Original author: Jia Liu
When it comes to the global storage industry, many people only know Samsung, SK Hynix, and Micron. But what many do not know is that the fourth largest company in the world is a Chinese enterprise that is yet to go public, named Changxin Memory Technologies (CXMT).
On July 9, Changxin Technology updated its latest IPO prospectus for the Science and Technology Innovation Board. It is scheduled to open for subscription on July 16, with payments due on the 20th, and it is hoped to begin trading on the Science and Technology Innovation Board by the end of the month.
The strongest storage leader in China is coming by the end of this month.
Data shows that revenue for the first quarter of 2026 is expected to reach 50.8 billion yuan, a year-on-year increase of 719%. Net profit is projected to be 33 billion yuan, averaging nearly 400 million yuan per day, surpassing the profitability of Kweichow Moutai. The market has given it a valuation of one trillion yuan, poised to become the first stock on the A-shares market by market capitalization.
However, as of the end of 2025, the company's accumulated losses are still 36.65 billion yuan.
This means that Changxin Memory has recovered nine years of losses in just half a year. This domestic chip company, which had suffered nearly a decade of continuous losses, has suddenly transformed into one of the most profitable hard-tech companies in the A-share market.
In the past week, the name Changxin Memory has frequently appeared in global technology media. Apple is lobbying the U.S. government for a special permit to include Changxin Memory in the memory supply chain for Mac and iPad. Google has also initiated an evaluation for procurement of Changxin DRAM, and reports have mentioned that HP and Dell are validating Changxin Memory DRAM, while Acer and Asus are urging Chinese partners to use more local memory chips. In the same week, Reuters revealed that Tencent signed a long-term supply agreement for server DRAM with Changxin Memory worth over 20 billion yuan, valid for three to five years. The client list disclosed in the Changxin prospectus also includes Alibaba Cloud, ByteDance, Lenovo, Xiaomi, OPPO, Vivo, and Honor.
Changxin, a leading enterprise in the global memory production field, ranks fourth globally after Hynix, Samsung, and Micron. Source: Reuters
For a time, Changxin Memory has become a favorite among almost all major domestic and international companies.
What has allowed Changxin to rise from zero to the fourth position globally? How much of the money it earns today is due to strength, and how much is luck? Can its trillion-yuan valuation hold?
The Path Forward for Changxin Memory: Zhaoyi Innovation
The story of Changxin Memory's beginnings begins with why China has always lacked DRAM.
There are many types of memory chips. NAND Flash is a type of storage that can retain data even when power is off, used in phone storage, SSDs, and USB drives; DRAM is the memory used for running, where data disappears when power is lost, but it is fast and operates as the workspace for CPUs, GPUs, mobile SoCs, and AI accelerator cards.
Both NAND and DRAM are bulk semiconductors, but DRAM is more like a steelmaking plant that requires high-precision manufacturing. Each generation of manufacturing processes must push capacitors, transistors, word lines, and bit lines to their limits while replicating billions of highly consistent cells on a single wafer. If there is any slight deviation, the yield drops and costs explode. When costs explode, during a memory downturn cycle, money flows out like water.
This is why the DRAM market later only had three giants: Samsung, SK Hynix, and Micron. Because this game is not easily entered by simply saying "I have a technology team," but rather requires enduring price wars, capacity battles, patent wars, equipment restrictions, and customer certifications over a cycle of more than ten years.
Changxin Memory started around 2016 and was established in Hefei. Its key figure is Zhu Yiming.
Zhu Yiming at a conference
The name Zhu Yiming is not unfamiliar in China's semiconductor circle.
Born in 1972, Zhu Yiming was a child prodigy, entering Tsinghua University at the age of 17 to study physics. Although his major was physics, he was highly skilled in programming, making over 300,000 yuan in a single year in the late 1990s by writing programs for other companies. Through his work, he learned that at that time, all chips were designed in the United States and believed that chip technology held higher value, so he went to the U.S. to study electronic engineering and graduated to work in Silicon Valley.
Zhu Yiming in his early years
In Silicon Valley, Zhu Yiming noticed that the storage industry had shifted from the U.S. to Japan, then to South Korea and Taiwan, and he predicted a future opportunity for mainland China, where it was also possible to create a "Chinese version of Samsung." He designed an SRAM (Static Random Access Memory) chip and returned to China in 2005 to establish a company called Zhaoyi Innovation.
