CXMT Deep Dive: DDR5 Ramp, RMB50bn First-Half Attributable Profit, and the Cycle Test for China’s DRAM Leader
目录
TL;DR
1. What CXMT Actually Sells: Memory Chips, Not a Concept Stock
2. The Real Focus of This Prospectus: The Income Statement Suddenly Looks Intimidating
III. Where Profit Comes From: Pricing, DDR5, and Scale Effects Working Together
IV. Revenue Mix Has Changed: From a Mobile Memory Base to Server DDR5 Optionality
V. Why 2023 Losses Were So Large: DRAM Fabs Carry Heavy Fixed Costs
VI. Customer Mix: High Concentration, But Not Yet Single-Customer Dependence
VII. RMB 34.5bn Fundraising Projects: Capital Mainly Goes to Upgrades, Not Storytelling
VIII. Offering Arrangement: Pricing Is Not Set Yet; For Now, Watch the Mechanism
IX. Three Scenarios: This Report Cannot Only Discuss the Upside
10. Compared With Global Peers, CXMT’s Strengths and Weaknesses Are Both Clear
11. The Most Easily Misread Point: First-Half Profit Is Very High, but the Full Year Cannot Be Linearly Extrapolated
12. Risk List: This Company Is Strong, but the Risks Are Not Soft
13. How to Define CXMT Before the Issue Price Is Announced
XIV. After the Offer Price Is Set, First Clarify the Earnings Basis
XV. ChangXin Post-Listing Tracking Checklist
XVI. Conclusion: The Key Question for ChangXin Technology Is How Much of the High Profit Can Be Retained
本内容基于公开资料和研报数据整理，不构成任何投资建议，不代表任何个人观点，仅供学习参考，请理性阅读
The core of CXMT’s prospectus is how a DRAM manufacturer long constrained by depreciation, R&D; spending, and a memory downcycle is releasing profit in 2026 through DDR5 volume ramp, price increases, and scale effects. The offer price has not been set, so valuation must wait for bookbuilding results, but the profit drivers, cycle risks, and use of IPO proceeds are already clear.
TL;DR
CXMT has entered a profit realization phase. Revenue reached RMB61.799bn in 2025, and 2026Q1 revenue already reached RMB50.8bn. The prospectus gives an even stronger estimate for 2026H1: revenue of RMB110-120bn, net profit attributable to shareholders of RMB50-57bn, and ex-one-off attributable net profit of RMB52-58bn. These figures show CXMT’s profit elasticity has now reached the scale of mainstream DRAM makers. But the company has not issued official full-year profit guidance, so investors should not simply double first-half profit to judge the full year.
The profit surge comes from three variables. DRAM price increases, DDR5 volume ramp, and capacity scale diluting fixed costs have together pushed up the income statement. In 2025, DDR-series revenue was already close to RMB20bn, while server application revenue also expanded meaningfully. This shows CXMT is shifting from a business led by mobile LPDDR toward higher-priced, higher-margin DDR5 and server memory.
Net profit needs to be unpacked by accounting line. CXMT’s 2025 net profit was RMB7.144bn, but net profit attributable to shareholders was only RMB1.875bn, while ex-one-off attributable net profit was RMB5.316bn. The reason is the complexity of structured entities, minority interests, fair-value changes, and other accounting items. To assess earning power, investors cannot focus on a single net profit number. They need to look at ex-one-off attributable net profit, gross margin, operating cash flow, and the pricing cycle together. In 2026Q1, ex-one-off attributable net profit reached RMB26.341bn, nearly five times the full-year 2025 level. This is the strongest financial signal in the prospectus.
CXMT’s risks are also clear: DRAM is a highly cyclical industry. Prices rise fast, and they can fall hard. The prospectus itself warns that the sharp growth in 2026H1 results may not be sustainable. In 2023, the company posted an attributable net loss of RMB16.34bn, mainly because industry downturn, price declines, depreciation, and inventory impairments hit at the same time. CXMT’s current profit should not be understood like that of a stable consumer electronics company. It is more like an asset-heavy cyclical manufacturer moving from capacity ramp into a pricing upcycle.
Valuation needs to wait for the offer price. Before the offer price is announced, the valuation question for CXMT comes down to three things: how high the offer price is versus annualized 2026H1 ex-one-off attributable net profit PE; whether the market is willing to give China’s DRAM leader a higher multiple on cyclical peak earnings; and whether the RMB29.5bn in IPO proceeds can turn technology platform and capacity upgrades into the next round of cost advantage. If priced off 2025 profit, it will look expensive. If priced off annualized 2026H1 profit, it will look cheap. The real question is how much of this profit cycle can be retained.
1. What CXMT Actually Sells: Memory Chips, Not a Concept Stock
CXMT makes DRAM. In plain terms, DRAM is the memory used when phones, computers, and servers are running. No matter how fast a CPU, GPU, or mobile processor calculates, it still needs DRAM to feed in the data being processed. If memory is insufficient, the system lags. If memory bandwidth is insufficient, the performance of AI servers and high-end terminals is constrained.
CXMT’s products are mainly divided into two lines. The first is DDR, mainly used in servers, desktops, laptops, and workstations. The second is LPDDR, mainly used in smartphones, tablets, thin-and-light laptops, and wearables. DDR is more focused on performance and capacity, while LPDDR is more focused on low power consumption and mobile devices. CXMT currently covers DDR4, DDR5, LPDDR4X, and LPDDR5/5X, and can also produce DRAM wafers, chips, and modules.
This matters for investment analysis. CXMT’s revenue does not come from a single product, nor is it only a bet on mobile memory. It already serves applications including mobile devices, servers, personal computers, and smart vehicles. Before 2025, LPDDR was the main revenue contributor because mobile customers were adopted earlier. After 2025, DDR5 and server revenue ramped quickly, and the company’s profit elasticity truly opened up.
CXMT is no longer at the stage of “whether China can make domestic DRAM.” The prospectus says the company has completed mass production across first- to fourth-generation process technology platforms, and has iterated from DDR4 and LPDDR4X to DDR5 and LPDDR5/5X. According to Omdia data, by DRAM sales in 2025Q4, CXMT’s global market share reached 7.67%. That is not global top three, but it is already enough to place the company inside the global DRAM supply structure.
China’s No. 1 and the world’s No. 4 DRAM manufacturer.
The investment implication is straightforward: CXMT is not a niche chip company. It has entered one of the most concentrated, capital-intensive, and price-cyclical semiconductor categories globally. Samsung Electronics, SK hynix, and Micron Technology have long held more than 90% of global market share. CXMT’s opportunity comes from the China market, domestic supply security, DDR5/LPDDR5 product upgrades, and tight global supply-demand. Its pressure comes from the same place: in this industry, scale, technology, funding, and customer qualification cannot fall behind.
