The most expensive "pre-emptive bet" in China's semiconductor industry

On the first trading day of Yuexin Semiconductor's listing, its market value exceeded 100 billion yuan.

On the morning of October 13, a chip company that has not yet turned a profit reached a market value of 100 billion yuan for the first time.

Yuexin Semiconductor landed on the ChiNext board with an offering price of 12.01 yuan. Its opening price jumped directly to 45 yuan, representing a 274.69% increase, corresponding to a total market value of about 129.6 billion yuan. After that, the stock price retreated somewhat, and finally closed at 38.80 yuan, still rising 223.06% on the first day of listing, with a total market value of about 111.7 billion yuan.

This is a set of data that is more than enough to create a visual impact.

Because a few months ago, the company's financial statements told a completely different story.

In the first half of 2026, Yuexin Semiconductor recorded operating revenue of 1.777 billion yuan, up 68.76% year on year, but its net profit attributable to shareholders of the parent company still posted a loss of 1.052 billion yuan. The loss after deducting non-recurring gains and losses further reached 1.163 billion yuan, while its gross profit margin stood at -24.46%.

If we extend the timeline further back, the losses are even more noticeable.

From 2023 to 2025, Yuexin Semiconductor's operating revenue reached 1.044 billion yuan, 1.681 billion yuan and 2.582 billion yuan respectively, but its net profit attributable to the parent company for the same period recorded losses of 1.917 billion yuan, 2.253 billion yuan and 2.346 billion yuan respectively. Adding the loss in the first half of this year, the company's cumulative attributable loss in the past three and a half years has exceeded 7.5 billion yuan.

Yet now, the capital market has priced this company at more than 110 billion yuan.

Based on its 25.82 billion yuan revenue in 2025 for rough calculation, the closing market value on the first listing day is equivalent to more than 40 times its revenue of last year; calculated at the nearly 130 billion yuan market value at opening, this figure once approached 50 times.

This is clearly no longer an investment calculated based on current profits.

In fact, for Yuexin, there is not even a price-to-earnings ratio to discuss for now.

What investors are buying is the future.

The question is, what kind of future can make a wafer fab that still has a negative gross profit margin at present obtain a valuation of 100 billion yuan?

01 The Critical Point of Yuexin

To understand Yuexin Semiconductor, you first need to understand the nature of the wafer fab business.

It is different from most asset-light technology companies.

For a software company, as long as it adds servers and engineers, after the revenue scale expands, the new cost does not necessarily grow at the same speed.

But this is not the case for wafer fabs.

Before producing the first wafer, it must invest huge sums of money in building clean rooms, purchasing equipment for lithography, etching, thin film deposition, ion implantation and other processes.

Once the equipment is put into use, depreciation will continue to be generated.

According to Yuexin's prospectus, its machinery and equipment are expected to have a service life of 6 years, with an annual depreciation rate of 15.83%. This means that for a wafer fab that is still in the capacity ramp-up stage, depreciation will still be included in costs even if the machines are not operating at full capacity.

This is also the important background for Yuexin's long-term losses.

What is really worth observing is not just the loss figures, but another indicator: capacity utilization rate.

In 2023, Yuexin Semiconductor had a production capacity of 449,500 wafers, with an actual output of 248,300 wafers, and the capacity utilization rate was only 55.23%.

By 2024, this figure rose to 84.79%.

In 2025, the company's production capacity reached 614,800 wafers, with an actual output of 592,600 wafers, and the capacity utilization rate further increased to 96.38%.

This set of data can better explain why the capital market is willing to give Yuexin such a high price than the net profit figure.

The most dangerous state for a wafer fab is not high depreciation, but that huge amounts of equipment have been built, but there are not enough orders to keep the machines running.

Once the utilization rate increases, the situation begins to change.

The same machine, the same plant, the same amount of depreciation, are spread over more wafers, so the unit manufacturing cost will decrease.

This change has already begun to appear in Yuexin's financial statements.

In the first half of 2025, the company's gross profit margin was still -54.70%; by the first half of 2026, the gross profit margin had improved to -24.46%.

It is still a negative number, but it has increased by about 30 percentage points in one year.

The explanation Yuexin gave in its prospectus is also very straightforward: first, the sales volume of high-priced products such as SiPho silicon photonics, BCD power management and HV high-voltage display drivers has increased, raising the average selling price; second, the expansion of production and sales scale has diluted the unit depreciation; third, the increase in the localization rate of raw materials has reduced material costs.

