Wuhan and Yangtze Memory waited twenty years; the cooling of the A-share memory sector narrows the IPO window for Changcun Holdings
Changcun Holdings' STAR Market IPO was accepted on August 21, with plans to raise 33 billion yuan; as of October 25 it remains at the inquiry stage, while the memory sector has shifted from a buying frenzy to a cold reception. Since July, the STAR 50 has fallen 40.85%, Biwin Storage, GigaDevice, and Longsys have all dropped more than 60%, and Changxin Technology has pulled back more than 30% from its August high. Changcun's net profit in Q1 2026 was 33.379 billion yuan, ranking third globally in NAND; if its valuation is revised downward, the same amount of financing will bring greater equity dilution.
Body · AI translation
[[**Storage sector cools, narrowing the IPO window for Yangtze Memory Holdings: the capitalization milestone that Wuhan's nearly two decades of semiconductor investment has been waiting for is now facing a dual test of falling valuations and industry cycles.**
**Key points**
**1. Memory and AI assets shift from scramble to cold shoulder**
Changxin Technology rose 465.82% on its first day, reaching a market value of 3.28 trillion yuan, but since July the STAR 50 has fallen 40.85%, with Biwin, GigaDevice and Longsys all down more than 60%; Dongshan Precision's Hong Kong margin financing was only 0.39x subscribed, and Solidigm suspended its IPO, showing simultaneous cooling in both primary and secondary markets.
**2. Yangtze Memory's earnings surge but its IPO window narrows**
Yangtze Memory Holdings was accepted for review on August 21 and plans to raise 33 billion yuan, but as of October 25 it remains in the inquiry stage; its net profit in the first quarter of 2026 was 33.379 billion yuan and it ranks third globally in NAND. If its valuation is revised down, the same fundraising amount will bring greater equity dilution.
**3. Wuhan's long-term industrial investment faces a timing mismatch in realizing returns**
Wuhan invested in building Wuhan Xinxin in 2006 and established Yangtze Memory in 2016, followed by the Tsinghua Unigroup restructuring and increased local state-owned capital; Wuhan Xinxin withdrew its IPO and Voyah fell 13.2% on its first day. If Yangtze Memory continues to wait, it may miss the current high-profit cycle for memory. 2026-10-25 20:15 Wuhan and Yangtze Memory waited twenty years, but the A-share party has dispersed. Super Focus Quick Overview
This article is from the WeChat public account: Super Focus foci, author: Sean
Memory chips are still rising in price, but the memory giants no longer want to go public.
On October 23, SK Hynix announced that it was suspending the Pre-IPO financing and IPO plans of its U.S. subsidiary Solidigm. This memory company, once reported to have a valuation of up to $150 billion and expected fundraising of up to $15 billion, ultimately decided to temporarily withdraw from the capital market.
Recently, capital market volatility has intensified, semiconductor stock prices have continued to adjust, and South Korean regulators have increased scrutiny over duplicate listings and minority shareholder rights, so continuing to push ahead with the IPO is clearly no longer worthwhile for SK Hynix.
On the same day, Dongshan Precision, which is sprinting toward a Hong Kong listing, also encountered a chilly subscription.
According to Zhitong Finance statistics, as of the end of bookbuilding on October 23, margin financing subscriptions provided by brokers such as Futu and Tiger for Dongshan Precision amounted to only HK$418 million, equivalent to 0.39 times the initial size of the Hong Kong public offering. Although 20 cornerstone investors committed in advance to subscribe to about $680 million, in the eyes of investors seeking margin financing for new shares, the AI industry chain signboard is no longer so attractive.
One company saw the poor market conditions and decided not to list for now, while another has already reached the issuance stage but found investors less enthusiastic than expected. And in Wuhan, there is another memory giant that attracts even more attention and has not even waited for the day of issue pricing.
On August 21, Yangtze Memory Holdings' STAR Market IPO application was accepted by the Shanghai Stock Exchange, with plans to raise 33 billion yuan. More than two months later, the review is still at the inquiry stage.
Wuhan has waited nearly twenty years for this day, but just as it finally brought Yangtze Memory to the doorstep of the capital market, investors who once warmly welcomed semiconductor companies seem to have started clearing the table.
01
Changxin has eaten its fill, while Yangtze Memory has not even seen the table
If we turn the clock back three months, Yangtze Memory might have faced a different scene.
