The Hidden Advantage of Dixinli

Let's talk about Deminli. Before that, I'd like to tell you a story.

01

In 2008, in Shenzhen, a man named Li Hu had been selling chips in Huaqiangbei for eight years. He mainly sold a small thing called "flash memory controller" and also dealt in memory devices.

In November of that year, he quit his job. He pulled in a partner and opened a small company with a registered capital of 100,000 yuan.

In the storage industry, the raw material is wafers. They are circular slices, and when cut, they become flash memory chips. Only a few international giants in the world can produce them.

Everyone wants the good stuff, and big buyers line up to sign long - term contracts. A newly opened small company can't squeeze into that line.

What Li Hu targeted was another kind of stuff.

Wafers are graded. The high - quality whole wafers are called "prime chips". There is another type whose quality level is not up to standard, and each batch is different. Big factories are too lazy to deal with them. In the industry jargon, they are called "scrap chips".

Scrap chips are cheap, and few people dare to take them. Each batch has different characteristics. If you take them back, you have to have real skills to turn them into usable products. If not, they are just a pile of waste.

He gambled on his own ability. More than a decade has passed, and this company has gone public. If you search for it online today, the most frequently mentioned word is still "gamble".

It converted most of its assets into wafers stored in the warehouse. Those who scold it say that it is just a gambler betting all its assets on the market.

Okay, that's the end of the story.

The small workshop that took in scrap chips back then and the "gambler" with a full - stocked warehouse of wafers today are the same company. What on earth has it done in these more than a decade? What kind of company is it exactly?

I've seen some people say that a large part of Deminli's revenue comes from reselling the purchased wafers without any processing. Buying and selling them as they are, isn't it just a middleman?

My suggestion on this matter is: first, admit it. The admission part is very simple. This revenue stream does exist, and it's not a small amount.

I checked. In the year of its listing, which was 2022, this "raw material reselling" business earned it about 380 million yuan, accounting for 30% of its total revenue that year. This item in the ledger is called "storage wafers and wafer - packaged chips", which includes both prime chips resold as they are and semi - finished products after rough processing.

By the first half of 2023, this revenue stream still accounted for 20%. For a company that claims to produce storage products, 20 yuan out of every 100 yuan of its income comes from selling raw materials. Those who scold it as a middleman have the figures to back them up.

After admitting it, let's analyze it. The key to the analysis still lies in the scrap chips.

Can you buy as many scrap chips as you want? No. There are only a few upstream original manufacturers. Scrap chips are by - products of their production lines. The quantity depends on luck, and they are often not sold separately.

What does it mean?

If you want to get scrap chips, the rule is to buy prime chips together. Deminli wants scrap chips. What should it do with the prime chips it gets as a package but doesn't need immediately? Resell them.

This is the first origin of the "raw material reselling" revenue stream: it's just dealing with the leftovers from the bundled purchase.

The second origin is simple: later, it expanded its skills, and a part of the prime chips were also processed into semi - finished products and then sold. Although it's still selling, the quality is different.

How has this revenue stream changed in recent years? The proportion has been decreasing all the way. As the production of solid - state drives and embedded storage increased, in the 2024 annual report, the revenue from selling wafers was no longer listed separately but was included in the "other" column, and the whole column only accounted for a small fraction.

So, the term "middleman" was appropriate for a while, but it becomes less and less accurate as time goes by.

One thing that has remained the same from beginning to end: whether it sells or not, when it makes purchases, it always focuses on one thing: cheapness. Scrap chips are much cheaper than prime chips, and they also have an advantage that prime chips don't have: stable prices. No matter how crazy the market is, their price fluctuations are small.

02

Of course, the low price of scrap chips doesn't come for free. Then the question arises: Why can the things with different characteristics in each batch turn into usable products in Deminli's hands?

This has to do with its secret weapon. What weapon? It's the thing that Li Hu sold for eight years before starting his business, the flash memory controller.

What is a controller?

A flash memory chip can't work on its own. Where to write data, how to bypass bad sectors, and how to correct errors all depend on a chip to give instructions. This chip is called a controller.

Let me give an analogy: This industry is like making clothes.