AI Drives a Sector-Wide Revaluation of Memory: Who Has the Most Pricing Power Across DRAM, NAND, SSD, and HDD, with Samsung, SK hynix, SanDisk, Western Digital, and Seagate Results Cross-Checking One Another
2. The Real Focus of This Prospectus: The Income Statement Suddenly Looks Intimidating
CXMT was still posting large losses in 2023, narrowed losses in 2024, turned profitable in 2025, and saw profit suddenly expand in 2026Q1. This is not financial magic. The core reason is that the DRAM industry shifted from a downcycle to an upcycle, while CXMT’s own capacity, yield, product mix, and customer adoption moved into a better position.
Start with the most important table.
This table can be broken down into four points.
First, CXMT’s revenue rose from RMB9.087bn in 2023 to RMB61.799bn in 2025, nearly 5.8x growth in two years. 2026Q1 revenue was RMB50.8bn, already close to 82% of full-year 2025 revenue. This shows the company is in a phase where both pricing and volume are moving upward.
Second, 2026Q1 net profit attributable to shareholders was RMB24.762bn, while ex-one-off attributable net profit was RMB26.341bn. Ex-one-off attributable net profit being higher than attributable net profit means non-recurring gains and losses dragged the attributable profit line in Q1. For investors, ex-one-off attributable net profit better reflects the earning power of the core business.
Third, 2026H1 net profit attributable to shareholders is expected at RMB50-57bn, with ex-one-off attributable net profit at RMB52-58bn. This is the most important profit anchor in the prospectus. It proves that under the current DRAM pricing environment, CXMT already has extremely strong profit elasticity.
Fourth, the prospectus does not provide a full-year profit forecast. It explicitly states the following:
The company has not prepared a profit forecast report.
So this deep dive cannot say “CXMT will definitely earn a certain amount in 2026.” A more reasonable formulation is: first-half profit has been given as a range, while the full year depends on DRAM prices in Q3 and Q4, customer pull-in, capacity ramp, and product mix.
III. Where Profit Comes From: Pricing, DDR5, and Scale Effects Working Together
CXMT’s profit did not appear out of nowhere. In 2023, it was constrained by three factors: low DRAM prices, capacity still ramping, and heavy fixed-asset depreciation. Starting in 2H25, all three reversed at the same time: prices rose, capacity utilization increased, and DDR5 and LPDDR5/5X scaled up.
DRAM is a highly standardized product. When prices rise, profit is amplified quickly because fixed costs such as production lines, fabs, and equipment depreciation do not rise in proportion to revenue. When prices fall, profit can also be wiped out quickly because those fixed costs remain.
CXMT’s gross margin changes in 2025 already show this logic.
The key point in this table is the direction of travel.
The DDR series saw a classic “volume and price both up” move in 2025. Unit price rose 61.00%, sales volume rose 282.22% by capacity, and gross margin recovered from -26.87% in 2024 to 41.89%. This means the DDR5 ramp did not only bring revenue; it also improved gross-profit quality.
LPDDR also recovered meaningfully. In 2025, LPDDR unit price rose 24.46%, sales volume rose 65.18%, and gross margin returned to 37.25%. Mobile-device demand remains CXMT’s revenue base, but servers and DDR5 are the areas where the profit slope changed more significantly.
CXMT’s own explanation is straightforward: in 2025, high-priced, high-margin DDR5 products ramped quickly, lifting the DDR series’ revenue contribution. Product-mix change is often more important than total revenue growth. Selling RMB10bn of low-margin mature products has a completely different income-statement impact from selling RMB10bn of high-margin DDR5.
IV. Revenue Mix Has Changed: From a Mobile Memory Base to Server DDR5 Optionality
CXMT used to look more like a mobile DRAM manufacturer, with LPDDR making up the bulk of revenue. In 2025, the DDR series began to rise materially.
In 2025, DDR series revenue reached RMB19.531bn, up more than fivefold year on year. This change is critical. After DDR5 entered servers and PCs, CXMT was no longer only exposed to mobile memory demand. Server DRAM is more sensitive to AI infrastructure, cloud capex, and data-center expansion; once prices rise, its profit elasticity is stronger than mobile LPDDR.
The application mix also shows this change.
Mobile devices remain the largest revenue source, accounting for 60.40%. This provides CXMT with a stable shipment base. Server revenue has already become the second-largest application source, accounting for 26.51%, and its growth rate is far higher than that of mobile devices. The prospectus gives a CAGR of 536.24% for server-application revenue.
These numbers show that CXMT’s second growth line has already emerged. It does not need to immediately own the global HBM leader narrative to enter the ordinary memory layer of AI infrastructure through server DDR5, RDIMM, MRDIMM, and higher-capacity modules. Commodity DRAM does not carry the same technology premium as HBM, but it is essential for server operations, and when prices rise it can still push up profit.
The Memory Tax Is Here: How AI Is Turning HBM, DRAM, and NAND into a Global Macro Bottleneck
V. Why 2023 Losses Were So Large: DRAM Fabs Carry Heavy Fixed Costs
CXMT posted a net loss attributable to shareholders of RMB16.340bn in 2023, and an adjusted net loss attributable to shareholders of RMB16.752bn. This loss should not simply be interpreted as evidence that the company’s products had no value. During the capacity ramp, DRAM fabs bear very heavy depreciation, R&D;, and inventory write-down pressure.
DRAM production lines are expensive. Fabs, equipment, cleanrooms, and process platforms all need to be funded upfront. Before capacity and yield are fully ramped, depreciation has already begun to enter costs. CXMT’s fixed-asset depreciation was RMB10.555bn in 2023, RMB14.875bn in 2024, and RMB24.680bn in 2025. Depreciation depresses profit, but it also shows that the company has already invested in heavy-asset production lines.
R&D; is also heavy. From 2023 to 2025, CXMT’s R&D; spending was RMB4.670bn, RMB6.341bn, and RMB9.593bn, respectively, totaling RMB20.605bn over three years. In 2025, it had 6,259 R&D; employees, accounting for 32.43% of total staff. This is not an asset-light design company, nor a chip brand relying only on outsourced manufacturing. It follows the DRAM IDM model.
One point needs to be made clear: accumulated losses do not mean the company has poor cash flow. In 2025, CXMT’s net operating cash flow was RMB36.520bn; in 1Q26, net operating cash flow was RMB42.566bn. The historical losses on the income statement mainly came from early heavy-asset investment, depreciation, R&D;, and the cyclical downturn. Now that prices have moved up, cash flow is already very strong.
