Home   |   News   |   Pengsheng Insights

CXMT’s Miraculous Turnaround: From Billions in Annual Losses to Daily Profit of ¥300 Million

 2026-06-01    150


Introduction

May 27, 2026, 9:00 a.m., Shanghai Stock Exchange.

Inside the Listing Review Committee meeting room, members of the SSE Listing Review Committee were leafing through hundreds of pages of review materials. Outside the room, the ChangXin Memory Technologies team held their breath, waiting for the result. Over the previous 18 months, this leading Chinese DRAM memory chip company had undergone one of the most dramatic financial reversals in business history – transforming from a “money‑burning behemoth” with cumulative losses exceeding RMB 36.6 billion and annual spending of tens of billions, into a “cash‑printing machine” that netted over RMB 24.7 billion in a single quarter.

The meeting concluded and the final decision was announced: it meets the issuance and listing conditions, and the review is approved (passed).

This is a Chinese company founded only ten years ago that has already broken into the top four of the global DRAM market from scratch. It carries one of the toughest pieces of the puzzle in China's semiconductor “bottleneck” areas, and has witnessed the trillion‑yuan paper profits behind a decade of “patient capital” accompaniment.

This article will deconstruct the myth of this comeback: from a “money‑burning black hole” with billions in losses to a quarterly profit of RMB 24.8 billion, what tailwinds did ChangXin Memory Technologies catch? And whose bets created this legend?


I

Breaking the Cocoon: Over RMB 30 Billion

in Losses Planted the Seeds of Revolution

The story begins in 2016.

That year, the global DRAM market was firmly held by the three giants – Samsung, SK Hynix, and Micron – together commanding over 90% market share, while domestic DRAM production was almost non‑existent. At that moment, Hefei, Anhui, did something that many at the time viewed as a “high‑stakes gamble” – partnering with GigaDevice to launch the domestic DRAM breakthrough project codenamed “506.” The first phase of the project required a total investment of RMB 18 billion, with Hefei Industry Investment contributing RMB 14.4 billion, or 80%.

At that time, the dangers of this track were well known: a single 12‑inch wafer fab cost billions of yuan, the R&D cycle for a high‑end chip took several years, and R&D spending ran into the billions. The global DRAM market had long been carved up by international giants, and for ChangXin Memory Technologies to carve out a path in the gap between the three behemoths seemed to have no shortcuts.

Over the next three years, ChangXin Memory Technologies validated the market’s judgment that this track was a “money burner.”

Prospectus data shows that from 2022 to 2024, ChangXin Memory Technologies recorded net losses attributable to parent companies of RMB 8.328 billion, 16.34 billion, and 7.145 billion, respectively, with cumulative net losses exceeding RMB 31.8 billion over three years. By the end of 2025, the company’s cumulative losses had reached RMB 36.65 billion. These are figures that would terrify any investor. The rhythm of losing tens of billions a year earned ChangXin Memory Technologies the label “money‑guzzling beast.”

But few noticed that buried beneath the enormous losses was a “revolutionary seed” capable of reshaping the industry landscape.

On the technology front, ChangXin adopted a highly bold “leapfrog R&D” strategy. Instead of following the incremental path of “starting low and moving high” that most Chinese chip companies had taken, it directly skipped several generations of process technology, completing mass production from first‑generation to fourth‑generation process technology platforms. Its product portfolio covers DDR4, LPDDR4X through DDR5 and LPDDR5/5X, with core products and process technologies reaching internationally advanced levels. This effectively crossed the technology gap that global peers had taken a decade to build.

Market share also steadily climbed: in Q2 2025, ChangXin’s global DRAM market share was 3.97%; by Q4, it had grown to 7.67%. Measured by capacity, shipments, and sales, ChangXin ranked first in China and fourth globally. Its three 12‑inch wafer fabs in Hefei and Beijing operate around the clock, with capacity growth rates among the fastest of comparable global companies.

But the speed at which the seed germinated still lagged far behind the pace of mounting losses. It wasn’t until the second half of 2025 that a superstorm from artificial intelligence completely changed the trajectory of ChangXin Memory Technologies.

II

Breakthrough: The Super Cycle

Ignited by AI Computing Power

In the summer of 2025, the hottest word in global tech was neither “iPhone” nor “Windows,” but “computing power.” The large model race drove exponential growth in demand for AI computing power, and DRAM – known as the “short‑term memory” of devices – is the central hub of any computing system.