Zhaoyi Innovation has been a pure design company from its inception, not producing chips itself and outsourcing all wafer manufacturing to foundries. Its main products are NOR Flash (storage chips for routers, game consoles' boot programs) and MCUs (microcontrollers for refrigerators, washing machines). Zhu Yiming and his team created China's first static memory and IP technology, the first serial flash product, and the first 32-bit general-purpose MCU based on the ARM Cortex-M3 architecture. By the time it went public in August 2016, it had already become the largest design firm for code flash memory chips in mainland China, with revenue growing from 1.489 billion yuan in its debut year to 9.203 billion yuan.
Light assets, high gross margin, without the need to build factories. This was the smartest way in China's IC design industry at that time.
But this model had a prerequisite: that foundries had adequate capacity and stable prices. In the second half of 2020, Zhu Yiming encountered a hurdle. SMIC's 8-inch capacity became severely constrained, forcing Zhaoyi Innovation to transfer some NOR Flash orders to Huahong Semiconductor’s 12-inch factory, incurring higher foundry costs. By the fourth quarter of 2020, the gross margin plummeted from a stable above 37% to 29.49%. The consequences of foundry capacity constraints were directly reflected in the profit statement.
During the same period, Zhaoyi Innovation was also dealing with another acquisition purchase. In 2019, it acquired Shanghai Silan Microelectronics for 1.7 billion yuan, at a premium of 16 times, resulting in a goodwill of 1.305 billion yuan. Silan was making under-screen fingerprint chips, and Zhu Yiming aimed to integrate the storage, controller, and sensor lines to form a platform design company. However, it was sued by Goodix Technology for patent infringement, and both sides became trapped in a price war, leading Silan to achieve only 58% of its performance commitment over three years. From 2020 to 2023, Zhaoyi Innovation recognized goodwill impairment for Silan for four consecutive years, totaling around 900 million yuan, diminishing the goodwill from 1.3 billion to a fraction. The path of horizontally acquiring another design company to expand categories was proven to be unworkable.
The lessons Zhu Yiming learned from these two incidents were: the ceiling for light-asset design companies does not lie in design capability, but in capacity. And the main battlefield for memory chips is not in NOR Flash, which only occupies 2.5% of the global market, but in DRAM. However, DRAM cannot be outsourced to general foundries for production like NOR Flash. The processes for DRAM are highly proprietary, with each manufacturer’s unit structure, capacitor design, and word line processes being custom-made. Samsung, SK Hynix, and Micron operate under a design and manufacturing integrated model. To enter the DRAM business, there is only one path: to build the factory oneself.
This is the backdrop and reason for the birth of Changxin Memory.
After BOE and NIO, the Third Big Bet in Hefei
Zhaoyi Innovation's financial statements and shareholder structure could not support the investment scale for building a DRAM factory. After all, phase one cost 18 billion yuan, the total investment exceeds 100 billion yuan, and it had been suffering losses for nearly a decade. This money could only come from another type of capital.
In 2016, the Hefei government extended an olive branch.
Hefei is commonly referred to as "the city that excels at venture capital." The most successful venture capital story in Hefei is BOE, which, being the world's largest LCD panel manufacturer with annual revenues exceeding 200 billion yuan, achieved a 70% market share in the global LCD market from Hefei.
In September 2008, BOE signed a framework investment agreement for a sixth-generation production line with Hefei. The total project investment was 17.5 billion yuan, with registered capital of 6 billion yuan funded by Hefei, which was injected through BOE's targeted share issuance, fully invested into the project company. Rather than merely providing a subsidy to BOE, Hefei leveraged local platform control, advanced capital, provided interest subsidies on loans, and matched energy and land provisions to support BOE through its toughest construction phase. Hefei continued to attract associated industries such as glass substrates, polarizers, drive chips, and equipment materials around BOE, earning the city the label of "New Display Capital."
The second typical story is NIO. In April 2020, NIO had just emerged from its life-and-death struggle in 2019, and Hefei’s construction investment and emerging industry investments made a total of 7 billion yuan in cash investments for NIO China, which then established its headquarters in the Hefei Economic and Technological Development Area. Later, NIO became one of the representative brands of high-end electric vehicles in China, reaching a market valuation that once exceeded 100 billion dollars.
Find strong asset leaders, replace simple subsidies with transaction structures, support companies through their construction periods, and then use the leaders to drive their supply chains to root locally.
By 2016, Hefei used the same investment logic to continue its next big bet.
Yuan Fei, a core figure in Hefei’s Industrial Investment, later said: "Whether it is Hefei Industrial Investment or the industry, everyone who participated in Changxin early on undertook unprecedented risks and pressures."