The real issue to track is whether depreciation and R&D; can be diluted by a larger revenue base. In 2025, this had already happened: gross margin rose from 5.58% in 2024 to 40.99%. The 1Q26 and 1H26 estimates show that scale effects are continuing to expand.
VI. Customer Mix: High Concentration, But Not Yet Single-Customer Dependence
ChangXin’s customer concentration is not low. In 2025, the top five customers accounted for 68.08% of principal operating revenue, versus 67.30% in 2024 and 74.12% in 2023. For a DRAM vendor, this is not surprising. Memory customers are mainly large accounts across smartphones, servers, PCs, modules, and distribution channels, and the industry is inherently concentrated.
The prospectus discloses that the company has entered the customer systems of Alibaba Cloud, ByteDance, Tencent, Lenovo, Xiaomi, Transsion, Honor, OPPO, vivo, and others. These represent two types of demand: mobile devices, which require power efficiency, stability, and supply assurance; and servers and cloud providers, which require capacity, bandwidth, reliability, and long-term supply.
The high distribution share is one point in this prospectus that deserves attention. In 2025, distribution revenue accounted for 85.38% of principal operating revenue. Selling semiconductor products through distribution channels is common, especially for standardized products such as memory; distributors help the company reach downstream customers quickly and accelerate turnover. But a high distribution share also means end-demand, inventory, and the real pace of customer pull-ins require ongoing cross-checking.
ChangXin has not disclosed any single customer contributing more than 50% of revenue, and says it does not have severe dependence on a small number of customers. Investors should next watch whether the direct-sales customer share can gradually rise, whether server customers and leading cloud providers can continue to be penetrated, and whether AI server-related revenue can move from “still relatively low” to a genuinely sizable source of profit.
VII. RMB 34.5bn Fundraising Projects: Capital Mainly Goes to Upgrades, Not Storytelling
This IPO involves total project investment of RMB 34.5bn, with RMB 29.5bn of proceeds planned for use. The project direction is highly focused, all centered on DRAM technology upgrades, production-line transformation, and forward-looking R&D.;
These three fundraising projects correspond to ChangXin’s next three core questions.
First, whether unit costs can continue to decline. DRAM is a scale- and process-driven industry. If the same wafer can produce more qualified chips, costs fall. If ChangXin can improve wafer output, yield, and equipment efficiency through process upgrades, gross margin will not rely only on price increases.
Second, whether high-end products can continue to scale. DDR5 and LPDDR5/5X have already shown that product-mix upgrades can lift gross margin. The next step is whether server DDR5, high-capacity modules, low-power high-end mobile memory, and future products can continue to be introduced into customer platforms.
Third, whether supply-chain security can become stronger. The prospectus mentions the introduction of domestic and new types of equipment, materials, and components. For domestic DRAM, the equipment, materials, EDA, components, and packaging-and-testing ecosystems directly determine capacity expansion and long-term stable supply.
This capital will not immediately turn into profit. The fundraising projects will bring more depreciation, R&D;, and upfront expenses. The real investment judgment is whether they can deliver higher output, lower costs, a higher share of newer-generation products, and stronger customer qualification.
VIII. Offering Arrangement: Pricing Is Not Set Yet; For Now, Watch the Mechanism
ChangXin’s offering combines strategic placement, offline bookbuilding, and online subscription. The issuance ratio is 10.00%, with an over-allotment option. The specific share capital, greenshoe, and timetable are set out in the table below to avoid overloading the text with numbers.
The most important variable has not yet been disclosed: the offer price. Without the offer price, there is no definitive market capitalization, issuance P/E, or issuance P/B. EPS in the prospectus is also temporarily calculated using the lower of 2025 net profit attributable to shareholders before and after non-recurring items divided by pre-issuance total share capital. Post-issuance valuation must wait until the price is determined.
Therefore, it is not yet possible to say “ChangXin is expensive” or “ChangXin is cheap.” For now, the framework can be laid out clearly.
IX. Three Scenarios: This Report Cannot Only Discuss the Upside
The judgment on ChangXin Technology must be placed within three scenarios. Because the DRAM industry is cyclical, valuation based on a single earnings point can easily mislead.
Among these three scenarios, the most likely outcome is neither “permanent prosperity” nor “an immediate return to 2023.” A more reasonable judgment is that the first half of 2026 has already been a strong-cycle window, and the market will treat part of the profit as cyclical peak earnings. ChangXin needs to prove that profit comes not only from pricing, but also from product-generation upgrades, server customers, capacity utilization, and cost reduction.
10. Compared With Global Peers, CXMT’s Strengths and Weaknesses Are Both Clear
The global DRAM industry has long been dominated by Samsung Electronics, SK hynix, and Micron Technology. CXMT’s share has reached 7.67%, putting it on the radar of major players, but it remains in catch-up mode versus the top three.
CXMT has three advantages.
First, China’s market demand is enormous. China is one of the world’s major DRAM consumption markets. Smartphones, PCs, servers, cloud computing, and smart vehicles all require memory. As the leading domestic DRAM company, CXMT has natural advantages in supply security and localized customer coverage.
Second, its product generation has reached the mainstream. After DDR5 and LPDDR5/5X entered mass production, CXMT is no longer relying only on older-generation products for shipments. The DDR series’ 2025 gross margin of 41.89% already shows that new products can support better earnings quality.
Third, it is highly scarce in the A-share market. A-shares have semiconductor equipment, materials, packaging and testing, and design companies, but very few true scale DRAM IDM targets. After listing, CXMT will likely become the core A-share proxy for the memory cycle, domestic DRAM, and AI server memory.
The weaknesses are equally clear.
First, compared with Samsung, SK hynix, and Micron, CXMT still has gaps in global customers, capacity scale, advanced generations, HBM, and ecosystem coordination. Its current strength is more in commodity DRAM and mainstream DDR/LPDDR products. It should not be directly described as an HBM leader.
Second, its earnings remain highly exposed to DRAM prices. The sharp profit improvement in 2025 and 1Q26 benefited from global undersupply and price increases. If prices fall, gross margin will come under pressure.
Third, it still has accumulated uncovered losses. At the end of 2025, undistributed profit was RMB -36.65 billion. This will not prevent the company from operating, but it will affect dividends and the market’s assessment of earnings stability.
11. The Most Easily Misread Point: First-Half Profit Is Very High, but the Full Year Cannot Be Linearly Extrapolated
CXMT expects 1H26 net profit attributable to shareholders of RMB 50-57 billion, an eye-catching number. Many people will naturally multiply it by two and imagine full-year attributable profit of more than RMB 100 billion. But the prospectus has already stated that the first-half performance estimate does not constitute an earnings forecast or performance commitment.
