📌AI Industry Chain Research Lab｜Issue 1

Many people follow AI, Nvidia, and TSMC daily, and know terms like CPU, GPU, and HBM, but few can clearly explain the relationships between them.
Without a clear understanding of the semiconductor industry chain, making money from AI is impossible.
Some still can't grasp why some semiconductor companies focus on design, others on manufacturing, and yet others only on packaging.
Today, I'll explain these issues in 5 minutes by connecting them through one main thread:
How does a grain of sand become a chip?
By understanding this thread, you'll not only grasp the semiconductor industry but also know where a company's value comes from.
🔔① What's the relationship between semiconductors, chips, and CPUs?
(See Image 01)
Many people's first mistake when encountering semiconductors is confusing these three terms.
In fact, they are hierarchical.
Semiconductor refers to the entire industry.
It includes all stages: materials, equipment, design, manufacturing, packaging, and testing.
A chip is a product made from semiconductor materials—essentially an integrated circuit with a large number of transistors.
A CPU is just one category of chips.
Besides CPUs, there are GPUs, memory chips, analog chips, RF chips, AI accelerator chips, and more.
So remember this:
Semiconductor is the industry, chip is the product, and CPU is just one type of chip.
Many people studying semiconductor stocks like to directly discuss specific companies.
But in reality, before discussing companies, it's more important to first establish this industry chain map.
Otherwise, it's like analyzing an auto company without knowing what components a car has—it's easy to get lost.
🔔② Why is it called "semi-conductor"?
(See Image 02)
Materials in the world can be roughly divided into three categories.
The first is conductors, such as copper, silver, and aluminum. Electric current can flow through them almost freely.
The second is insulators, such as plastic, rubber, and glass, which hardly conduct electricity.
Semiconductors fall in between.
Their biggest feature is not that they "conduct a little," but that their conductivity can be artificially controlled.
The most widely used material in modern chips is silicon.
Silicon itself is not a particularly good conductor, but by doping it with elements like boron or phosphorus, its conductivity can be precisely controlled.
Transistors were invented using this property.
It can be said that without silicon, there would be no modern computers. That's why the entire industry is called the semiconductor industry.
🔔③ How does a grain of sand become a chip?
(See Image 03)
The starting point of a chip is ordinary quartz sand. After high-temperature purification, high-purity polycrystalline silicon is obtained.
But at this stage, it cannot yet be used to make chips.
Because the crystal arrangement inside polycrystalline silicon is disordered, causing interference when electrons move.
Engineers then use a process called the Czochralski method to slowly pull the polycrystalline silicon into a single crystal silicon ingot. Only then can electrons move stably along the designed path.
Next, the silicon ingot is sliced into thin circular wafers less than 1 mm thick.
These are the most important basic material in the entire semiconductor industry—wafers. Many people mistakenly think wafers are chips.
In fact, they are not. A wafer is more like a blank sheet of paper.
All circuits are first drawn on this blank sheet.
Finally, it is cut into individual chips.
So, in the future, if you see a company's business mentioning "silicon wafers" or "wafers," they are not selling chips but the most basic raw material for chip manufacturing.
🔔④ How are chips "etched" out?
(See Image 04)
Having a wafer is far from enough. What truly determines a chip's performance is the subsequent manufacturing process.
Many people think chips are "produced."
Actually, more accurately, they are sculpted layer by layer.
First, the chip design company completes the circuit design. Then, the fab evenly coats the wafer surface with photoresist. Next, using a lithography machine, the designed circuit pattern is "exposed" onto the wafer surface.
Which areas need to be retained and which removed are already designed in advance.
Then, using etching equipment, the unwanted parts are "corroded" away little by little.
After that, through processes like deposition, ion implantation, and CMP polishing, new materials are stacked layer by layer.
Then lithography, etching, and deposition are repeated.
Advanced chips often require repeating this process hundreds of times.