The original manufacturers weave the cloth. Module manufacturers like Deminli are tailors. They buy cloth to make clothes. They have no say in the cloth price. Scrap chips are like leftover pieces of cloth. The controller is like a pair of scissors.

Whether the unruly leftover pieces of cloth can be tamed depends entirely on the scissors.

Most tailors buy ready - made scissors. There are scissor shops in the market. Taiwan's Phison and Silicon Motion have been making scissors for decades. Their scissors are good and can be used right away for cutting.

There are two costs: one is that you have to pay for each pair; the other is that the ready - made scissors are made according to the specifications of whole bolts of cloth and can't handle the leftover pieces.

Deminli did something that seemed stubborn at that time: it made its own scissors.

A few years after its establishment, it invested all its assets and set up a team to make scissors. It wasn't until 2016 that the first batch of its own controller chips were mass - produced and shipped.

When I was looking through its annual report, I saw a sentence that I thought was very wonderful. The annual report itself stated that the company chose to use scrap chips to make products in a differentiated way at the startup stage, and self - developed the controller to adapt to the characteristics of scrap chips.

In other words, these scissors were made specifically for cutting leftover pieces of cloth.

I've seen some people say that Deminli's controllers are low - end products and are not the same as those from real controller manufacturers. This statement needs to be taken with a grain of salt.

The "low - end" part refers to its origin, and it's not wrong. The first batch of scissors were used for rough work, cutting small items like memory cards and USB flash drives. Being rough was its starting point.

But over the years, it has been constantly refining its technology. It has developed from scissors for memory cards to scissors for solid - state drives, covering the mainstream types. Having a low starting point is one thing, and whether it is making progress is another. Those who scold it and those who praise it each have their own views.

As for the part about "not being the same", it's actually correct.

Phison and Silicon Motion are scissor shops. They make scissors and sell them to tailors all over the world. Deminli never sells its scissors. It never intended to be a scissor shop. It's a tailor who makes its own scissors, and it's a different business.

There is another misunderstanding, even bigger than the previous two. Let me clarify it briefly.

Some people, once they hear "self - developed chips", pin their hopes on it for domestic storage. Hold on; what it makes are scissors.

It designs the scissors' blueprints by itself and entrusts the manufacturing to foundries like SMIC and UMC. The cloth is still woven by the original manufacturers. A tailor who can make scissors is still a tailor. It's the wrong person to expect to weave the cloth.

Okay, now we can understand the logic of this business model.

When others make clothes, they buy cloth at the market price and have to keep spending money on scissors. What about Deminli? It uses leftover pieces of cloth that are sold in piles and makes its own scissors. It has survived for more than a decade with this model, but it has also been restricted by it for more than a decade: leftover pieces of cloth can't be used to make large items. No matter how refined the shop is, it's still a small shop.

03

The need to buy cloth in whole bolts forced Deminli to transform into a different company.

At the end of 2021, the end of the year before its listing, the raw materials stored in its warehouse were only worth 78 million yuan on the books. What does that mean?

For a company with an annual revenue of about one billion yuan, the cloth in its warehouse is only enough for a few weeks of production. It stated frankly in its documents that with limited capital, it had to make the money circulate.

Deminli, which was doing the leftover - piece business, was actually one of the tailors in the industry that stocked the least amount of cloth. It had a hand - to - mouth, small - scale operation.

The turning point came in 2022.

It went public in July of that year and raised funds. Around this time, it was determined to produce large - scale products like solid - state drives. The customers for large - scale products are industrial customers, with large orders and strict delivery deadlines. If you don't have enough cloth in your warehouse, why would they give you the orders?

From that year on, it started to buy cloth. And it bought more and more aggressively every year.

I looked through the accounts of recent years. At the end of 2022, the inventory in its warehouse, including both cloth and half - made clothes, was worth more than 700 million yuan. At the end of 2025, it was seven billion yuan. It increased nearly ten times in three years.

This is not the end. I checked the latest data. As of the end of March 2026, the inventory in its warehouse was about 12 billion yuan, accounting for two - thirds of its total assets. Six or seven out of ten pockets of its assets were full of cloth.

Those who scold it as a gambler are looking at this.

Its gambling logic is not hard to understand. The price of storage cloth has a cycle, rising and falling every few years. When the price hits the bottom, the whole industry pulls back. The lower the price, the less willing they are to buy.