After all, DRAM is harder than panels and complete vehicles, with higher risks related to technology, equipment, yield, patents, and export controls. But from the perspective of regional industrial organizations, it is like BOE and NIO; while it involves short-term cash outflows, if successful in the long term, it can greatly change the supply-demand relationship of an industry chain.
Advancing the integrated circuit industry in Hefei
Hefei Industrial Investment contributed 14.4 billion yuan to the first phase of Changxin Memory, with a total project investment exceeding 100 billion yuan.
In July 2018, Zhaoyi Innovation announced that Zhu Yiming resigned from the position of General Manager, retaining only the Chairman role, and officially took up the position of Chairman and CEO of Changxin Memory. This is known in the industry as Zhu Yiming’s "second entrepreneurship."
The current shareholder list already indicates the stakes in this gamble. The top five shareholders are Qinghui Electric at 21.67% (fully controlled by Hefei State-owned Assets), Changxin Integrated at 11.71%, the Great Fund Phase II at 8.73%, Hefei Jixin at 8.37%, and Anhui Provincial Investment at 7.91%. Hefei State-owned Assets holds over 36% in total.
Among the other shareholders, Alibaba Cloud holds 3.85% (investing 6.1 billion yuan for a valuation of 158.4 billion yuan), and Zhaoyi Innovation holds 1.8%. The list of investors also includes the National Adjustment Fund, China International Capital Corporation, Junlian Capital, China Merchants Capital, Yunfeng Fund, Tencent, and Alibaba.
Both money and momentum are in place. But the "second entrepreneurship" of Zhu Yiming still lacks one thing: core technology.
The Technical Legacy of a Bankrupt German Company
Around 2016, Samsung, SK Hynix, and Micron collectively controlled over 90% of the global DRAM market share, a result of over a decade of elimination tournaments.
The DRAM industry has been slashing players every few years since the 1980s.
Japan once accounted for over 80% of global DRAM capacity, and by the 2010s, only Elpida was left struggling, eventually acquired by Micron in 2012. Europe once had Qimonda, a subsidiary of Infineon, but it went bankrupt in 2009.
Zhu Yiming's solution lay in this German company that had already declared bankruptcy in 2009.
The name Qimonda comes from two words. "Qi" means "energy" in Chinese, and "monda" means "world" in Latin. The direct meaning: the key to unlocking the world.
This name is beautiful. The fate of this company was tragic.
Qimonda was spun off from Infineon in May 2006 and then listed on the NYSE on August 9 of the same year under the ticker QI. By the time of its IPO, it was already one of the major suppliers of memory products globally and a leader in 300mm wafer manufacturing technology. In 2008, Qimonda had completed the development of a 46nm stacked process product based on Buried Word Line technology, boosting production capacity by 100% compared to the previous generation's 58nm process, just waiting for mass production.
Qimonda DDR2/GDDR chip products
But then the financial crisis hit. DRAM prices plummeted, and Samsung expanded production even at a loss to crush its competitors. Qimonda's new technology had yet to go into mass production when its funding chain broke. It declared bankruptcy in 2009. The last light of Europe’s largest memory manufacturer went out. The Munich R&D center was empty, with 12,000 employees scattering, absorbed by Samsung, Micron, and SK Hynix. In 2012, the bankruptcy administrator of Qimonda began selling 7,500 patents.
Zhu Yiming saw his opportunity in this history.
Qimonda perished due to cycles, not due to technology. The 2.8TB of technical documents and tens of thousands of patents left behind are a legacy that had been buried for nearly a decade. Today’s DRAM products from Samsung, Micron, and SK Hynix can actually trace their roots back to Qimonda's buried word line and cell capacitor structures, which indirectly flowed into the three giants' process systems through partnerships with firms like Winbond and Nanya.
However, obtaining this legacy and using it to build a factory presents significant challenges, especially with a bloody lesson from experience.
For China, 2016 was the "first year" of a resurgence in memory. Before this, China’s DRAM industry had completely declined, facing a technical monopoly by foreign firms from the U.S. and South Korea, with no countermeasures.
At that time, when China launched its memory national team, aside from Changxin, there were two parallel routes: Yangtze Memory Technologies was focused on NAND Flash, and Fujian Jinhua Integrated Circuit was developing DRAM.