It does not constitute an earnings forecast or performance commitment.
This sentence must be taken seriously.
DRAM prices do not move in a linear trend. Customer restocking, AI server pull-in, capacity allocation by global leaders, smartphone and PC demand, and inventory cycles all affect pricing. CXMT’s very high first-half profit shows that current pricing and product mix are very favorable; it does not automatically prove that the second half will be equally strong.
The right tracking method is to watch four numbers.
One post-issuance metric should also be added: market capitalization and PE implied by the issue price. If the issue price is low, the market may treat CXMT as a high-beta domestic DRAM leader. If the issue price is high, the market will ask more stringent questions about earnings sustainability.
12. Risk List: This Company Is Strong, but the Risks Are Not Soft
There is no need to talk around CXMT’s risks. The prospectus has already laid them out clearly. What investors really need to do is translate the risks into trackable numbers.
First, the DRAM price cycle. The root cause of CXMT’s 2023 loss was the DRAM industry downturn, a sharp price decline, inventory write-downs, and fixed-cost pressure. One of the root causes of the current profit surge is also price increases. The benefit and the risk in this industry come from the same variable.
Second, uncertainty in AI demand. AI server and cloud capex have lifted DRAM demand, but AI application deployment, improvements in model efficiency, and cloud spending cadence can all change. If AI capex slows, DRAM prices will be affected.
Third, capacity expansion by global leaders. When Samsung Electronics, SK hynix, and Micron see profit improvement, they will increase capital expenditure. New capacity will not be released immediately, but if future demand fails to keep up with supply, the industry will return to oversupply.
Fourth, depreciation and expenses from the fundraising projects. The fundraising projects can improve long-term competitiveness, but they will also increase depreciation, amortization, and expenses in the short and medium term. Only if capacity ramp-up, yield improvement, and product upgrades are fast enough will the projects translate into profit rather than expense pressure.
Fifth, the distributor share and customer concentration. A high distributor-model share can accelerate sales and turnover, but end-market real demand and inventory must be continuously verified. The top five customers account for close to 70%, making customer structure stability important.
Sixth, pressure to catch up technologically. CXMT has already achieved DDR5 and LPDDR5/5X, but global DRAM technology continues to evolve. Advanced processes, capacity, speed, power consumption, yield, module form factors, and future forward-looking DRAM technologies all require continued catch-up.
13. How to Define CXMT Before the Issue Price Is Announced
The most appropriate characterization of CXMT is: “China’s DRAM leader entering a profit realization window.” It is neither a pure concept stock nor a stable dividend asset. It is a capital-intensive, highly cyclical, technology catch-up semiconductor manufacturer with strong domestic scarcity value.
Looking only at 2025, CXMT had just emerged from losses. Revenue was RMB 61.799 billion, net profit attributable to shareholders was RMB 1.875 billion, and recurring net profit attributable to shareholders was RMB 5.316 billion. Historical accumulated losses had not yet been filled.
Looking at 1Q26 and the first-half estimate, CXMT resembles a cyclical leader with extremely strong earnings leverage. First-half net profit attributable to shareholders is expected to be RMB 50-57 billion, and recurring attributable net profit is expected to be RMB 52-58 billion, putting it in an extremely high profit range.
Looking longer term, CXMT also resembles a capital-intensive technology company still catching up with the global top three. The fundraising projects total RMB 34.5 billion, R&D; investment over three years was RMB 20.605 billion, and fixed-asset depreciation continues to increase. It needs continuous investment to maintain technology generations, capacity scale, and customer qualifications.
None of these three perspectives can be dropped. Looking only at domestic scarcity underestimates cycle risk; looking only at the 2023 loss misses the 2026 earnings shift; annualizing only first-half profit overestimates the stability of peak-cycle earnings.
XIV. After the Offer Price Is Set, First Clarify the Earnings Basis
ChangXin’s IPO is likely to see the widest valuation disagreement because different earnings bases produce completely different P/E ratios. The company was already profitable in 2025, but net profit attributable to shareholders was only RMB1.875 billion, while recurring net profit attributable to shareholders was RMB5.316 billion. In 1H26, expected profit suddenly jumps to more than RMB50 billion. Using 2025, the profit base still looks low; using annualized 1H26, earnings look very substantial. Both bases need to be considered. Looking at either one in isolation can easily be misleading.
Pre-IPO total share capital is 60.193 billion shares. Post-IPO total share capital is 66.881 billion shares, and 67.884 billion shares after full exercise of the greenshoe. On a rough post-IPO share-count basis, 1H26 net profit attributable to shareholders of RMB50-57 billion implies 1H EPS of about RMB0.75-0.85; simply annualized, that implies RMB1.49-1.70. After full exercise of the greenshoe, 1H EPS is about RMB0.74-0.84, or roughly RMB1.47-1.68 annualized.
This calculation should not be treated as a full-year forecast, but it helps frame market pricing once the offer price is released.
Once the offer price is released, it can be read this way:
First, if the offer price implies a very high P/E on 2025 earnings, that does not necessarily mean the pricing is unreasonable, because 2025 does not yet fully reflect 2026 pricing and DDR5 volume ramp.
Second, if the offer price implies a very low P/E on annualized 1H26 earnings, that does not necessarily mean it is cheap, because 1H earnings themselves sit within a strong upcycle window.
Third, mid-cycle earnings are the most useful reference point. The prospectus does not provide this figure directly. It needs to be calibrated later using 3Q26 and 4Q26 gross margin, pricing, and inventory impairment. ChangXin’s first post-listing re-rating will very likely revolve around whether 2026 earnings can be partially capitalized.
Fourth, A-share scarcity will increase market attention. ChangXin is not an ordinary chip design stock, nor a traditional consumer electronics component company. It is a rare DRAM IDM leader in the A-share market. Scarcity can support a valuation premium, but if the offer price has already fully priced in the strong 1H profit, subsequent earnings validation will become more demanding.
XV. ChangXin Post-Listing Tracking Checklist
After ChangXin lists, investors should not just watch the stock price every day. More importantly, they need to track whether fundamentals are continuing along the strong-cycle path outlined in the prospectus.
First, watch the offer price and post-IPO market cap. The offer price determines which earnings basis the market is using to value ChangXin. If the market cap already fully reflects annualized 1H26 earnings, stronger 3Q and 4Q data will be needed to support it.
Second, watch 3Q revenue and gross margin. The 1H26 profit range has already been provided, and 3Q will test whether elevated DRAM pricing is continuing.