In other words, the first variable that the market is betting on now is not "when will Yuexin start making profits", but whether the gross profit margin can quickly cross the zero line after a wafer fab with utilization rate close to full capacity continues to increase product prices and optimize product structure.

This is the first layer of logic behind the 100-billion-yuan valuation.

02 The Answer to the 100-Billion-Yuan Valuation Lies in "Silicon Photonics"

If Yuexin were just another ordinary mature process wafer fab, the 100-billion-yuan valuation would still be hard to explain.

What the capital market is really interested in is its second label: silicon photonics.

Yuexin's existing processes do not pursue the most advanced 3nm or 5nm nodes.

The process nodes of its two existing wafer fabs mainly cover 180nm to 55nm, which belong to mature processes; the under-construction Phase IV project of Yuexin is planned to cover 65nm to 22nm.

But mature processes do not equal low-end positioning.

A large number of products such as analog chips, power management chips, automotive electronics, industrial control systems, CMOS image sensors and power devices do not require the most advanced process nodes.

AI has added a new variable to this logic.

What data centers really need is not only GPUs.

When computing clusters expand from thousands of GPUs to tens of thousands or even more chips, how to transmit data between chips has become a new bottleneck.

As a result, high-speed optical modules, CPO, silicon photonics and optoelectronic integration have gradually entered the core chain of AI infrastructure investment.

Yuexin happens to occupy a position in this track.

The company has formed process platforms for mixed-signal, high-voltage display drivers, CIS, embedded storage, BCD and SiPho silicon photonics. Citing data from Frost & Sullivan in its prospectus, as of the end of April 2026, Yuexin is a rare player among enterprises in mainland China that have the mass production capacity for 12-inch silicon photonic wafers. The company also stated that the end customers of its silicon photonics business include well-known cloud service vendors in the industry.

A very important valuation change has emerged here.

In the past, when the capital market looked at mature process wafer fabs, it focused more on production capacity, utilization rate and wafer prices.

Now, it is looking at which production capacities can be connected to the AI industrial chain.

This means that for the same 12-inch mature process wafer, the economic value of different process platforms can be completely different.

If it mainly produces traditional consumer electronics chips with fierce competition, it is closer to a general manufacturing business.

If it can undertake high value-added products such as silicon photonics, optoelectronic integration and high-end BCD, it will have more technology growth attributes.

Yuexin obviously wants to enter the latter category.

This is also why the company has prioritized the planning of processes such as silicon photonics and optoelectronic integration, embedded storage, image sensors, analog-to-digital mixing, radio frequency and OLED display drivers in its Phase IV production line.

The Phase IV project has a planned production capacity of 40,000 wafers per month.

Adding the planned 80,000 wafers per month of its two existing factories, after all construction is completed, Yuexin's planned monthly production capacity will reach 120,000 wafers — nearly twice the actual monthly production capacity of 63,300 wafers at the end of 2025.

The 100-billion-yuan price that the capital market is paying today has largely already priced in this "Yuexin with 120,000 wafers per month" in advance.

Rather than the current Yuexin with a monthly capacity of 63,300 wafers.

03 The Other Side of Doubling Production Capacity

But the real problem also starts right here.

Doubling production capacity sounds like a growth story.

For wafer fabs, it is also a risk story at the same time.

Because every additional factory means new equipment, new depreciation and new fixed costs.

If demand grows as expected and revenue grows faster than depreciation, profits will improve rapidly.

But if the newly added capacity encounters industry downturn, falling wafer prices or slower-than-expected customer verification progress, the new capacity may also continue to amplify losses.

Yuexin does not avoid this point.

The company estimated in its prospectus materials that its operating revenue in the first three quarters of 2026 will be about 2.87 billion to 2.97 billion yuan, but its net profit attributable to the parent company will still record a loss of 1.49 billion to 1.59 billion yuan.

In other words, even if the company's revenue maintains a rapid growth rate of more than 60% this year, it is still far from achieving real profitability.

Yuexin's own forecast is to achieve overall break-even as early as 2029.

To achieve this goal, the company's revenue forecast is very aggressive.

In its reply to the regulatory inquiry letter, Yuexin estimated that its operating revenue from 2026 to 2029 will reach about 3.551 billion yuan, 7.093 billion yuan, 9.278 billion yuan and 12.47 billion yuan respectively.

