the New Shepard NS-34 mission.

In the public market, SpaceX submitted a confidential IPO draft to the SEC in April 2026, targeting a valuation of 1.75 trillion dollars, which would become the largest IPO in human history; Rocket Lab reported over 200 million dollars in revenue for Q1 2026, serving as the most direct alternative to SpaceX when it is unavailable.

Once SpaceX goes public, the pricing coordinates for the entire space sector will need to be rewritten.

You should listen to Brother Sun

Piecing together Sun Yuchen's statements over the past two years, "AI, robots, and blockchain have reached the iPhone moment" is his judgment on embodied intelligence. "Robot armies, robot police" is his prediction about autonomous weaponized AI. "The fusion of AI, robots, and spatial computing" is his bet on the next generation of human-machine interfaces. "The Earth is too small; it's our home" reflects his perspective shift after flying past the Kármán line.

These four things combined present the full picture of physical AI.

Over the past twenty years, the internet has changed the way information flows. WeChat has replaced writing letters, Taobao has replaced market gatherings, and Douyin has replaced television.

But the underlying rules of the physical world remain unchanged: workers are still workers, factories are still factories.

In the next twenty years, AI may change the operational methods of the real world itself. Factories may have humanoid robots that stand and work without needing rest, autonomous vehicles may flow along the roads, drone swarms may thunder on the battlefield, and the first "residents" landing on the moon and Mars are likely to be the early-arriving AI robots.

The young man who, in 2016, urged everyone not to buy houses has now flown past the Kármán line.

And most of us are still waiting for the next Yanjiao.