You once bought NVIDIA and sold it after a 30% gain, feeling a bit clever.
Then it surged another 120%, and you stared at the candlestick chart for five minutes, growing more and more frustrated.
This is actually the typical behavior of most retail investors.
In 2022, someone did the complete opposite of the general public. Instead of buying NVIDIA, he invested in a company that supplies NVIDIA, a little-known, $700 million market cap company where 90% of people have never heard of, $AXTI. The stock was priced at $12 at the time, but later surged to over $70.
This individual is called Serenity, and he has become a sensation in the investment community this year. Out of the 35 positions he publicly disclosed, 31 saw gains, with a return of 22,500%, even Bloomberg and Reuters are now covering his tweets.
After going through dozens of his posts, I discovered that what he did was completely different from the general public buying NVIDIA.
What did the general public do when buying NVIDIA? Looked at the PE ratio, examined the financial report growth, heard news about the AI boom, saw the foreign capital inflow, and then placed an order. I could do this blindfolded.
What did Serenity do before buying $AXTI?
He started from NVIDIA's GPU, drawing an industry chain map. To run a GPU, you need a data center, which requires optical modules. The core component in optical modules is a laser, and the raw material for the laser is called indium phosphide. Then he did something I had never even thought of—he checked the global production capacity distribution of indium phosphide.
Globally, there are two companies capable of large-scale production of indium phosphide substrates. $AXTI holds a share of one-fourth to one-third.
In other words, as long as AI chips continue to be manufactured in the future, all optical module manufacturers will have to source raw materials from these two companies. And this structure will not change in the short term because from factory construction to mass production, the cycle is annual.
Then he also flipped through the company's patent documents, customer list, production capacity cap, and upstream sources. After checking everything, he finally placed the order.
Perilla Leaf Theory
He gave this set of tactics a name called Perilla Leaf Theory.
He said, when you go to a high-end sushi restaurant, everyone is focusing on that piece of fatty tuna belly. But the ingredient that the kitchen can absolutely not run out of is perilla leaf. Without the fatty tuna, the menu will lose several dishes, but if there's no perilla leaf, the whole restaurant can close down.
In the AI industry chain, NVIDIA, Microsoft, and OpenAI are the fatty tuna belly. The perilla leaf is a material name that is hard to come by, a low-profile company with a market value of over ten billion that no analyst covers, a part of the supply chain with only two global suppliers, and if it's missing, the entire chain will come to a halt.
Breaking down this theory, it's actually three steps.
Step One: Start from the top-level demand and ask layer by layer
AI boom → GPU demand surges → GPU manufacturing requires lithography machines → The core component of lithography machines is the lens → Who globally manufactures the lens? It's Zeiss, the only one. Keep going. Which small Japanese factory possibly supplies the special glass for Zeiss's lenses?
Ask the same question at each level: This level needs to function, what is the irreplaceable item in the next level?
Most people might stop at the second level, starting to discuss whether NVIDIA's PE ratio is high. But Serenity breaks it down to the fifth or sixth layer, tracing it until reaching the smallest market cap, most unfamiliar company.
Step Two: Count the players in this company's segment globally
Three or more players, pass! Because of sufficient competition, there is no pricing power.
Two players, watch closely.
One player or a de facto monopoly, that's it.
The logic is simple. The more AI expands, the more money flows upstream along the industry chain. As the tide rises, all ships rise with it.
But if there is only one ship in a certain segment, not only does it rise with the tide, but it can also turn around and threaten downstream, saying, "You can only use me, I call the shots."
Step Three: Before Buying, Share Your Analysis and Wait for Criticism
No need to seek approval; we are going against the grain, specifically inviting knowledgeable individuals to provide counterarguments.
Some pointed out when a logical step was taken too quickly, he went back to reevaluate. If someone mentioned a missing supplier, he went back to fill in the gap. Only when all loopholes were closed and no one could criticize anymore, did he place the order.
As he himself once said, ChatGPT will not challenge you. If you feed your analysis to AI, it will always agree with you. So, you must present your findings to a real person.
$SIVE: Second Validation of the Same Method
He used this method more than once.
In 2025, he heavily invested in $SIVE, a Swedish semiconductor company specializing in lasers. With a market capitalization of over a dozen billion U.S. dollars, priced in Swedish Krona, and hardly covered by U.S. analysts.
Why did he target it? Because the next-generation optical communication architecture for data centers, called CPO, has a physical limitation that silicon, a material, cannot emit light. If silicon cannot emit light, how can data be transmitted? It must use an external laser module. $SIVE produces high-power continuous wave lasers, serving as the external light source for CPO.
He once again followed that chain: NVIDIA GPU → Data Center Expansion → CPO Optical Interconnect → Surge in External Light Source Demand → $SIVE is one of the few companies globally capable of supplying.
After buying in, this stock nearly increased twentyfold.
RPI: Wall Street Overlooked New Demand
Turning to his February 2026 pick of the Raspberry Pi.
RPI, a UK company, produces inexpensive single-board computers, selling for $35 each, used by children to learn programming. Wall Street analysts unanimously expected a 14% annual revenue growth.
He presented a different figure: 55%.