In 2017, Micron simultaneously sued UMC and Jinhua in the U.S. and Taiwan, accusing three employees who had moved from Micron to UMC of stealing Micron’s DRAM trade secrets, with one person accused of stealing over 900 technical documents. In October 2018, the U.S. Department of Commerce placed Fujian Jinhua on the entity list for national security reasons, implementing export controls. UMC immediately announced a suspension of technical cooperation with Jinhua. Equipment supply was cut off, technical cooperation frozen, and the project was halted. The U.S. Department of Justice simultaneously filed criminal charges against Jinhua and UMC for economic espionage, facing potential penalties exceeding 20 billion dollars. It wasn’t until the end of 2023 that Micron and Jinhua finally reached a global settlement, which had spanned six years.
The lesson from Fujian Jinhua is clear: after being embroiled in a commercial secret dispute with Micron, it was placed on the entity list, leading to equipment supply cuts and halted projects. If a new DRAM company steps on an IP landmine, commercial clients will hesitate to use it, equipment suppliers will hesitate to supply it, and the international market could be killed by litigation.
Therefore, when the bankruptcy manager of Qimonda began selling its 7,500 patents, Zhu Yiming quickly acted, acquiring over ten million files of DRAM technical documents (around 2.8TB of data), during which there were already 16,000 patent applications. He then spent about 2.5 billion dollars on a comprehensive redesign of the original architecture. He advanced Qimonda's 46nm process all the way to the 10nm level. He also signed patent licensing agreements with WiLAN subsidiary Polaris and the U.S. company Rambus, covering about 5,000 U.S. patents and applications.
Initially, among the three storage routes in China, Hefei Changxin was the least favored compared to Yangtze Memory Technologies, which had the backing of Tsinghua Unigroup, and Fujian Jinhua, which was closer to Taiwan and had comparatively advanced technology. In comparison, Hefei had no advantages.
But nobody expected that Changxin's production ramp-up speed would be so quick, a rarity in the entire semiconductor industry.
Founded in 2016, within 14 months it completed the construction of a 12-inch wafer fab. In January 2018, the first phase of the fab was completed, and equipment installation began. By the end of 2018, the 19nm 8Gb DDR4 engineering samples were produced. In the first half of 2019, testing wafers of 15,000 pieces were completed.
On September 20, 2019, at the World Manufacturing Conference, Changxin announced the formal mass production of 8Gb DDR4 chips using the 19nm process. This was China’s first self-developed DDR4 memory chip. The first phase design capacity is 120,000 wafers per month, with total investment of about 150 billion yuan, the largest single industrial project in Anhui Province. In December of the same year, LPDDR4X was mass-produced and shipped.
The leader of the expert group of the National Major Special Project 01 and Director of the Microelectronics Institute at Tsinghua University, Wei Shaojun, commented: This marks the realization of mass production technology breakthroughs for China in the memory chip field, obtaining independent production capacity for this critical strategic component. This has never happened before.
However, mass production is not the end. DRAM customers will not place orders based on a single demo; mobile phone, PC, and server manufacturers need to validate stability, compatibility, longevity, temperature, voltage, and failure rates, and also assess consistency in mass production. The most challenging phase for Changxin was from 2020 to 2023: capacity ramp-up, customer validation, process iteration, all while enduring continuous losses and facing escalating U.S. export controls.
Changxin LPDDR5X and DDR5 display samples
By 2025, Changxin had completed the mass production of process technology platforms from the first to the fourth generation. The proportion of high value-added products like DDR5 and LPDDR5 significantly increased. In November, new DDR5 chips and modules were launched, with peak speeds of 8000Mbps and maximum single-chip capacities of 24Gb, with industry evaluations approaching the high-end products of Samsung and SK Hynix. In 2026, it accelerated the reduction of DDR4 capacity, fully transitioning to DDR5 and LPDDR5, with expectations that over 90% of its production capacity will shift to new processes throughout the year.
Science and Technology Innovation Board IPO, Aiming for the Top Market Value in A Shares
Changxin Technology Group submitted its IPO application to the Shanghai Stock Exchange's Science and Technology Innovation Board on December 30, 2025, marking the first acceptance under the pre-review mechanism for the pilot IPO on the Science and Technology Innovation Board. It plans to raise 29.5 billion yuan, setting a record as the largest proposed fundraising amount for an IPO on the Science and Technology Innovation Board, second only to SMIC's actual fundraising of 53.2 billion yuan in 2020. The sponsors are China International Capital Corporation and CITIC Construction Investment, with Deloitte as the auditor and Shanghai Jintiancheng as the legal counsel.
