From Apple to Nvidia: Why Dongshan Precision Has No Choice but to Go to Hong Kong

[[**After its Hong Kong IPO prospectus once lapsed, Dongshan Precision refiled and passed the listing hearing, planning to raise US$2 billion to US$3 billion. The company entered the optical module sector through its RMB 5.935 billion acquisition of Source Photonics, with first-half profit surging 290%, but a 64% debt ratio and a US$2.7 billion expansion bill force it to raise funds in Hong Kong to shore up its finances.**

**Key Points**

**1. A Hong Kong listing is the inevitable choice for debt repayment and expansion**

The company has total liabilities of RMB 38.549 billion, an asset-liability ratio of 64.5%, and short-term borrowings approaching RMB 9 billion; at the same time, it plans to invest more than US$2.7 billion in expansion over the next two to three years. A-share financing capacity is limited, and Hong Kong is the most effective channel for supplementing US dollar capital and lowering leverage.

**2. The Source Photonics acquisition ignited performance, but at the cost of soaring leverage**

In 2025 it acquired 100% of Source Photonics for RMB 5.935 billion in cash. In the first half of 2026, optical module revenue was RMB 5.351 billion with a gross margin of 39.2%, driving net profit attributable to shareholders up 290% year on year; but the acquisition was completed with debt, directly pushing up the debt ratio and adding new goodwill impairment risk.

**3. Customer concentration is abnormally high, and Apple dependence remains unresolved**

In 2025, its largest customer, "Customer A," accounted for 46.5% of revenue, and the top five customers together accounted for 64.4%. Fluctuations in Apple supply chain orders can determine the company's profit or loss in a single year, and its shift toward the Nvidia supply chain has not changed the fundamentals of a company defined by its major customers.

**4. Goodwill and cyclical risks compound, narrowing room for trial and error**

This article is from the WeChat public account: Xie Pu Notes, author: Xie Pu

May 18, 2026, HKEXnews: the status of Dongshan Precision's prospectus changed from "processing" to "lapsed," exactly 6 months after submission, not a day more

October 11, 2026: refiled and passed the hearing, with listing as early as that month

US$2 billion–US$3 billion: the planned fundraising size reported by Bloomberg

On May 18, 2026, on HKEXnews, the status of Suzhou Dongshan Precision's prospectus quietly changed. From "processing" to "lapsed."

From the application submitted on November 18, 2025, to that day, exactly 6 months, not a day more. Under HKEX rules, if the review process is not completed within 6 months, the application documents automatically become void. Half a year of work by the four sponsors—UBS, Haitong International, GF, and CITIC—was reduced to zero.

A hot AI computing concept stock—third in the world in PCB, second in the world in flexible boards, and having just swallowed optical module company Source Photonics—had the first step of its Hong Kong listing put on pause just like that.

Five months later, the tide turned. On October 11, Dongshan Precision refiled and passed the hearing. According to Bloomberg, it plans to raise US$2 billion to US$3 billion and list as early as that month.

Now the question arises. Dongshan Precision has been listed on the A-share market for 15 years, has a market value of more than RMB 260 billion, and had net operating cash flow of RMB 5.3 billion in 2025. Why must this company, which is not short of fame and not exactly poor on paper, go ring the bell in Hong Kong again?

To answer this question, we start with a small town by Lake Tai in Suzhou 46 years ago.

01 Sheet Metal Factory: The Beginning of Everything

In 1980, in Dongshan Town, Suzhou, Yuan Fugen opened a small sheet metal stamping factory.

What is sheet metal? Cutting, punching, bending, and welding metal sheets into casings and frames. TV antennas, communication equipment chassis, hardware parts for home appliances—this is the most inconspicuous link in manufacturing, with low added value and hard-earned money.

Hard-earned money, gathered bit by bit, can also lead to great achievements. The essence of this business is just two things: deliver on time and make no mistakes. Relying on these two things, Yuan Fugen turned a small workshop into "Suzhou Dongshan Sheet Metal Co., Ltd.," registered in 1998.

In 2004, when the SME board of the Shenzhen Stock Exchange opened, Yuan Fugen judged that the capital market was the next step and handed the task of pushing for a listing to his second son, Yuan Yonggang. In 2007, the company was renamed "Dongshan Precision," brought in strategic shareholders, and completed its shareholding reform.

In April 2010, Dongshan Precision listed on the Shenzhen Stock Exchange and raised RMB 1.04 billion. In the same year as the listing, the company's power transition was smoothly completed: Yuan Yonggang became chairman, his elder brother Yuan Yongfeng became general manager, and their father stepped back.

What happened next defined the company's character for the following 15 years—it did not grow much business on its own, but the businesses it bought were one bigger than the next.

In 2016, it spent about US$600 million to buy Nasdaq-listed MFLEX, acquiring the flexible printed circuit (FPC) business and a ticket into Apple's supply chain;

In 2018, it spent another roughly US$300 million to buy PCB manufacturer Multek from Flextronics, filling out its rigid board capabilities.