On July 27, domestic DRAM leader Changxin Technology officially listed on the STAR Market at an issue price of 8.66 yuan and closed at 49 yuan on its first trading day, surging 465.82%. Even more exaggerated, Changxin Technology's total market value reached 3.28 trillion yuan that day, overtaking ICBC in one stroke to top the A-share market value rankings, with full-day turnover as high as 141.2 billion yuan.
To grab a ticket to this capital feast, investors practically broke down the door.
Data shows that Changxin Technology's online subscription attracted 9.4288 million valid accounts, with valid subscription shares exceeding 816.9 billion, and the final winning rate was only 0.47%. Even though the issuance size had already set a STAR Market record, it could not stop the enthusiasm of funds. Based on the first-day closing price, winning one lot of 500 shares meant earning 20,170 yuan.
Institutions were also busy: 10,907 accounts managed by 285 offline investors participated in the new-share subscription, with valid subscription shares exceeding 12 trillion, and top quantitative private funds such as JiuKun and High-Flyer also appeared on the allocation list.
Ultimately, Changxin Technology raised about 66.6 billion yuan through the IPO, setting a record issuance size on the STAR Market.
For the Hefei state-owned capital behind it, this was even more a capital feast ten years in the making.
Public information shows that when Changxin Technology was founded in 2016, Hefei state-owned capital was an important founding and supporting force. Since then, the company has continuously received funding support from local industrial capital, the National Big Fund and other investment institutions, gradually building a domestic DRAM manufacturing system.
When Changxin Technology's market value exceeded 3 trillion yuan on its first trading day, the equity held by the relevant state-owned capital systems in Anhui and Hefei also gained trillion-yuan-level market pricing. From starting in 2016 to topping the A-share market value rankings in 2026, the Hefei industrial capital behind Changxin Technology did catch a good meal.
But the capital market's dinner table turns faster than anyone. From July 1 to October 22, the STAR 50 Index fell 40.85% cumulatively, while the ChiNext Index fell 40.66% and the Shanghai Composite Index fell 10.66% over the same period.
Memory concept stocks, which had already been rising for more than a year, became the hardest-hit area. During the period, Biwin Storage fell 67.73%, GigaDevice fell 66.04%, Longsys fell 65.62%, and Demingli fell 63.84%.
Changxin Technology naturally could not remain unaffected either. As of October 23, Changxin Technology closed at 42.85 yuan, a drawdown of more than 30% from its August stage high of 61.80 yuan.
In fact, it is not just A-share memory stocks; investors' attitude has already spread to all technology sectors.
In September 2025, Dongshan Precision completed its acquisition of Source Photonics and formally entered the optical module business. In the first half of 2026, its optical module business revenue reached 5.351 billion yuan, accounting for about 19.2% of total revenue and becoming a new source of growth for the company. In addition, it also has high-end PCB, electronic circuits and other businesses, all of which are important links in the AI server industry chain.
Under the market conditions of a few months ago, these stories put together would have been enough to attract considerable capital attention. But by October 20, when Dongshan Precision was soliciting subscriptions for its Hong Kong listing, the market had already changed its face.
The maximum offering price for Dongshan Precision's Hong Kong shares was HK$93.8, but by the end of subscription on October 23, its margin financing subscription ratio was only 0.39 times. Weiyi Zhizao, which closed its subscription in the same period, was oversubscribed by more than 31 times in margin subscriptions.
The reason subscription conditions fell to freezing point is very direct: investors are no longer willing to give money to a technology stock that has continued to fall sharply.
On October 23, the company's A-shares closed at 114.22 yuan, a drawdown of nearly 60% from the 52-week high of 280.08 yuan. For an AI industry chain company preparing to list in Hong Kong, while its home-market stock price keeps falling, it is naturally difficult for investors to muster the same IPO subscription enthusiasm they had a few months earlier.
In addition to Solidigm, the capital operations of enterprises such as Australian data center operator Firmus Grid have also been affected by AI valuation concerns and changes in the market environment.
This has made the situation of semiconductor companies particularly delicate. On the one hand, AI infrastructure construction is still driving memory demand, NAND and DRAM prices are rising, and the operating performance of international memory giants such as Samsung, SK Hynix and Micron remains strong; on the other hand, the capital market has begun to worry that the valuations previously assigned to the AI industry chain may be too high.
For companies that have already listed, a falling stock price can still be followed by waiting for a rebound; for companies that have not yet completed issuance, however, the financing window may keep narrowing while they wait. And Changcun is the one that has not yet taken a seat at the table.