The IPO process experienced a setback. In March 2026, the review was suspended due to expired financial data. On May 17, the prospectus was updated to resume the review. On May 27, it was approved by the Listing Committee of the Shanghai Stock Exchange and registered on the same day. It took 148 days from acceptance to meeting the review.
Use of proceeds: 7.5 billion yuan for the technical upgrade of the DRAM wafer manufacturing mass production line, 13 billion yuan for the DRAM technology upgrade project, and 9 billion yuan for forward-looking technology research and development projects, with a total project investment of 34.5 billion yuan. The market generally expects it to be officially listed in July or August 2026.
The current valuation anchor is the last round of financing before the IPO: Alibaba Cloud invested 6.1 billion yuan for a 3.85% stake, corresponding to an overall valuation of 158.4 billion yuan.
However, the market's expectations far exceed this figure. Institutions estimate that Changxin's net profit attributable to its parent company will reach between 150 billion to 200 billion yuan in 2026, resulting in a market value corresponding to 3 trillion yuan at a price-to-earnings ratio of 20 times, with some institutions valuing it at over 4 trillion yuan. From 158.4 billion to the trillion range, this span can be seen as a pressure test of global capital on China's storage industry.
Global DRAM revenue market share from 2025 to the first quarter of 2026, with Changxin rising from 3% to 8%. Source: Counterpoint Research
Meanwhile, Changxin's financial curve may also represent the most dramatic reversal in A-share semiconductors.
In 2022, it lost 8.3 billion yuan. In 2023, it lost 16.3 billion yuan. In 2024, it lost 7.1 billion yuan. The capital expenditures over three years were 43.7 billion, 71.2 billion, and 49.7 billion yuan, respectively. By the end of 2025, it had accumulated losses of 36.65 billion yuan.
Then, the memory super cycle arrived. Starting in the second half of 2025, global DRAM demand surged, and prices continued to rise significantly. The National Development and Reform Commission’s price monitoring center indicated that from September 2025 to February 2026, both DRAM and NAND prices hit record highs since 2016. TrendForce data showed that in the first quarter of 2026, DRAM contract prices rose by 93% to 98% quarter-on-quarter. Goldman Sachs raised its 2026 annual increase forecast from 150% to 250%-280%.
Changxin rode this wave of price tsunami. In 2025, its net profit attributable to the parent was 1.875 billion yuan, marking its first turnaround profit. In the first quarter of 2026, it surged with revenues of 50.8 billion yuan, net profit of 33 billion yuan, and net profit attributable to the parent company of 24.8 billion yuan. The company expects revenues for the first half of 2026 to be between 110 billion and 120 billion yuan, with net profit attributable to the parent company between 50 billion and 57 billion yuan.
It earned back nine years of losses in half a year. But looking at it calmly, the primary reason for the current profits remains the cycle.
In the first quarter of 2026, Changxin’s actual shipment volume grew by only about 11%. The average selling price increased by about 57%. The quarter-on-quarter increases in the ASP in the first two quarters were 63% and 68%, respectively. The improvement in profits mostly came from the upward trend of the DRAM price cycle, rather than Changxin catching up to the three giants in terms of costs, yield, and product structure all at once.
This is something all investors must view with caution when considering Changxin's valuation: in today's windfall, the cycle's contribution outweighs structural advancements. Viewing Changxin as a "stable growth stock" may easily overestimate the sustainability of profits. Understanding it as a "new variable emerging in the cycle" is more aligned with the real situation.
However, Changxin still faces many challenges ahead.
Zhu Yiming is 54 years old this year. Since entering Tsinghua at the age of 17, he has been on the semiconductor path for 37 years. Zhaoyi Innovation allowed him to prove that Chinese people can design memory chips, while Changxin Memory enabled him to prove that Chinese people can manufacture memory chips.
Samsung used its loss-making expansion to kill Qimonda during the 2008 financial crisis and forced Elpida to close down through price wars in 2012. When the next downturn cycle arrives, Changxin will sit at the same table, facing more threats.
By that time, what determines Changxin's fate will not be the currently reported earnings of 400 million yuan a day but rather whether its yields are low enough, customer stickiness is strong enough, product line breadth is sufficient, and cash reserves are thick enough.
After all, during the same period, Samsung reported an operating profit of about 260.4 billion yuan, and SK Hynix about 183.6 billion yuan. Changxin's scale still differs by an order of magnitude, indicating significant growth potential remains.
Zhu Yiming is more aware of this than anyone else.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