Three deals transformed a sheet metal factory into an electronic circuits company of "flexible boards + rigid boards + display modules." The subsequent rankings of second in the world in FPC and third in the world in PCB were all built on these two acquired businesses.

When it comes to borrowing a chicken to lay eggs—and golden eggs at that—the Yuan brothers are experts.

In 2023, Dongshan Precision acted again, spending RMB 1.383 billion to acquire JDI's Suzhou Jingduan, entering automotive displays; that same year, it turned around and acquired the US Aranda factory to get closer to North American automotive customers. Later, in 2025, it splashed out EUR 100 million to acquire French auto parts supplier GMD.

Then came the biggest bet. In June 2025, Dongshan Precision announced it would acquire 100% of Source Photonics for no more than RMB 5.935 billion in cash.

Source Photonics is one of only three companies in the world capable of vertically integrating "optical chip design—manufacturing—packaging—optical modules." Its self-supply rate for 200G EML optical chips is 99%, its 800G optical modules are being delivered in volume, and its 1.6T products have passed Nvidia's validation and are shipping in small batches.

Interestingly, this deal had a small sideshow: the first company to set its sights on Source Photonics was not Dongshan Precision but the real estate company Vantone Development. In June 2024, Vantone offered USD 324 million to buy it, but the cross-industry leap was too great, and the deal fell through in January 2025. Five months later, Dongshan Precision raised its bid by nearly double and plucked this peach.

Even if a deal fails, goodwill remains—Wang Jianshi and Wang Hongyu have Vantone Development to thank.

Vantone Development's "price bump" raised the price, but the deal was still a bargain. The capital market voted on this deal with real money—three months after the announcement, Dongshan Precision's share price rose from RMB 31 to RMB 86, a gain of 178%. Measured from the low of RMB 21.55 in April 2025, it surged to a peak of RMB 222 within a year—a 10x gain.

03 Financial Checkup: Cash Flow Is Real, and So Is the Leverage

Having understood Dongshan Precision's development history, we can now give it a checkup. Let's look at the good parts first.

In 2025, Dongshan Precision's revenue was RMB 40.125 billion, surpassing RMB 40 billion for the first time in its history, up 9.12% year on year; net profit attributable to shareholders was RMB 1.386 billion, up 27.67%.

Even more striking, actually, is net operating cash flow: RMB 5.307 billion—nearly 4 times net profit, showing that the money earned is real cash, not receivables.

The real upheaval came in the first half of 2026: revenue of RMB 27.798 billion, up 64% year on year; net profit attributable to shareholders of RMB 2.957 billion, a surge of 290% year on year. Half a year's profit was more than twice the full year of 2025.

Why did profit suddenly explode? Because Source Photonics was consolidated. In the first half, the optical module business generated revenue of RMB 5.351 billion, accounting for 19.2%, with a gross margin of 39.2%—while the company's overall gross margin had long been only 12% to 14%. A business accounting for one-fifth of revenue contributed more than half the profit. Optical modules lifted the entire company's gross margin from 13.1% to 19.4%.

Having covered the good side, let's look at the other half of the checkup report. First, the debt problem.

As of the end of 2025, Dongshan Precision's total liabilities were RMB 38.549 billion, surging 42% year on year, with an asset-liability ratio of 63.98%, which rose further to 64.5% in the first half of 2026. Short-term borrowings were RMB 8 billion, approaching RMB 9 billion in the first quarter of 2026; the books also held RMB 9.8 billion in inventory and RMB 9.6 billion in receivables.

Beyond liabilities, there is the speed of spending.

The RMB 5.935 billion for Source Photonics was only the down payment: in June 2025, the board approved an optical chip and optical module capacity expansion plan with total investment of USD 1.2 billion; in August, it was increased to USD 1.7 billion. Add Multek's USD 1 billion planned investment in high-end PCB capacity, the Thailand base, and the Mexico base—over the next two to three years, the company plans to spend more than USD 2.7 billion.

Dongshan Precision really does make money, and it also spends without blinking. Now the question is—where does the money come from?

In June 2025, the company conducted a private placement to brothers Yuan Yonggang and Yuan Yongfeng, raising RMB 1.4 billion. Of that RMB 1.4 billion, RMB 500 million was the brothers' own funds, and the remaining more than RMB 900 million was borrowed by the two from banks and third parties. The actual controllers staking their entire fortunes and taking on debt to inject funds—this action itself answers the question of whether the company lacks money.

Looking back now at that prospectus that lapsed in May, the use of proceeds was written plainly: building and upgrading capacity (high-speed EML chips, optical modules, the Zhuhai Chaoyi high-end PCB production line), potential strategic investments and acquisitions, repaying part of existing debt, and supplementing working capital.

So let's move to the next question: A-shares allow fundraising, so why must it be Hong Kong? Three reasons.

First, the capacity expansion is in Thailand and Mexico, and buying equipment and building factories require US dollars. Renminbi raised on the A-share market has friction costs when taken abroad, while Hong Kong provides direct access to offshore funds. Second, overseas revenue accounts for more than 80%, and customers are global cloud providers; an internationalized shareholder register and financing platform are part of the business. Third, and most practically: the 64% debt ratio needs equity financing to bring down, the A-share private placement has just been done once with limited quota, and Hong Kong's USD 2 billion to USD 3 billion is the largest bucket of fresh blood this machine can find.