In fact, Changcun's performance alone is already enough to make investors sit up and take notice. The prospectus shows that in 2023 the company's operating revenue was 18.744 billion yuan, with a net loss attributable to shareholders of 19.181 billion yuan; in 2024 revenue grew to 45.203 billion yuan, with net profit of 6.771 billion yuan; in 2025 revenue reached 63.185 billion yuan, with net profit of 14.211 billion yuan. By the first quarter of 2026, the company's revenue reached 47.042 billion yuan, and net profit attributable to shareholders was as high as 33.379 billion yuan.
In just three years, Changcun grew from a company losing nearly 20 billion yuan a year into a global memory giant earning more than 30 billion yuan in a single quarter.
Its industry position is also rising in tandem. According to data disclosed in the prospectus, in 2025 Changcun ranked third in the global NAND Flash market by both sales amount and shipments. Such results, if placed half a year ago, would probably have filled many investors with imagination.
At present, Changcun is still in the queue. According to information disclosed by the Shanghai Stock Exchange, Changcun Holdings' IPO was accepted on August 21 and entered the inquiry stage on September 3. As of October 25, the review status remained "under inquiry" and had not yet entered the listing committee review and registration stages.
Of the 33 billion yuan it plans to raise, 20.8 billion yuan will be used for technology upgrades on mass-production lines, and the other 12.2 billion yuan for research and development. Technology upgrades need funding, capacity expansion needs funding, and next-generation NAND research and development also needs funding. International competitors will not slow their own research and development pace just because A-shares have fallen.
More importantly, IPO valuation is no longer the easy question it was a few months ago.
If Changcun's pre-issuance valuation reaches 800 billion yuan, the new shares corresponding to the 33 billion yuan raised would account for about 4% of post-issuance share capital; if the valuation drops to 400 billion yuan, the proportion would approach 8%.
For the same amount of financing, how much the market is willing to pay may directly affect how much equity existing shareholders need to give up.
Changcun's profits are growing rapidly, but this growth happens to occur during a high-prosperity cycle in the memory industry. Once the listing timetable continues to be pushed back, even if the capital market regains its enthusiasm, it may not still catch the same memory price cycle.
It is almost certain that even though both are among the twin giants of domestic memory, Changxin took ten years to catch a capital feast, while Changcun may need to spend even longer, and by then only leftovers may remain.
However, perhaps more anxious than Changcun is Wuhan, which has accompanied it for nearly twenty years.
In 2006, the Hubei provincial, Wuhan municipal and East Lake High-tech Zone governments jointly invested 10 billion yuan to launch the Wuhan Xinxin 12-inch wafer manufacturing project. At that time, the global memory chip market was still firmly in the hands of international giants such as Samsung, Toshiba, SK Hynix and Micron.
The starting point of Wuhan Xinxin was to build a wafer fab with advanced process capabilities and lay the foundation for the local integrated circuit industry.
However, investment in semiconductor manufacturing is far more staggering than in ordinary industrial projects. A wafer production line often requires billions or even tens of billions of yuan, and equipment such as lithography, etching and thin-film deposition is expensive. After the line is put into operation, it must also continuously face pressure from depreciation, yield and process upgrades.
What is more troublesome is that chip production is not a matter of making money simply by building factories. Technology needs to be accumulated, products need to be validated, and production capacity needs to ramp up. By the time one generation of products is finally mature, international competitors may have already launched the next generation.
Wuhan Xinxin groped its way through this long industrial cultivation period for ten years, and in 2016, the opportunity finally came.
That year, the national memory base project was officially launched in Wuhan Optics Valley, with planned investment of up to US$24 billion, equivalent to about RMB 160 billion.
The project planned to build on Wuhan Xinxin's existing 12-inch wafer manufacturing capabilities and bring in capital from the National Integrated Circuit Industry Investment Fund, Hubei Guoxin Fund, Hubei Science and Technology Investment and Tsinghua Unigroup to build a memory industry system covering chip design, manufacturing, packaging and testing.
In July of the same year, Yangtze Memory Technologies was officially established. For Wuhan, this was not just one more semiconductor company, but finally an opportunity to transform the wafer manufacturing accumulated over the previous decade into a globally competitive memory industry.
However, wresting a share of the market from international giants was not easy.
In 2018, Yangtze Memory released its self-developed Xtacking architecture, attempting to improve memory density and product performance by manufacturing the peripheral circuits and memory cells separately and then bonding the wafers; in 2019, the company announced mass production of 64-layer 3D NAND; in 2020, its 128-layer product was successfully developed; thereafter, it gradually advanced Xtacking 3.0, 4.0 and higher-generation NAND products.