The DRAM market immediately entered an unprecedented state of supply‑demand imbalance. The three giants – Samsung, SK Hynix, and Micron – all shifted advanced capacity toward the more profitable HBM (High‑Bandwidth Memory) segment, actively compressing conventional DRAM capacity. This strategic tilt opened up a massive market gap for ChangXin Memory Technologies – a crack in a track that had been tightly squeezed by the three behemoths.

Shortages drove prices soaring. Statistics show that global DRAM prices rose continuously from the summer of 2025, with DDR5 prices accumulating a staggering 414% increase. Counterpoint Research data showed that from Q1 2025 to Q1 2026, the average contract price of 1GB consumer DDR4 surged from $1.8 to $11.4, a more than six‑fold increase in one year.

It was precisely in this window that ChangXin Memory Technologies, with years of accumulated mass‑production capability and technological reserves, perfectly rode the wave.

For the full year 2025, revenue reached RMB 61.799 billion, up 155.6% year‑on‑year, with net profit attributable to parent companies of RMB 1.875 billion, turning losses into profits ahead of schedule. Even more astonishing was the Q1 2026 data. The prospectus showed revenue of RMB 50.8 billion, up 719.13% year‑on‑year; net profit of RMB 33.012 billion, up 1268.45%; and net profit attributable to parent companies of RMB 24.762 billion, up 1688.30%. At this rate, it earned about RMB 275 million per day. In contrast, in Q1 2025, the company still posted a net loss attributable to parent of RMB 1.559 billion – a world of difference in just one year.

What does that mean? In terms of single‑quarter net profit attributable to parent, ChangXin Memory Technologies would rank 13th among more than 5,000 listed A‑share companies, ahead of China Life and Sinopec, and even surpassing CATL's RMB 20.74 billion net profit for the same period.

Profitability continues to accelerate. ChangXin Memory Technologies expects revenue of RMB 110‑120 billion in the first half of 2026, up more than six times year‑on‑year; net profit attributable to parent is projected at RMB 50‑57 billion, up more than 20 times. That means in just half a year, it will not only have filled the cumulative loss hole of over RMB 36.6 billion, but also have ample surplus profit.

What underpinned this dramatic reversal – from losing tens of billions a year to earning nearly RMB 300 million a day?

The prospectus gives the core explanation: continued growth in global computing demand, capacity reallocation by major players leading to DRAM shortages, sustained and sharp price increases, combined with the company's growing production and sales scale and continuous product mix optimisation, all drove this explosion.

But the deeper reason is: while the whole industry focused on HBM and other high‑end memories, ChangXin Memory Technologies, with its years of accumulated mass‑production capability and cost control, filled the supply gap in the traditional DRAM market. Industry sources indicate that the Hefei industrial cluster, combined with domestic equipment substitution, gives ChangXin a unit cost 20‑30% lower than Korean manufacturers, giving it a significant price‑competition advantage – that is the true secret behind its transformation from a “money‑burning black hole” to a “cash cow.”

But if we only see that ChangXin Memory Technologies “caught the wave,” we would miss the most moving chapter of this capital story: a decade ago, who made the first heavy bet, with no expectation of return, to build from scratch a track that didn't even exist?

III

The Move: Hefei's Decade‑Long

Accompaniment and the Victory of Patience

In 2016, when the predecessor of ChangXin Memory Technologies began construction of its 12‑inch wafer fab in the Hefei Economic Development Zone, how many could have imagined the scene ten years later?

By then, Hefei had already made its name with its investment in BOE. During the 2008 global financial crisis, while most cities cut fiscal spending, Hefei devoted nearly one‑third of its annual fiscal revenue to bring in the BOE project, filling the gap in display panels. And memory chips were the key “chip” piece in Hefei’s “chip‑display‑vehicle‑integration” strategy.

DRAM R&D is a “money‑burning marathon.” The long‑term monopoly of the DRAM market deterred countless investors, but Hefei’s state capital chose to double down. Looking at ChangXin’s equity structure: the largest shareholder, Hefei Qinghui Jidian, holds 21.67%; the second, ChangXin Integration, holds 11.71%; and the fifth, Anhui Provincial Investment, holds 7.91%. Behind Qinghui Jidian are Hefei Economic Development Zone state‑owned assets and the Hefei Industry Investment system, while ChangXin Integration is 100% owned by the Hefei SASAC through Hefei Industry Investment.

Rough estimates show that Hefei state‑owned entities collectively hold more than 35% of the shares. That means that during the years when ChangXin Memory Technologies was losing money year after year and was least favoured, Hefei state capital always stood behind it. In 2023 and 2024, the company accumulated losses of RMB 23.48 billion – such sustained capital injection is not an easy decision for any local government.

