A Decade of Support Yields Trillions: The Biggest Winner of Hefei Growth Xin's IPO

Original Title: "A Decade of Support Yields Trillions! Uncovering the Biggest Winner of Changxin's IPO------Hefei"
Original Author: Xu Chao, Wall Street Insight

A city bets on an "impossible" project, suffers losses for ten consecutive years, burns through 36.6 billion, and finally reaps a trillion-dollar return------this is not a novel, this is Hefei.

On July 27, 2026, Changxin Technology (688825) officially landed on the STAR Market, closing at 49 yuan, a staggering increase of 465.82% from the issue price, with a market value surpassing 3.2 trillion yuan, overtaking Industrial and Commercial Bank of China to become the "king of market value" in A-shares. This largest DRAM chip manufacturer in China and the fourth largest globally, after ten years of establishment, stood under the spotlight of the capital market.

Behind this enterprise, the city that has supported it for a decade is quietly settling a historic bill. Based on Hefei's state-owned assets system holding approximately 36.79%, the corresponding market value has exceeded 1.2 trillion yuan. With the massive market value increase brought by Changxin Technology, Hefei's total market value in A-shares has surpassed 4 trillion yuan, making it the second city in the Yangtze River Delta in terms of A-share market value------a "gamble" from a central provincial capital city has written the most shocking footnote in the history of industrial investment in China.

Two Decades of Commitment from One Person

To understand why Hefei dared to invest, one must first understand the person Zhu Yiming.

Zhu Yiming, from Yancheng, Jiangsu, entered Tsinghua University in 1989, and after obtaining his master's degree, went to the United States for further study, transitioning to the semiconductor field at Stony Brook University. After graduation, he entered Silicon Valley, working as a project manager at a memory chip company.

There, he saw an alarming fact: memory chips are the most consumed and standardized category of semiconductors, the "food" for almost all electronic devices, yet Chinese players have long been absent in this arena.

In 2004, he made a life-changing decision------to resign and return to China to start a business. The initial funding was 920,000 USD, raised by several Tsinghua alumni. After the Spring Festival in 2005, he founded what would later become Gigadevice in a two-story unfinished building in Tsinghua Science Park. Instead of directly competing with giants like Samsung and SK Hynix, he entered the "niche" market of NOR Flash to complete his initial accumulation. In 2016, Gigadevice successfully went public.

But Zhu Yiming's ambition goes far beyond this. He once said: "If we compare a computer to a crown, the CPU is the jewel on the crown, and the memory is the base of the crown." "Whoever leads memory technology can dominate the entire integrated circuit industry."

Creating a Chinese version of "Samsung Electronics" has been his ultimate goal since day one of his entrepreneurial journey.

In 2016, the opportunity arose.

Hefei Took on the 'Deadlock' Others Dared Not Touch

At that time, 96% of the global DRAM market was firmly held by the three giants: Samsung, SK Hynix, and Micron, with China's self-production capacity almost zero. The DRAM industry has high demands for capital, talent, and technology, and its strong cyclical nature leads to severe price fluctuations, making losses almost an unavoidable "entrance tax."

Hefei, at that time, was not wealthy. But Hefei decided------to go for it.

This led to the strategic project later named "506": a total investment of approximately 150 billion yuan for Changxin's 12-inch memory wafer manufacturing base. The first phase had a total investment of 18 billion yuan, with Hefei Industrial Investment contributing 14.4 billion yuan, accounting for a staggering 80%. This amount in 2016 was almost equivalent to selling everything they had.

What is even more commendable is that Hefei chose to be a "true supporter."

During the darkest times when Changxin Technology faced years of losses, accumulating over 36.6 billion in deficits, Hefei's state-owned assets did not retreat or withdraw their investment. Even when other investors exited at the end of 2024, Hefei proactively contributed nearly 2 billion to take over old shares. A responsible person from Hefei Industrial Investment once bluntly stated the essence of this logic:

"In weak links of the chip industry chain, the probability of achieving capital returns in the short term is very low; it must be large capital, long cycles, and even cross several cycles to ultimately realize value investment."