Today, these years look like just a few technology nodes in Yangtze Memory's development history, but at the time, every step forward meant large-scale R&D investment and a fight with Samsung, SK Hynix, Micron and other companies over technology and cost advantages.
What is more, Yangtze Memory's technological catch-up was accompanied by external supply chain restrictions.
In December 2022, the United States added Yangtze Memory to the Entity List, and access to advanced semiconductor equipment, components and technology faced stricter restrictions. This meant that a semiconductor company that already required continuous heavy investment also had to bear the extra costs of supply chain re-adaptation and equipment localization.
However, before all the technical problems were solved, capital problems came knocking first.
In the early days of Yangtze Memory's establishment, Tsinghua Unigroup was an important investment and industrial integration entity, and its first chairman, Zhao Weiguo, also tried to use large-scale capital operations to build a semiconductor industry map covering memory, chip design and manufacturing.
Unfortunately, before Yangtze Memory could reach its profit peak, Tsinghua Unigroup fell into a debt crisis first. In 2021, Tsinghua Unigroup entered bankruptcy reorganization proceedings. This crisis left Yangtze Memory's equity relationships and future financing arrangements facing new variables.
In the end, Hubei local capital stepped forward again. Under Tsinghua Unigroup's reorganization arrangement that year, Hubei Science and Technology Investment acquired the relevant Yangtze Memory project assets for RMB 5.1 billion in cash and assumed the corresponding guarantee liabilities.
After that, Hubei local capital continued to increase its stake. In February 2023, Yangtze Memory's registered capital increased from RMB 56.275 billion to RMB 105.27 billion, an increase of nearly 87%. Among them, Hubei Changsheng subscribed RMB 30.069 billion, leaping to become the single largest shareholder of Yangtze Memory.
From initially participating in factory construction, to taking over during the Tsinghua Unigroup reorganization, to continuing to increase capital in 2023, local state-owned capital has run through almost all of Yangtze Memory's most important development stages. This also means that Yangtze Memory's listing will directly affect the market value of an important asset held by Hubei local industrial capital.
However, what the devoted Wuhan venture capital investors got was not always good news. In 2020, the blow-up of the Wuhan Hongxin project poured a basin of cold water on local semiconductor investment enthusiasm.
This company, founded in 2017, had planned to invest RMB 128 billion to build an advanced-process chip production line and had brought in Chiang Shang-yi, who had worked at TSMC and SMIC, as an executive.
On one side was a hundred-billion-yuan investment plan, and on the other was international semiconductor industry talent; it looked somewhat like an attempt to replicate a chip giant.
However, by 2020, Wuhan Hongxin was exposed as having serious funding problems. According to the major project construction plan previously disclosed by Wuhan, by the end of 2019, the Hongxin project had completed cumulative investment of RMB 15.3 billion. But in the second half of 2020, when National Business Daily conducted an on-site visit, it found that construction of the plant had stalled and some subcontractors' project payments could not be settled for a long time. The hundred-billion-level project that had initially been highly expected ultimately failed to land as planned.
For Wuhan, the lesson left by Hongxin was especially heavy: the semiconductor industry needs patient capital, but huge investment may also fail to form real production capacity for a long time because of the project's own operational and governance problems.
While the Hongxin project was stalled, Yangtze Memory was still struggling to advance technological iteration.
Wuhan could only continue to bet, but even if a company truly produces products, reaching the capital market does not mean the story will go smoothly. This year's IPO experience of Wuhan Xinxin illustrates the problem well.
In September 2024, Wuhan Xinxin submitted an application to the Shanghai Stock Exchange for a STAR Market listing, planning to raise RMB 4.8 billion. According to the company's business positioning, Wuhan Xinxin mainly focuses on specialty memory, mixed-signal and 3D integration, and has strong competitiveness in the NOR Flash wafer foundry field.
This is a semiconductor company with years of technological accumulation, actual production lines and a business foundation. Yet after a year and a half of review, on May 19, 2026, Wuhan Xinxin and its sponsor chose to withdraw the IPO application, and the Shanghai Stock Exchange subsequently terminated the review. Coincidentally, on the same day, Yangtze Memory Holdings officially launched IPO tutoring filing.