According to Yuan Fei, then General Manager of Hefei Industry Investment, the first phase of ChangXin Memory Storage construction required RMB 18 billion, of which the Hefei municipal government contributed three‑quarters and GigaDevice contributed one‑quarter. During the industry downturn, the presence of Anhui provincial and Hefei municipal state capital not only appeared frequently in multiple rounds of investment in ChangXin Memory Technologies, but also attracted a number of social capital followers, helping the company survive the early “money‑burning” stage.

This is the most vivid illustration of “patient capital.” While the market voted with its feet and sentenced the company to death with short‑term losses, Hefei chose to believe in long‑term value. After a decade of accompaniment, the victory of patience is now paying off.

Based on the shareholder structure disclosed in the prospectus, ChangXin Memory Technologies’ post‑IPO valuation is expected to be in the range of RMB 2‑3 trillion. With Hefei state‑owned entities holding about 37%, for every RMB 1 trillion in market capitalisation, Hefei state capital would hold about RMB 370 billion in paper value – an amount close to the magnitude of Hefei’s 2025 GDP (RMB 1.42 trillion).

But Hefei is not the only winner. ChangXin’s capital landscape is far grander than anticipated.


IV

The Relay: A Nine‑Round

Capital Relay Race

From its angel round in 2018 to its final pre‑IPO round in June 2025, ChangXin Memory Technologies completed nine financing rounds, raising tens of billions of yuan in total. The leap in valuation numbers charts the full course of Chinese semiconductor capital from disbelief to feverish enthusiasm: Series C in 2021, valuation approaching RMB 40 billion; financing in 2022, valuation about RMB 107.8 billion; in March 2024, a strategic financing of RMB 10.8 billion at a pre‑money valuation of about RMB 140 billion; and by the most recent round in June 2025, post‑money valuation climbed to RMB 158.3 billion.

The jump from a valuation of nearly RMB 10 billion to the 100‑billion level took less than five years. By the date of the prospectus, the company had 60 institutional shareholders.

Behind this capital relay is a full‑chain funding map that includes national industrial capital, local state capital, bank AICs, broker direct investment, and insurance funds.

National Team: The second phase of the National Integrated Circuit Industry Investment Fund (Big Fund II) holds 8.73% as the third largest shareholder, the highest among external institutional shareholders. The entry of Big Fund II itself is a symbol of ChangXin’s industrial status – the core platform for promoting integrated circuit development at the national level placed its heaviest bet on this domestic DRAM leader.

Banking Team: The collective ambush of the five major banks’ AICs is one of the most striking capital stories. In June 2024, ChangXin completed its eighth capital increase, with the five AICs entering en masse: CCB Investment contributed RMB 1.3 billion, BOC Asset and BOCOM Investment each RMB 600 million, ICBC Financial Investment RMB 1 billion, and ABC Investment added more. In total, the five AICs invested RMB 3.9 billion, accounting for 36% of the round’s RMB 10.8 billion total. At that time, ChangXin’s post‑money valuation was about RMB 150 billion.

Two years later, this “patient capital” is reaping unprecedented returns. If ChangXin’s post‑IPO market capitalisation reaches RMB 2‑3 trillion, the market value of these bank AIC holdings will reach tens of billions or even hundreds of billions of yuan, potentially yielding more than 10 times paper profit relative to their entry costs. Against the backdrop of shrinking margins in traditional credit businesses, the AICs’ forward‑looking layout in hard tech has opened up an entirely new profit growth curve for the major state‑owned banks.

Also noteworthy in this round is a group of “special investors” – the core technical team of ChangXin, who collectively hold about 4.2% through employee stock ownership platforms. Valued at a RMB 2 trillion market cap, these engineers and management would see their stakes exceed RMB 80 billion. From low‑salary perseverance during the “money‑burning” years to today’s wealth realisation, their story is the most touching micro‑note in this capital relay.

Insurance Capital: Early positioning. Six insurance entities – Hexie Health, China Life Investment, PICC Capital, Sunshine Life, China Post Life, and PICC Technology Innovation – directly hold stakes in ChangXin. Based on a RMB 3 trillion valuation, Hexie Health’s stake would be worth RMB 40.5 billion, and China Life Investment and PICC Capital each over RMB 20 billion.