How did he arrive at this? He delved into the developer community. He found that a large number of AI developers on GitHub were starting to deploy AI Agents using Raspberry Pi, with the related repositories' growth curve almost vertical.
He analyzed the purchasing discussions on various forums, developer growth trends, and then worked backward. The Wall Street revenue model completely failed to take this new demand into account, missing at least 40 percentage points.
Within two days of the tweet, RPI's stock price almost doubled. Two months later, upon the financial report release, the actual growth was 58%. The Wall Street consensus was 14%.
Three cases, $AXTI, $SIVE, RPI, share the exact same underlying logic.
Find a position in the industry chain where pricing is not yet in place but demand has already been secured.
Three Most Common Pitfalls for Retail Investors
At this point, I would like to discuss the three most common pitfalls that retail investors fall into when buying stocks and what problems this method can actually solve.
First Pitfall: Chasing Highs and Selling Lows, Always Holding the Bag
Seeing a hot sector, a stock on the rise, news hyping it up, and then jumping in. Only to see it start falling after you jump in. As it falls, you can't hold on and end up selling. After you sell, it rises again, leaving you feeling confused...
Why? Because when something is so hot that even someone like me, who doesn't delve deep into research, has heard about it, its pricing has long been maxed out by the market.
First-tier and second-tier companies, Nvidia, Microsoft, TSMC, with analysts worldwide watching them and funds globally buying in, why would I think I'm faster than others?
Serenity's method is to dig deeper. Go down to the third tier and below, where analysts don't cover, institutions don't hold positions, the market cap is too small, and the liquidity is too poor for them to enter. In these areas, there are plenty of pricing loopholes to exploit.
This solves one problem: You don't have to compete for speed with the world's smartest people; you slowly climb in a race where no one is competing against you.
Second Pitfall: Buying, Panicking When It Drops, Panicking When It Rises
The reason for panic is actually one, you don't know how much the thing you bought is really worth.
You may have bought because it looked like it was going to rise. But if it looks like it's going to drop, will you sell? Without your own judgment anchor, your mindset will collapse with any market movement.
The Serenity approach forces you to answer three questions before placing an order: Is this node irreplaceable? How many global suppliers are there? Is downstream demand rising or falling?
Once you answer these three questions, you have an independent logic anchor unaffected by stock prices. So, there's no need to panic over a short-term drop unless the answers to those three questions change.
This solves one problem: Price movements are not based on faith but on the industrial chain diagram in your hands.
The Third Pitfall: Everyone sees the same information
PE ratio, ROE, financial report growth rate, northbound funds, top shareholders' list. When you open any stock trading app, everyone sees the same interface. All this information is already reflected in the stock price.
What does Serenity look at? Patent databases, supplier directories, customs export data, developer discussions in industry forums, and technical roadmap comparisons in academic papers.
These things are free and public, but most retail investors have never looked at them in their lifetime.
This solves one problem: When your information source is different from others, your judgment may differ from theirs.
There are also things this method cannot solve
Of course, this method is not omnipotent, and there are a few things it cannot solve.
· First, Serenity has made mistakes. UPWK -35%, HIMS -50%, CRCL -45%. Being right with the method does not mean every transaction is correct. No matter how accurately the industrial chain is depicted, problems in company operations, sudden industry shifts, or policy changes can all lead to a crash.
· Second, Serenity is anonymous. Former AI scientist, Nature paper author, rejected an offer from NVIDIA – these titles are all self-proclaimed, and no one can verify them. He buys only micro-cap stocks, a single tweet from him can drive up the price, and followers jumping in can further boost it. He has never disclosed when he sells.
· Third, his entire portfolio is based on two predictions: CPO will become the sole technical roadmap for data centers, and humanoid robots will reach a billion-unit scale. If either of these is overturned, the logic behind many of his targets will collapse.
His approach tells you this position is neck-deep, but whether the neck itself will always be there is for you to judge.
If We Apply This to Crypto?
I tried applying this approach to the Crypto track.
Start breaking down from the Meme Launchpad. Platforms like Pump.fun → What are they dependent on? Liquidity Pools → Where does the liquidity for the pools come from? Keep digging further down.
In plain language, what this method can offer you is: When you are used to digging from the top down, you won't FOMO just because something is pumping. You will instinctively ask, at what layer is it? Who is upstream? Who is upstream of the upstream? Where is that inevitable bottleneck where only one or two players operate and everyone is interconnected?
I pondered for a long time, reviewed the situation extensively, and suddenly had a moment of enlightenment.
I didn't discover a secret to wealth or anything; it's because I suddenly realized that every purchase made by many people, whether it's a stock or a coin, is all circulating in the first layer of the industry chain. Whatever is trending, everyone's decision-making radius doesn't extend beyond that self-selected page of the trading app.
Therefore, the Stock God Serenity's strategy isn't to directly tell you what to buy, and certainly not to urge you to FOMO along with him.
What we need to do is another thing, pull you out from the crowded first layer where everyone is and let you see that the pricing on the fifth layer is yet to be discovered by the market.
Welcome to join the official BlockBeats community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Official Twitter Account: https://twitter.com/BlockBeatsAsia