Next, Wuhan's semiconductor assets underwent another round of adjustment. In June, Changcun Holdings planned to transfer its 39% stake in Wuhan Xinxin to a Guanggu state-owned capital platform, driving an adjustment of control over Wuhan Xinxin. By September, the relevant equity transfer was formally completed. Hubei Science and Technology Investment, through platforms such as Guanggu Semiconductor under its umbrella, regained control of Wuhan Xinxin.
After this adjustment, Changcun Holdings continued to push forward its STAR Market IPO, while Wuhan Xinxin remained within the Guanggu local state-owned capital system, developing its specialty-process wafer foundry business.
For Wuhan, the business layouts of the two enterprises are gradually becoming clear, but the capitalization process still requires waiting.
From Wuhan Xinxin's filing in 2024, to the withdrawal of its application in 2026, and then to Changcun Holdings taking over the IPO baton, the process of listing Wuhan's semiconductor assets has yet to see an endpoint.
On March 19 this year, Voyah, also rooted in Wuhan, officially listed on the Hong Kong Stock Exchange. This high-end new energy vehicle brand under Dongfeng saw sales exceed 150,000 units in 2025 and achieved a full-year net profit of RMB 1.017 billion, finally emerging from its previous phase of consecutive losses.
But on its first day of trading, Voyah's share price still fell 13.2%, leaving its market value at only about HKD 24 billion. Previously, Dongfeng Group shares delisted through privatization, making Voyah a new independent listing platform in an attempt to give new energy assets clearer market pricing, yet the price investors offered was clearly not as enthusiastic as the company and shareholders had hoped.
From the halt of Hongxin, to Wuhan Xinxin's withdrawal of its IPO, and then to Voyah's cold reception on listing, the twists and turns Wuhan has experienced in major industrial projects over the years are enough to write another story of the capital markets.
Compared with Hefei, which in recent years has frequently attracted capital market attention through projects such as BOE and Changxin Technology, Wuhan's industrial accumulation is similarly deep, but the process of converting long-term investment into public market value appears more tortuous.
Today, Changcun may be Wuhan's most anticipated card. It is no longer the semiconductor startup that once worried about survival, yet the capital markets have chosen precisely this moment to begin receding.
For Changcun, listing can continue to raise funds for next-generation technology and production line upgrades; for Wuhan, it carries another layer of meaning.
Over the past two decades, local capital has continuously invested in the semiconductor industry. From the earliest Wuhan Xinxin, to the national memory base, and then to the equity takeover after Tsinghua Unigroup's restructuring, capital has continued to settle into production lines, plants, technology and equity. If Changcun can list smoothly, these long-term industrial assets will also reach an important capitalization milestone.
A company's listing may also open new funding space for the next round of Wuhan's industrial investment, but at present, what lies before Wuhan is no longer simply whether to list.
Continuing to push forward with the offering may mean facing pressure on financing and equity dilution brought by lower market valuations; if it waits for the market to recover, the storage industry's high-profit cycle may not last until then, because what the storage industry never lacks is cycles.
In 2023, Changcun was still losing nearly RMB 20 billion; in the first quarter of 2026, net profit exceeded RMB 33 billion. No one can guarantee that today's boom and high profits will continue indefinitely.
Moreover, Changcun's next round of technology investment cannot stop.
When Wuhan invested in building the plant back then, international storage giants were several product generations ahead of Chinese companies; today Changcun has ranked among the global top three, yet still has technological and scale gaps to close with competitors such as Samsung and SK Hynix.
The closer a company gets to the global first tier, the less likely the next round of investment can be reduced, which makes Wuhan's position seem particularly delicate.
Back then, when the industrial foundation was weak, it needed Wuhan to put up money; later, when shareholders faced a debt crisis, it needed Wuhan to take over; now that the company finally has global competitiveness, the capital markets have begun to consider it too expensive.
For SK Hynix, Solidigm's suspension of its IPO is merely deferring a financing plan to a later date, but behind Changcun lies nearly twenty years of waiting for Wuhan. Building the wafer fab in 2006, founding Yangtze Memory in 2016, undergoing Tsinghua Unigroup's restructuring in 2021, local state-owned capital continuing to increase its stake in 2023, and finally reaching the IPO inquiry stage in 2026.
Hefei waited ten years for a capital feast, and Wuhan has waited twenty years and seems about to reach the table, only that when it is Changcun's turn, the diners have all lost their appetites.
-END-Channel: Finance and Economics. This content is published with the author's authorization; the views represent only the author and do not represent Huxiu's position. If you have objections or complaints about this manuscript, please contact tougao@huxiu.com. Everyone who is changing and wants to change the world is on the Huxiu APP[ ]