Finally, tens of billions of transistors are built on a piece of silicon the size of a fingernail.
This is the true process of a chip's birth.
At this point, a wafer has completed the most complex manufacturing steps.
But it still cannot be used directly.
Why?
Because it is still a "bare die."
🔔⑤ Why can't chips be sold directly after being made?
(See Image 05)
After thousands of steps such as lithography, etching, and deposition, a wafer is finally finished. But at this stage, it still cannot be installed into a computer or a phone.
The reason is simple:
It's too fragile.
A real chip is actually only a few square millimeters to tens of square millimeters in size.
Once cut, it becomes a piece of bare silicon.
It has no protective layer, no pins, and cannot connect to a motherboard.
So, two final steps remain:
Packaging and testing.
Packaging is not just "wrapping it up."
It also performs three important tasks:
First, protect the chip.
Second, help with heat dissipation.
Third, connect the chip to external circuits.
Finally, after testing, the chip's performance, power consumption, and stability are confirmed to meet requirements.
Only then is a chip truly ready for sale.
Many people think packaging is just the final step.
In reality, in the AI era, advanced packaging has become one of the most important technologies in the entire industry chain.
Why?
Because GPUs are getting larger, HBM is increasing, and chiplets are becoming more complex.
Packaging no longer just determines whether a chip can be used; it also determines the upper limit of chip performance.
Therefore, in recent years, advanced packaging has become one of the hottest directions in the industry.
🔔⑥ Why is the division of labor in the semiconductor industry becoming increasingly detailed?
(See Image 06)
If you observe the semiconductor industry, you'll notice a very interesting phenomenon. Almost no company can do everything itself.
Why?
The answer is two words:
Too expensive.
Building an advanced fab often requires tens of billions of dollars in investment. Developing a generation of advanced process nodes takes years.
Plus, equipment, materials, and process technology—each link requires long-term accumulation.
As a result, the entire industry has gradually formed a professional division of labor. Every company concentrates its resources on the part it does best.
This is the reason behind today's semiconductor industry chain structure.
🔔⑦ Why does AI drive the entire industry chain?
(See Image 08)
Many people think: AI market movement equals Nvidia's movement.
In fact, that's just one part of the industry chain.
An AI server is not just about GPUs.
It also requires:
CPUs for scheduling,
HBM for high-speed memory,
PCBs for connectivity,
High-speed switches for communication,
Optical modules for transmission,
And advanced packaging to integrate them all.
If any link falters, the entire AI server cannot function properly.
So for every additional dollar invested in AI, not only GPU manufacturers benefit, but the entire semiconductor chain.
That's why, in the past two years, we haven't only seen Nvidia rise.
Companies like TSMC, Broadcom, Micron, SK Hynix, Samsung Electronics, Applied Materials, and ASML have also continuously benefited.
🔔⑧ Studying a semiconductor company: First answer one question
(See Image 09)
When we look at a semiconductor company, don't rush to look at its PE ratio or stock price.
First ask yourself:
Where does it stand in the industry chain?
Because the position in the chain determines what kind of profit it makes.
Material companies earn money from consumables.
Equipment companies earn money from selling machines.
Design companies earn money from intellectual property.
Fabs earn money from manufacturing capability.
Packaging companies earn money from advanced processes.
Different positions mean completely different business models.
Understanding this makes the valuation logic of many companies very clear.
Final Words
Many people study AI by focusing on only one company.
But what truly drives the AI revolution is never a single company.
It is a complete industry chain spanning materials, equipment, design, manufacturing, packaging, servers, and cloud computing.
Understanding this chain means you no longer just see stock prices—you see the underlying logic of capital flow in the entire AI era.
In the next issue, we will continue to break down one of the most confusing topics:
What are the differences between CPU, GPU, NPU, FPGA, and ASIC?
Why is AI training almost inseparable from GPUs?
Why is the inference chip space seeing a proliferation of options?
Where is the real competition among AI chips?