Third, watch the revenue mix from DDR5 and servers. The real improvement in ChangXin’s earnings quality comes from volume growth in DDR5, servers, and higher-generation products. Mobile LPDDR is the base; server DDR5 is the upside.
Fourth, watch inventory impairment losses. If DRAM prices fall, inventory write-downs may re-enter the income statement. The lesson from 2023 should not be forgotten.
Fifth, watch the progress of fundraising investment projects. If technology upgrades and forward-looking R&D; proceed smoothly, they will support the next round of cost reduction and product upgrades. If progress is slow, depreciation and expenses will hit the financial statements first.
Sixth, watch global peer capex. The expansion pace of Samsung, SK hynix, and Micron will determine industry supply-demand after 2027. If global supply again overwhelms demand, ChangXin will also find it difficult to stay immune.
SK hynix Deep-Dive Update: Is It Still Expensive? DDR5 Takes Over Price Increases, Korean Exports Surge, and Samsung Catch-Up Risk
Samsung Electronics Deep-Dive Update: 2Q26 Core Profit Beats Expectations, HBM4 Exceeds USD1 Billion, and Valuation Reset After the First Mobile Loss
Micron Deep Dive: Can Peak Earnings Be Capitalized, and How AI Memory LTAs Are Rewriting the Cycle Discount
XVI. Conclusion: The Key Question for ChangXin Technology Is How Much of the High Profit Can Be Retained
ChangXin Technology’s prospectus already provides a clear picture of the company. It is the DRAM leader in mainland China, with three 12-inch DRAM wafer fabs and products covering DDR5 and LPDDR5/5X. Its global share had already reached 7.67% in 2025. The story is no longer simply “can China make domestic DRAM,” but rather “can China’s domestic DRAM leader generate stable profits in a global memory upcycle.”
Expected 1H26 net profit attributable to shareholders of RMB50-57 billion is the strongest financial anchor for this company during the IPO stage. This figure will attract the market, and it will also create debate. Optimists will focus on domestic scarcity, DDR5 volume ramp, AI servers, and tight supply; more cautious investors will focus on the DRAM cycle, elevated pricing, depreciation pressure, and historical losses.
The most prudent judgment is this: ChangXin has already proved that it can earn substantial profits during a strong upcycle. The next step is to prove that it can retain sufficient profit when the cycle rolls over. Before the offer price is released, there is no need to rush to conclude whether it is cheap or expensive. Only when the offer price, listed market cap, 3Q pricing, and server revenue mix are all available will ChangXin’s first round of market pricing have a real anchor.
This company is worth writing about and worth tracking. It is not a target that can be explained clearly in a generic domestic substitution article. After ChangXin Technology lists, the A-share market will, for the first time, have a DRAM leader of real scale as a sample. Every subsequent quarter will become a key window for observing whether China’s memory industry can move from capacity catch-up to profit realization.CXMT Deep Dive: DDR5 Ramp, RMB50bn First-Half Attributable Profit, and the Cycle Test for China’s DRAM Leader
目录
TL;DR
1. What CXMT Actually Sells: Memory Chips, Not a Concept Stock
2. The Real Focus of This Prospectus: The Income Statement Suddenly Looks Intimidating
III. Where Profit Comes From: Pricing, DDR5, and Scale Effects Working Together
IV. Revenue Mix Has Changed: From a Mobile Memory Base to Server DDR5 Optionality
V. Why 2023 Losses Were So Large: DRAM Fabs Carry Heavy Fixed Costs
VI. Customer Mix: High Concentration, But Not Yet Single-Customer Dependence
VII. RMB 34.5bn Fundraising Projects: Capital Mainly Goes to Upgrades, Not Storytelling
VIII. Offering Arrangement: Pricing Is Not Set Yet; For Now, Watch the Mechanism
IX. Three Scenarios: This Report Cannot Only Discuss the Upside
10. Compared With Global Peers, CXMT’s Strengths and Weaknesses Are Both Clear
11. The Most Easily Misread Point: First-Half Profit Is Very High, but the Full Year Cannot Be Linearly Extrapolated
12. Risk List: This Company Is Strong, but the Risks Are Not Soft
13. How to Define CXMT Before the Issue Price Is Announced
XIV. After the Offer Price Is Set, First Clarify the Earnings Basis
XV. ChangXin Post-Listing Tracking Checklist
XVI. Conclusion: The Key Question for ChangXin Technology Is How Much of the High Profit Can Be Retained
本内容基于公开资料和研报数据整理，不构成任何投资建议，不代表任何个人观点，仅供学习参考，请理性阅读
The core of CXMT’s prospectus is how a DRAM manufacturer long constrained by depreciation, R&D; spending, and a memory downcycle is releasing profit in 2026 through DDR5 volume ramp, price increases, and scale effects. The offer price has not been set, so valuation must wait for bookbuilding results, but the profit drivers, cycle risks, and use of IPO proceeds are already clear.
TL;DR
CXMT has entered a profit realization phase. Revenue reached RMB61.799bn in 2025, and 2026Q1 revenue already reached RMB50.8bn. The prospectus gives an even stronger estimate for 2026H1: revenue of RMB110-120bn, net profit attributable to shareholders of RMB50-57bn, and ex-one-off attributable net profit of RMB52-58bn. These figures show CXMT’s profit elasticity has now reached the scale of mainstream DRAM makers. But the company has not issued official full-year profit guidance, so investors should not simply double first-half profit to judge the full year.
The profit surge comes from three variables. DRAM price increases, DDR5 volume ramp, and capacity scale diluting fixed costs have together pushed up the income statement. In 2025, DDR-series revenue was already close to RMB20bn, while server application revenue also expanded meaningfully. This shows CXMT is shifting from a business led by mobile LPDDR toward higher-priced, higher-margin DDR5 and server memory.
Net profit needs to be unpacked by accounting line. CXMT’s 2025 net profit was RMB7.144bn, but net profit attributable to shareholders was only RMB1.875bn, while ex-one-off attributable net profit was RMB5.316bn. The reason is the complexity of structured entities, minority interests, fair-value changes, and other accounting items. To assess earning power, investors cannot focus on a single net profit number. They need to look at ex-one-off attributable net profit, gross margin, operating cash flow, and the pricing cycle together. In 2026Q1, ex-one-off attributable net profit reached RMB26.341bn, nearly five times the full-year 2025 level. This is the strongest financial signal in the prospectus.
CXMT’s risks are also clear: DRAM is a highly cyclical industry. Prices rise fast, and they can fall hard. The prospectus itself warns that the sharp growth in 2026H1 results may not be sustainable. In 2023, the company posted an attributable net loss of RMB16.34bn, mainly because industry downturn, price declines, depreciation, and inventory impairments hit at the same time. CXMT’s current profit should not be understood like that of a stable consumer electronics company. It is more like an asset-heavy cyclical manufacturer moving from capacity ramp into a pricing upcycle.