This is not a gamble; it is a profound understanding of industrial rules by a city and a clear judgment of the national strategic needs.

China's DRAM: A Dangerous Breakthrough from Zero to One

The path Changxin has taken is far more perilous than outsiders imagine.

Fujian Jin Hua, established at the same time as Changxin, halted its hundreds of billions investment project on the eve of mass production due to a lawsuit from Micron for stealing trade secrets. Changxin chose another path------through legal negotiations, at the cost of "hundreds of millions of dollars," obtained over 10 million DRAM technical documents, 2.8TB of core data, and numerous implementation licenses for Infineon's DRAM technology from the bankrupt German storage giant Qimonda.

In 2018, Zhu Yiming made a shocking decision in the capital market: he resigned as general manager of Gigadevice to fully take on the role of chairman and CEO of Changxin Technology, pledging------not to take a penny in salary or bonus until the project is profitable.

In September 2019, a year later, Changxin Technology launched its self-designed and produced 8Gb DDR4 chip, marking a historic breakthrough of the DRAM industry in mainland China from "zero to one."

But "from zero to one" is just the entry ticket. The real test arrived in 2023.

That year, global DRAM prices plummeted over 40%, and shipments of smartphones and PCs declined, plunging the industry into a deep downturn. The three giants leveraged their cost advantages to implement a "counter-cyclical" strategy, maintaining high shipment volumes and further squeezing new players.

Changxin lost money on every chip sold, yet continued to accelerate the breakthrough of 1x nm process technology and overcome key technical barriers for DDR5 mass production. The company recorded a loss of 16.34 billion that year, the highest since its establishment, with a cumulative deficit of 36.65 billion over ten years.

Any commercial entity would have exited at this point.

But Hefei did not. During Changxin's years of losses, Hefei's state-owned assets chose to invest repeatedly, provide resources, and supply ammunition.

That year, the Hefei Municipal People's Congress Standing Committee reviewed and approved a capital increase and expansion proposal. By the end of 2024, Country Garden Venture Capital exited, and Hefei's state-owned platform contributed nearly 2 billion to take over old shares without any hesitation.

Behind this, Hefei established a systematic error-tolerance mechanism: projects must be reviewed by the Financial Committee of the People's Congress, and major decisions must be voted on by the Standing Committee. As long as due diligence is compliant and procedures are in place, even if the project ultimately incurs losses, decision-makers do not bear personal responsibility. It is reported that Hefei has never penalized any unit or individual for failed industrial investments.

It is precisely this "affordable loss" institutional guarantee that allowed Hefei to become truly patient capital while other cities hesitated.

Earning 400 Million Daily, Filling a Decade of Losses in a Quarter

The turning point quietly arrived in 2025.

The demand for AI computing power completely ignited a super cycle in storage. An AI server uses 3 to 5 times more DRAM than a traditional server, while Samsung, SK Hynix, and Micron shifted their production capacity to more profitable HBM, significantly widening the supply gap for conventional DRAM.

Changxin Technology just completed the product iteration from DDR4 to DDR5, with the capacity utilization rate of three 12-inch wafer fabs steadily increasing from 85% to 95%. Surging demand, shrinking supply, and released capacity------three favorable factors combined to create a textbook-level "Davis Double Play."

In 2025, Changxin Technology achieved annual profitability for the first time, with a net profit attributable to the parent company of 1.875 billion yuan.

In the first quarter of 2026, revenue reached 50.8 billion yuan, with a net profit attributable to the parent company of 24.762 billion yuan, a year-on-year increase of 1688%. This translates to nearly 400 million yuan earned daily. At this pace, in less than half a year, Changxin had almost filled all the losses from the previous decade.

At this moment, Hefei's initial investment of 14.4 billion yuan, along with continuous follow-up investments over the past decade, finally realized its value.

Beyond the Trillion Floating Profit: The Industrial Restructuring of a City

The 1 trillion on paper is just the tip of the iceberg of Hefei's returns.