Deminli does the opposite. The lower the price, the more aggressively it buys cloth. The logic is simple: the cloth is still the same, but the price tag at the bottom is the cheapest. This is the same logic as when it picked up leftover pieces of cloth, except that the stakes are hundreds of times higher.

How can the cheaply bought cloth turn into profit? There is an accounting trick here that is worth spending a minute to understand.

It records the cloth at the price it was stocked. The cloth stocked at the bottom price is recorded at that price in the books. When the market price rises, the clothes are sold at the new market price, but the cost of the cloth is still calculated at the old price. The difference between the selling price and the cost is all its profit. The earlier and more it stocks, the greater the difference.

It sounds great, right? But look at it from the other side.

What if the market turns around? The more than 10 billion yuan worth of cloth in the warehouse has to be re - valued according to the falling market price. The amount of the reduction will be directly deducted from the profit. The more it stocks, the greater the deduction.

Players with full - stocked warehouses tasted this pain during the last downward cycle.

It's not over yet. The money for buying cloth is not all its own. Its operating cash flow has been negative for four consecutive years since 2022. How to fill the hole? By borrowing money and raising funds from the market from time to time. In short, it is borrowing money, keeping a full - stocked warehouse, and gambling that the market will be on its side.

At this point, does it seem like a pure gambler? Don't rush to make a judgment. There is another aspect to consider.

Downstream customers are paying it deposits. The amount of this prepaid money was only tens of millions on the books two years ago. By the end of the first quarter of this year, it was about 1.3 billion yuan. Customers are willing to pay in advance and wait in line for the goods.

That is to say, a considerable part of the cloth in the warehouse already has a buyer, not just stocked up waiting for a favorable market.

This is the most contradictory and dramatic part of Deminli. The same warehouse of cloth can be seen as a bold gamble or as a pre - locked order. Both views can be supported by the accounts.

Which view is more accurate depends on the future market situation, and no one can guarantee it for it.

There is one thing that doesn't need to wait for the market to unfold. The analogy of making clothes can't hold up anymore: no tailor in the world would use two - thirds of their assets to stock up on cloth.

In a real tailor's shop, the valuable things are skills and regular customers. In Deminli's shop, the valuable thing is the cloth itself. This "difference" is exactly the most fundamental difference between this business and a tailor's shop: the cloth is their life.

04

All three tailors stock up on the cloth that is as important as their lives, but they spend their money in completely different ways. I've seen some people say: Deminli has no workshop and no brand. How can it sit at the same table as Jiangbolong and Baiwei?

It really doesn't have a workshop.

There is no need for outsiders to verify this. It made it clear at the performance briefing in May this year: In the packaging process of storage modules, it mainly uses external subcontractors. The self - owned production lines it controls are the testing process to ensure quality and the SMT integration process.

Translated into the language of tailors:

It outsources the sewing work to external factories. It only controls two aspects: inspecting the cloth when purchasing and ironing the clothes before delivery. The most important tools in a tailor's shop are not in its hands for a long time.

It has a clearer understanding of the brand. Its exact words in the investor records are: The company does not operate consumer - grade storage brands. Its products are mainly supplied to external terminal brand customers.

Who are these customers?

I checked the public list of approved customers. Brands like Netac, aigo, and Eaget that you've seen on the shelves are all on the list. You've probably used products it made. The brand belongs to others, but the clothes are made by it.

Interestingly, its general manager, Du Tiejun, used to work at Netac and worked his way up to the general manager of the storage division. Making clothes for his former employer's brand shows how small this industry circle is.

Not having a workshop and a brand seems like a shortcoming. To put it bluntly, it only has a limited amount of money and can only invest in one direction.

The three tailors started at a similar level, but they spent their money in completely different ways.

One spent money on the brand and the storefront, making its name well - known to consumers. Another spent money on building its own workshop, bringing the sewing work back in - house.

Deminli spent money on two things: making scissors and stocking up on cloth.

Investing the same amount of money in different areas has resulted in three completely different companies on the balance sheet. One is most like a brand - owner, one is most like a factory, and Deminli is least like a "factory". Its assets consist of blueprints on one hand and cloth on the other.

Why can it sit at the same table? That's why. When the market is good, the one with a full - stocked