In simple summary: the body that has grown fat through acquisitions needs the Hong Kong market for a blood transfusion.

04 Risk List: Three Blades Hanging Overhead

All right, according to the report, we should now go look at the risk page of the prospectus. Dongshan Precision has three uncertainties.

First, something called Customer A.

The prospectus shows that in 2025, Dongshan Precision's top five customers accounted for 64.4% of revenue, and the largest customer—the "Customer A" everyone knows who that is—accounted for 46.5%. Over a longer stretch, the largest customer contributed 29.81% of revenue in 2017, and in 2023 that figure once reached 55.81%.

The fate of the Apple supply chain has never been in suppliers' hands. An increase or decrease in one Apple order can determine a company's profit or loss for a year; the cautionary tales of Goertek and OFILM are all written in A-share candlestick charts. Dongshan Precision used Source Photonics to open the doors to Nvidia and cloud providers, but a 46.5% concentration means that in the first half of its transformation, it is still a company defined by Apple.

The second uncertainty is the resonance between goodwill and liabilities.

The acquisition was made with 5.935 billion in cash, and the premium portion will sit on the books as goodwill. Source Photonics' technology iteration cycle is 18 months—800G is thriving, 1.6T has arrived, and 3.2T is on the way. If it falls behind in the next-generation optical chip competition, goodwill impairment will directly eat into profits. This is not a hypothetical: in 2024, the company's net profit fell from 1.965 billion the previous year to 1.085 billion. This company's income statement has historically been unable to withstand major impairment charges.

In the CSRC's five supplementary inquiries in January 2026, goodwill impairment and cash flow pressure were prominently listed; the same inquiry also named controlling shareholder equity pledges and data export compliance. That lapsed prospectus was, in a sense, "dragged out" by these problems.

The third uncertainty is cycles and geopolitics.

AI capital expenditure is the engine of this story. The global optical module market grew from $11.2 billion in 2021 to $24.5 billion in 2025, and industry forecasts put it at $108.6 billion by 2030—but every scenario of "if cloud providers slow capital expenditure" would land unchanged on orders for optical chips and high-end PCBs. At the same time, with 80% of revenue overseas, Thai capacity ramping up, and a Mexican base under construction, exchange rates, tariffs, and export controls each directly affect the income statement.

In Hu Xijin's words, this is a complex globalization.

05 Outlook: From "Earning Apple's Money" to "Earning Nvidia's Money"

Setting risks aside, the cards in Dongshan Precision's hand are indeed the best in its 46-year history.

According to China Insights Consultancy data, it is the only supplier in the world with capabilities in PCBs, optical chips, and optical modules at the same time. Together, these three account for about 9% to 14% of the material cost of an AI server—in most AI servers, second only to GPUs in cost share. PCBs handle signal transmission inside servers, optical modules handle interconnection between servers, and optical chips sit at the highest-value point of the entire chain. Vertical coverage from boards to modules to chips means it is not selling a component but a segment of computing infrastructure.

The pace is also materializing: the first phase of AI PCB capacity will be released in Q3 2026; Source Photonics' 1.6T optical modules are expected to ship in volume in Q4 2026; and the $1.7 billion capacity expansion will land in Changzhou, Thailand, and Yancheng.

But it must be made clear: this is projection, not fact. Capacity ramp-up, validation by major customers, and positioning in next-generation technology each carry the risk of falling behind. From a "consumer electronics base" to an "AI computing hardware platform," there lies an entire cycle that requires burning money, time, and yield rates. The 5.351 billion in optical module revenue in the first half is a beautiful starting point, but what it must prove is that this number can survive competition in the next generation of optical interconnect technology and keep expanding.

In 1980, the sheet metal factory in Dongshan Town earned its first pot of gold with stamped parts; from 2016 to 2018, the Yuan brothers spent $900 million to buy back FPC and PCB, giving the company a second blood transfusion; in 2025, Source Photonics at 5.9 billion yuan is the third.

Each blood transfusion follows the same posture: spot a track it cannot grow on its own, take out cash, take on debt, and then use its listed-company status to raise money from the capital market to replenish blood. In 2010 it was the Shenzhen Stock Exchange; in 2026 it is the Hong Kong Stock Exchange.

So that prospectus, lapsed and reborn, is essentially not an IPO but the company's very way of survival—M&A is its engine, and the capital market is its fuel tank.

The engine is being swapped from internal combustion to electric, and the fuel tank must change with it. The only problem is that this time, the $2.7 billion expansion bill and the 64% debt ratio leave it with far less room for trial and error than the sheet metal factory had in 1980, and far less than in 2016.

The cards are the best cards, but the road is a bit narrow. It needs a bet. Channel: Finance and Economics. This content comes from the internet; original link. The views represent only the author and do not represent Huxiu's position. For copyright issues, please contact hezuo@huxiu.com, and we will verify and handle it promptly. Everyone who is changing the world and wants to change it is on the Huxiu APP [ ]