Valuation needs to wait for the offer price. Before the offer price is announced, the valuation question for CXMT comes down to three things: how high the offer price is versus annualized 2026H1 ex-one-off attributable net profit PE; whether the market is willing to give China’s DRAM leader a higher multiple on cyclical peak earnings; and whether the RMB29.5bn in IPO proceeds can turn technology platform and capacity upgrades into the next round of cost advantage. If priced off 2025 profit, it will look expensive. If priced off annualized 2026H1 profit, it will look cheap. The real question is how much of this profit cycle can be retained.
1. What CXMT Actually Sells: Memory Chips, Not a Concept Stock
CXMT makes DRAM. In plain terms, DRAM is the memory used when phones, computers, and servers are running. No matter how fast a CPU, GPU, or mobile processor calculates, it still needs DRAM to feed in the data being processed. If memory is insufficient, the system lags. If memory bandwidth is insufficient, the performance of AI servers and high-end terminals is constrained.
CXMT’s products are mainly divided into two lines. The first is DDR, mainly used in servers, desktops, laptops, and workstations. The second is LPDDR, mainly used in smartphones, tablets, thin-and-light laptops, and wearables. DDR is more focused on performance and capacity, while LPDDR is more focused on low power consumption and mobile devices. CXMT currently covers DDR4, DDR5, LPDDR4X, and LPDDR5/5X, and can also produce DRAM wafers, chips, and modules.
This matters for investment analysis. CXMT’s revenue does not come from a single product, nor is it only a bet on mobile memory. It already serves applications including mobile devices, servers, personal computers, and smart vehicles. Before 2025, LPDDR was the main revenue contributor because mobile customers were adopted earlier. After 2025, DDR5 and server revenue ramped quickly, and the company’s profit elasticity truly opened up.
CXMT is no longer at the stage of “whether China can make domestic DRAM.” The prospectus says the company has completed mass production across first- to fourth-generation process technology platforms, and has iterated from DDR4 and LPDDR4X to DDR5 and LPDDR5/5X. According to Omdia data, by DRAM sales in 2025Q4, CXMT’s global market share reached 7.67%. That is not global top three, but it is already enough to place the company inside the global DRAM supply structure.
China’s No. 1 and the world’s No. 4 DRAM manufacturer.
The investment implication is straightforward: CXMT is not a niche chip company. It has entered one of the most concentrated, capital-intensive, and price-cyclical semiconductor categories globally. Samsung Electronics, SK hynix, and Micron Technology have long held more than 90% of global market share. CXMT’s opportunity comes from the China market, domestic supply security, DDR5/LPDDR5 product upgrades, and tight global supply-demand. Its pressure comes from the same place: in this industry, scale, technology, funding, and customer qualification cannot fall behind.
AI Drives a Sector-Wide Revaluation of Memory: Who Has the Most Pricing Power Across DRAM, NAND, SSD, and HDD, with Samsung, SK hynix, SanDisk, Western Digital, and Seagate Results Cross-Checking One Another
2. The Real Focus of This Prospectus: The Income Statement Suddenly Looks Intimidating
CXMT was still posting large losses in 2023, narrowed losses in 2024, turned profitable in 2025, and saw profit suddenly expand in 2026Q1. This is not financial magic. The core reason is that the DRAM industry shifted from a downcycle to an upcycle, while CXMT’s own capacity, yield, product mix, and customer adoption moved into a better position.
Start with the most important table.
This table can be broken down into four points.
First, CXMT’s revenue rose from RMB9.087bn in 2023 to RMB61.799bn in 2025, nearly 5.8x growth in two years. 2026Q1 revenue was RMB50.8bn, already close to 82% of full-year 2025 revenue. This shows the company is in a phase where both pricing and volume are moving upward.
Second, 2026Q1 net profit attributable to shareholders was RMB24.762bn, while ex-one-off attributable net profit was RMB26.341bn. Ex-one-off attributable net profit being higher than attributable net profit means non-recurring gains and losses dragged the attributable profit line in Q1. For investors, ex-one-off attributable net profit better reflects the earning power of the core business.
Third, 2026H1 net profit attributable to shareholders is expected at RMB50-57bn, with ex-one-off attributable net profit at RMB52-58bn. This is the most important profit anchor in the prospectus. It proves that under the current DRAM pricing environment, CXMT already has extremely strong profit elasticity.
Fourth, the prospectus does not provide a full-year profit forecast. It explicitly states the following:
The company has not prepared a profit forecast report.
So this deep dive cannot say “CXMT will definitely earn a certain amount in 2026.” A more reasonable formulation is: first-half profit has been given as a range, while the full year depends on DRAM prices in Q3 and Q4, customer pull-in, capacity ramp, and product mix.
III. Where Profit Comes From: Pricing, DDR5, and Scale Effects Working Together
CXMT’s profit did not appear out of nowhere. In 2023, it was constrained by three factors: low DRAM prices, capacity still ramping, and heavy fixed-asset depreciation. Starting in 2H25, all three reversed at the same time: prices rose, capacity utilization increased, and DDR5 and LPDDR5/5X scaled up.
DRAM is a highly standardized product. When prices rise, profit is amplified quickly because fixed costs such as production lines, fabs, and equipment depreciation do not rise in proportion to revenue. When prices fall, profit can also be wiped out quickly because those fixed costs remain.
CXMT’s gross margin changes in 2025 already show this logic.
The key point in this table is the direction of travel.
The DDR series saw a classic “volume and price both up” move in 2025. Unit price rose 61.00%, sales volume rose 282.22% by capacity, and gross margin recovered from -26.87% in 2024 to 41.89%. This means the DDR5 ramp did not only bring revenue; it also improved gross-profit quality.
LPDDR also recovered meaningfully. In 2025, LPDDR unit price rose 24.46%, sales volume rose 65.18%, and gross margin returned to 37.25%. Mobile-device demand remains CXMT’s revenue base, but servers and DDR5 are the areas where the profit slope changed more significantly.
CXMT’s own explanation is straightforward: in 2025, high-priced, high-margin DDR5 products ramped quickly, lifting the DDR series’ revenue contribution. Product-mix change is often more important than total revenue growth. Selling RMB10bn of low-margin mature products has a completely different income-statement impact from selling RMB10bn of high-margin DDR5.