Brokerage: Multi‑layered hidden positions. China Merchants Securities, through direct and indirect holdings, holds about 505 million shares, the largest among publicly disclosed brokerages. China Securities, one of the sponsors, indirectly holds about 88.96 million shares through its subsidiary; CICC, Guotai Haitong, Founder Securities, Huaan Securities, and Guoyuan Securities also hold indirect stakes through various means. Adding underwriting and sponsorship fees estimated at RMB 380‑750 million, plus long‑term capital gains from sponsor follow‑on investments, the depth of brokerages’ positioning on ChangXin’s chessboard far exceeds market recognition.

This is a perfect resonance of national will, local strategy, and market‑oriented capital. From any angle, ChangXin Memory Technologies is the most eye‑catching IPO case in China’s capital market in recent years – bar none.

V

The Journey Ahead: The Real Battlefield

Beyond the Cycle

The successful review on May 27 pushed this comeback story to a new climax.

The RMB 29.5 billion fundraising scale is second only to SMIC’s RMB 53.2 billion, making it the second largest since the opening of the STAR Market. The funds will be directed to three main areas: the technology upgrade and transformation of memory wafer manufacturing production lines (RMB 7.5 billion), DRAM memory technology upgrades (RMB 13 billion), and forward‑looking R&D for DRAM (RMB 9 billion). Expanding production, catching up in technology, and aiming for the future – ChangXin’s debut on the capital market landed three heavy punches.

But looking down from the peak of the legend, the comeback story also raises three deeper questions that must be examined.

First, how long can the tailwind last?

The fundamental driver of this performance surge is the supply‑demand imbalance caused by the AI computing wave. ChangXin explicitly warns in the prospectus: if future AI‑driven DRAM demand falls short of expectations, and at the same time new global capacity is released intensively, the industry supply‑demand relationship may change again, affecting the company’s performance.

This is a challenge that all companies reliant on cycle‑driven profits must face head‑on. When Samsung, SK Hynix, and Micron complete their capacity reallocation, the HBM market saturates, and conventional DRAM supply returns to balance, can ChangXin maintain its nearly RMB 300 million daily profit? Only time will tell.

Second, is the technological moat a defence, or just the starting line for the pursuer?

ChangXin’s current market share is 7.67%, while the combined share of the three giants – Samsung, SK Hynix, and Micron – exceeds 90%. The gap remains huge. R&D investment, talent reserves, and process iteration speed are all long‑term challenges. From 4th‑generation process to more advanced nodes, the path of catch‑up is long.

At the same time, ChangXin’s breakthrough sends a highly valuable signal to China’s chip industry: at least in DRAM – once called the “jewel in the crown of semiconductors” – a Chinese company has broken into the global first tier. Once that door is opened, it can never be closed.

From a broader perspective, ChangXin’s rise is reshaping the profit distribution pattern of the global memory industry. In 2025, China’s imports of memory chips still amounted to about USD 120 billion, with a domestic substitution rate of less than 15%. For every one percentage point increase in global market share, ChangXin reduces import dependence by nearly USD 10 billion annually. This comeback is not just a commercial success at the company level, but a key move in the national supply chain security strategy.

ChangXin’s IPO is by no means merely a victory for the capital market. It is a critical milestone in a decade‑long industrial breakthrough, a landmark on China’s semiconductor journey from “bottleneck” to “domestic substitution.” Looking back to 2016 when the “506 project” was launched, the domestic DRAM field was almost a blank slate; today, three 12‑inch wafer fabs in Hefei and Beijing operate day and night, and domestically produced memory chips have entered the global mainstream supply chain. That itself is a comeback.


Conclusion

A decade of patient accompaniment, and the flowers bloom at last. When ChangXin Memory Technologies officially lists on the STAR Market, it will no longer be just “China’s hope for DRAM,” but a truly competitive player that has been validated by the global market.

The next chapter of the legend will be written by China’s own memory giant.

Before writing that chapter, let us look back at ChangXin’s growth curve: founded in 2016, first domestic DDR4 put into production in 2019, billions in losses in 2022, turned losses into profits in 2025, and net over RMB 24.7 billion in a single quarter in 2026. In ten years, it has sprung from the depths of loss to earn nearly RMB 300 million per day. The pace of the turnaround has stunned the entire Chinese capital market. When it stands on the STAR Market, it will be not only a coming‑of‑age ceremony for an enterprise, but also a historic footnote for a nation’s industrial breakthrough.



Related Articles
Copyright © 2026 Asia Pacific Pengsheng Certified Tax Agents Co., Ltd. All rights reserved.  粤ICP备19025653号