In other words, its 2029 revenue needs to reach nearly 5 times its 2025 level.

And there is a detail that the market often tends to ignore.

The company's prospectus clearly states that the forecast of achieving profitability as early as 2029 is based on multiple assumptions including sustained revenue growth, rising gross profit margin, capacity ramp-up as planned, smooth R&D transformation of new products, and access to a reasonable level of government subsidies. The company also specifically noted that this profit forecast includes the income brought by government subsidies.

In the first half of 2026, the non-recurring gains and losses attributable to shareholders of the parent company of Yuexin was 111 million yuan, the main source of which was government subsidies included in the current profit and loss.

This means that what investors are facing is not a realized profit model, but a series of assumptions that must be valid consecutively: the capacity must ramp up smoothly, the utilization rate must remain at a high level, the proportion of high value-added products such as silicon photonics and BCD needs to continue to increase, the selling price of wafers cannot drop significantly, and the depreciation formed by new capacity must be diluted by larger revenue.

At the same time, the industry must not enter a severe mature process price war.

If any part of these assumptions is missing, the 2029 profitability timeline may be pushed back.

This is also why simply defining Yuexin as an "AI chip company" can easily lead to misjudgment of this investment.

It is first and foremost a wafer manufacturing enterprise, and the final competition in wafer manufacturing has never been about concepts, but about yield, utilization rate, product structure and cost.

04 SMIC and Hua Hong Provide Two Additional Reference Benchmarks

Putting Yuexin into China's wafer foundry industry, the tension of its 100-billion-yuan valuation will become more obvious.

SMIC achieved sales revenue of 67.323 billion yuan in 2025, with a monthly capacity equivalent to more than 1 million 8-inch standard logic wafers, a capacity utilization rate of 93.5% and a gross profit margin of 22%.

In other words, SMIC's revenue scale is about 26 times that of Yuexin in 2025, and it has already established positive gross profit margin and profitability.

Hua Hong, which has a feature process route closer to Yuexin's, recorded sales revenue of 717.5 million US dollars in the second quarter of 2026, with a gross profit margin of 16.5%; the company stated that it maintained high capacity utilization in the quarter, with both product volume and price rising.

While Yuexin's gross profit margin in the first half of 2026 is still -24.46%.

This does not necessarily mean that Yuexin is overvalued.

But it illustrates one thing: the valuation that the market is giving Yuexin now is not based on the profitability it has already achieved as a mature industry leader.

It is actually a "growth right valuation".

Investors are paying in advance for the price of the company after it grows from 2.5 billion yuan revenue to 10 billion yuan revenue, from negative gross margin to positive gross margin, and from a regional wafer fab to an important national player in feature processes.

This is also why Yuexin's stock price on the first day of listing looks so abnormal.

Judging from the profit statement, it is very expensive.

Judging from the industrial scarcity, the market is willing to give it time.

05 The 100-Billion-Yuan Market Value Is Essentially a Bet on the Period After 2029

The capital logic of China's semiconductor industry in the past few years has undergone a subtle change.

The first stage was to solve the problem of "whether we have it or not".

Whether we have wafer fabs, whether we have equipment, whether we have materials, whether we have domestic chips.

Capital was willing to pay a very high price for the process from 0 to 1.

But with more and more production lines completed, the industry is entering the second stage: after the facilities are built, can they make profits?

This is especially important for Yuexin.

Its significance to Guangdong's semiconductor industry is not difficult to understand.

Guangdong is one of the most important electronic manufacturing bases in China, but it has long lacked 12-inch wafer manufacturing capacity that matches the scale of its terminal industry. Yuexin is the first domestically cultivated 12-inch wafer manufacturing enterprise in Guangdong to achieve mass production, which itself carries strong industrial strategic value.

But the capital market will not only price industrial strategic value in the long run.

Stock prices will eventually return to cash flow.

Therefore, what is really worth paying attention to on Yuexin's first listing day is not the 274% increase.

Not even the peak market value of nearly 130 billion yuan.

What is really worth tracking is a set of increasingly specific figures in the next three years: when will the gross profit margin turn positive? In the process of monthly production capacity growing from 63,300 wafers to 120,000 wafers, can the utilization rate still remain above 90%? How much revenue can high value-added businesses such as silicon photon