IV. Revenue Mix Has Changed: From a Mobile Memory Base to Server DDR5 Optionality
CXMT used to look more like a mobile DRAM manufacturer, with LPDDR making up the bulk of revenue. In 2025, the DDR series began to rise materially.
In 2025, DDR series revenue reached RMB19.531bn, up more than fivefold year on year. This change is critical. After DDR5 entered servers and PCs, CXMT was no longer only exposed to mobile memory demand. Server DRAM is more sensitive to AI infrastructure, cloud capex, and data-center expansion; once prices rise, its profit elasticity is stronger than mobile LPDDR.
The application mix also shows this change.
Mobile devices remain the largest revenue source, accounting for 60.40%. This provides CXMT with a stable shipment base. Server revenue has already become the second-largest application source, accounting for 26.51%, and its growth rate is far higher than that of mobile devices. The prospectus gives a CAGR of 536.24% for server-application revenue.
These numbers show that CXMT’s second growth line has already emerged. It does not need to immediately own the global HBM leader narrative to enter the ordinary memory layer of AI infrastructure through server DDR5, RDIMM, MRDIMM, and higher-capacity modules. Commodity DRAM does not carry the same technology premium as HBM, but it is essential for server operations, and when prices rise it can still push up profit.
The Memory Tax Is Here: How AI Is Turning HBM, DRAM, and NAND into a Global Macro Bottleneck
V. Why 2023 Losses Were So Large: DRAM Fabs Carry Heavy Fixed Costs
CXMT posted a net loss attributable to shareholders of RMB16.340bn in 2023, and an adjusted net loss attributable to shareholders of RMB16.752bn. This loss should not simply be interpreted as evidence that the company’s products had no value. During the capacity ramp, DRAM fabs bear very heavy depreciation, R&D;, and inventory write-down pressure.
DRAM production lines are expensive. Fabs, equipment, cleanrooms, and process platforms all need to be funded upfront. Before capacity and yield are fully ramped, depreciation has already begun to enter costs. CXMT’s fixed-asset depreciation was RMB10.555bn in 2023, RMB14.875bn in 2024, and RMB24.680bn in 2025. Depreciation depresses profit, but it also shows that the company has already invested in heavy-asset production lines.
R&D; is also heavy. From 2023 to 2025, CXMT’s R&D; spending was RMB4.670bn, RMB6.341bn, and RMB9.593bn, respectively, totaling RMB20.605bn over three years. In 2025, it had 6,259 R&D; employees, accounting for 32.43% of total staff. This is not an asset-light design company, nor a chip brand relying only on outsourced manufacturing. It follows the DRAM IDM model.
One point needs to be made clear: accumulated losses do not mean the company has poor cash flow. In 2025, CXMT’s net operating cash flow was RMB36.520bn; in 1Q26, net operating cash flow was RMB42.566bn. The historical losses on the income statement mainly came from early heavy-asset investment, depreciation, R&D;, and the cyclical downturn. Now that prices have moved up, cash flow is already very strong.
The real issue to track is whether depreciation and R&D; can be diluted by a larger revenue base. In 2025, this had already happened: gross margin rose from 5.58% in 2024 to 40.99%. The 1Q26 and 1H26 estimates show that scale effects are continuing to expand.
VI. Customer Mix: High Concentration, But Not Yet Single-Customer Dependence
ChangXin’s customer concentration is not low. In 2025, the top five customers accounted for 68.08% of principal operating revenue, versus 67.30% in 2024 and 74.12% in 2023. For a DRAM vendor, this is not surprising. Memory customers are mainly large accounts across smartphones, servers, PCs, modules, and distribution channels, and the industry is inherently concentrated.
The prospectus discloses that the company has entered the customer systems of Alibaba Cloud, ByteDance, Tencent, Lenovo, Xiaomi, Transsion, Honor, OPPO, vivo, and others. These represent two types of demand: mobile devices, which require power efficiency, stability, and supply assurance; and servers and cloud providers, which require capacity, bandwidth, reliability, and long-term supply.
The high distribution share is one point in this prospectus that deserves attention. In 2025, distribution revenue accounted for 85.38% of principal operating revenue. Selling semiconductor products through distribution channels is common, especially for standardized products such as memory; distributors help the company reach downstream customers quickly and accelerate turnover. But a high distribution share also means end-demand, inventory, and the real pace of customer pull-ins require ongoing cross-checking.
ChangXin has not disclosed any single customer contributing more than 50% of revenue, and says it does not have severe dependence on a small number of customers. Investors should next watch whether the direct-sales customer share can gradually rise, whether server customers and leading cloud providers can continue to be penetrated, and whether AI server-related revenue can move from “still relatively low” to a genuinely sizable source of profit.
VII. RMB 34.5bn Fundraising Projects: Capital Mainly Goes to Upgrades, Not Storytelling
This IPO involves total project investment of RMB 34.5bn, with RMB 29.5bn of proceeds planned for use. The project direction is highly focused, all centered on DRAM technology upgrades, production-line transformation, and forward-looking R&D.;
These three fundraising projects correspond to ChangXin’s next three core questions.
First, whether unit costs can continue to decline. DRAM is a scale- and process-driven industry. If the same wafer can produce more qualified chips, costs fall. If ChangXin can improve wafer output, yield, and equipment efficiency through process upgrades, gross margin will not rely only on price increases.
Second, whether high-end products can continue to scale. DDR5 and LPDDR5/5X have already shown that product-mix upgrades can lift gross margin. The next step is whether server DDR5, high-capacity modules, low-power high-end mobile memory, and future products can continue to be introduced into customer platforms.
Third, whether supply-chain security can become stronger. The prospectus mentions the introduction of domestic and new types of equipment, materials, and components. For domestic DRAM, the equipment, materials, EDA, components, and packaging-and-testing ecosystems directly determine capacity expansion and long-term stable supply.
This capital will not immediately turn into profit. The fundraising projects will bring more depreciation, R&D;, and upfront expenses. The real investment judgment is whether they can deliver higher output, lower costs, a higher share of newer-generation products, and stronger customer qualification.
VIII. Offering Arrangement: Pricing Is Not Set Yet; For Now, Watch the Mechanism
ChangXin’s offering combines strategic placement, offline bookbuilding, and online subscription. The issuance ratio is 10.00%, with an over-allotment option. The specific share capital, greenshoe, and timetable are set out in the table below to avoid overloading the text with numbers.
The most important variable has not yet been disclosed: the offer price. Without the offer price, there is no definitive market capitalization, issuance P/E, or issuance P/B. EPS in the prospectus is also temporarily calculated using the lower of 2025 net profit attributable to shareholders before and after non-recurring items divided by pre-issuance total share capital. Post-issuance valuation must wait until the price is determined.