Ten years ago, the area where Changxin's factory is located in the northwestern suburbs of Hefei was still a mix of farmland and wasteland. Today, massive gray and white factories stretch for hundreds of meters, with dense silver ducts, corridors, and industrial pipelines intertwining in the air; around the factory, research buildings, employee apartments, canteens, commercial centers, fast food restaurants, and supermarkets have opened one after another, humorously dubbed the "Changgang CBD."

By the end of 2025, Changxin Technology's total number of employees had reached 19,300, with over 6,000 in research and development, mostly aged between 25 and 35, and the majority holding master's degrees or higher. These young, highly educated, and strong purchasing power employees are fundamentally changing the consumption structure and urban atmosphere of the surrounding area.

The changes at the industrial chain level are even more profound.

Relying on the leading effect of Changxin Technology, Hefei has gathered over 450 integrated circuit companies, forming a complete industrial chain from design, manufacturing to packaging and testing, becoming one of the few cities in the country with a full integrated circuit industrial chain. In 2016, the output value of Hefei's integrated circuit industry chain was only about 18 billion yuan; by 2025, this figure had reached 151.4 billion yuan, a growth of 7.4 times.

More noteworthy is the industrial synergy effect. Changxin's memory chips, along with BOE's panels, NIO's and BYD's electric vehicles, together constitute Hefei's industrial landmark of "chip-screen-vehicle integration," forming an interdependent and deeply integrated industrial ecosystem------Jinghe Integrated Circuit Co. produces panel display driver chips for BOE, while Jiefa Technology supplies automotive-grade MCU chips to BYD and NIO, accelerating the formation of the internal circulation of the industrial chain.

Why Others Cannot Replicate the 'Hefei Model'

After Changxin Technology's IPO, external attention once again turned to the "Hefei Model." However, in reality, about 50 investigation teams flood into Hefei each month, writing millions of words in research reports, yet they have never managed to cultivate a truly replicable sample.

Hefei itself is clear: this model has four prerequisites, and none can be missing.

Sufficient financial backing. In 2008, Hefei invested 6 billion in BOE, equivalent to 20% of its fiscal revenue that year. The Changxin project bore the pressure of 366 billion in losses over ten years. Without corresponding fiscal maneuvering space, this question cannot be answered at all.

A robust error-tolerance mechanism. Hefei was the first in the country to establish a "duty exemption" system, where project decisions only need to comply with procedures and due diligence; even if there are losses, decision-makers do not bear personal responsibility. The local government has never penalized any unit or individual for investment failures, which is the institutional confidence to "dare to invest."

Accurate industrial judgment. Every move Hefei made occurred at the coldest moments in the industry------BOE during the global panel industry losses, NIO when its stock price fell to 1 USD and 18 cities turned it away, and Changxin when there were no Chinese players in the global DRAM market. This counter-cyclical layout relies on years of systematic judgment of industrial trends, rather than chasing fads.

A wide policy window. Hefei has caught the golden decade of China's manufacturing industry climbing from low-end to mid-to-high-end. The demand for domestic substitution is real and urgent. As pointed out by the macroeconomic team led by Song Xuetao at Guojin Securities, Changxin "caught up" with the overlapping demands of domestic substitution, storage security, and AI expansion, which itself indicates the foresight of national strategic planning.

The Paradigm Shift in Urban Development in China

Behind Changxin Technology's IPO, a more macro proposition is emerging: beyond land finance, urban development needs a new engine.

Hefei's path provides an answer: use state-owned assets as early capital, then leverage the capital market for amplification, constructing a system that can continuously produce good companies. From 2015 to 2021, during the real estate boom, Hefei's total land transfer revenue was approximately 551.6 billion yuan; whereas just from Changxin Technology alone, the market value of Hefei's state-owned assets has approached 1 trillion yuan.

This is not just Hefei's story, but a paradigm shift in the logic of urban competition in China: from "attracting investment" to "cultivating industries," from "land finance" to "equity finance," from "planting a big tree" to "growing a forest."

The bell ringing for Changxin Technology's IPO is the final answer to Hefei's decade of patient capital and a question that more cities cannot avoid.