Therefore, it is not yet possible to say “ChangXin is expensive” or “ChangXin is cheap.” For now, the framework can be laid out clearly.
IX. Three Scenarios: This Report Cannot Only Discuss the Upside
The judgment on ChangXin Technology must be placed within three scenarios. Because the DRAM industry is cyclical, valuation based on a single earnings point can easily mislead.
Among these three scenarios, the most likely outcome is neither “permanent prosperity” nor “an immediate return to 2023.” A more reasonable judgment is that the first half of 2026 has already been a strong-cycle window, and the market will treat part of the profit as cyclical peak earnings. ChangXin needs to prove that profit comes not only from pricing, but also from product-generation upgrades, server customers, capacity utilization, and cost reduction.
10. Compared With Global Peers, CXMT’s Strengths and Weaknesses Are Both Clear
The global DRAM industry has long been dominated by Samsung Electronics, SK hynix, and Micron Technology. CXMT’s share has reached 7.67%, putting it on the radar of major players, but it remains in catch-up mode versus the top three.
CXMT has three advantages.
First, China’s market demand is enormous. China is one of the world’s major DRAM consumption markets. Smartphones, PCs, servers, cloud computing, and smart vehicles all require memory. As the leading domestic DRAM company, CXMT has natural advantages in supply security and localized customer coverage.
Second, its product generation has reached the mainstream. After DDR5 and LPDDR5/5X entered mass production, CXMT is no longer relying only on older-generation products for shipments. The DDR series’ 2025 gross margin of 41.89% already shows that new products can support better earnings quality.
Third, it is highly scarce in the A-share market. A-shares have semiconductor equipment, materials, packaging and testing, and design companies, but very few true scale DRAM IDM targets. After listing, CXMT will likely become the core A-share proxy for the memory cycle, domestic DRAM, and AI server memory.
The weaknesses are equally clear.
First, compared with Samsung, SK hynix, and Micron, CXMT still has gaps in global customers, capacity scale, advanced generations, HBM, and ecosystem coordination. Its current strength is more in commodity DRAM and mainstream DDR/LPDDR products. It should not be directly described as an HBM leader.
Second, its earnings remain highly exposed to DRAM prices. The sharp profit improvement in 2025 and 1Q26 benefited from global undersupply and price increases. If prices fall, gross margin will come under pressure.
Third, it still has accumulated uncovered losses. At the end of 2025, undistributed profit was RMB -36.65 billion. This will not prevent the company from operating, but it will affect dividends and the market’s assessment of earnings stability.
11. The Most Easily Misread Point: First-Half Profit Is Very High, but the Full Year Cannot Be Linearly Extrapolated
CXMT expects 1H26 net profit attributable to shareholders of RMB 50-57 billion, an eye-catching number. Many people will naturally multiply it by two and imagine full-year attributable profit of more than RMB 100 billion. But the prospectus has already stated that the first-half performance estimate does not constitute an earnings forecast or performance commitment.
It does not constitute an earnings forecast or performance commitment.
This sentence must be taken seriously.
DRAM prices do not move in a linear trend. Customer restocking, AI server pull-in, capacity allocation by global leaders, smartphone and PC demand, and inventory cycles all affect pricing. CXMT’s very high first-half profit shows that current pricing and product mix are very favorable; it does not automatically prove that the second half will be equally strong.
The right tracking method is to watch four numbers.
One post-issuance metric should also be added: market capitalization and PE implied by the issue price. If the issue price is low, the market may treat CXMT as a high-beta domestic DRAM leader. If the issue price is high, the market will ask more stringent questions about earnings sustainability.
12. Risk List: This Company Is Strong, but the Risks Are Not Soft
There is no need to talk around CXMT’s risks. The prospectus has already laid them out clearly. What investors really need to do is translate the risks into trackable numbers.
First, the DRAM price cycle. The root cause of CXMT’s 2023 loss was the DRAM industry downturn, a sharp price decline, inventory write-downs, and fixed-cost pressure. One of the root causes of the current profit surge is also price increases. The benefit and the risk in this industry come from the same variable.
Second, uncertainty in AI demand. AI server and cloud capex have lifted DRAM demand, but AI application deployment, improvements in model efficiency, and cloud spending cadence can all change. If AI capex slows, DRAM prices will be affected.
Third, capacity expansion by global leaders. When Samsung Electronics, SK hynix, and Micron see profit improvement, they will increase capital expenditure. New capacity will not be released immediately, but if future demand fails to keep up with supply, the industry will return to oversupply.
Fourth, depreciation and expenses from the fundraising projects. The fundraising projects can improve long-term competitiveness, but they will also increase depreciation, amortization, and expenses in the short and medium term. Only if capacity ramp-up, yield improvement, and product upgrades are fast enough will the projects translate into profit rather than expense pressure.
Fifth, the distributor share and customer concentration. A high distributor-model share can accelerate sales and turnover, but end-market real demand and inventory must be continuously verified. The top five customers account for close to 70%, making customer structure stability important.
Sixth, pressure to catch up technologically. CXMT has already achieved DDR5 and LPDDR5/5X, but global DRAM technology continues to evolve. Advanced processes, capacity, speed, power consumption, yield, module form factors, and future forward-looking DRAM technologies all require continued catch-up.
13. How to Define CXMT Before the Issue Price Is Announced
The most appropriate characterization of CXMT is: “China’s DRAM leader entering a profit realization window.” It is neither a pure concept stock nor a stable dividend asset. It is a capital-intensive, highly cyclical, technology catch-up semiconductor manufacturer with strong domestic scarcity value.
Looking only at 2025, CXMT had just emerged from losses. Revenue was RMB 61.799 billion, net profit attributable to shareholders was RMB 1.875 billion, and recurring net profit attributable to shareholders was RMB 5.316 billion. Historical accumulated losses had not yet been filled.
Looking at 1Q26 and the first-half estimate, CXMT resembles a cyclical leader with extremely strong earnings leverage. First-half net profit attributable to shareholders is expected to be RMB 50-57 billion, and recurring attributable net profit is expected to be RMB 52-58 billion, putting it in an extremely high profit range.
Looking longer term, CXMT also resembles a capital-intensive technology company still catching up with the global top three. The fundraising projects total RMB 34.5 billion, R&D; investment over three years was RMB 20.605 billion, and fixed-asset depreciation continues to increase. It needs continuous investment to maintain technology generations, capacity scale, and customer qualifications.
None of these three perspectives can be dropped. Looking only at domestic scarcity underestimates cycle risk; looking only at the 2023 loss misses the 2026 earnings shift; annualizing only first-half profit overestimates the stability of peak-cycle earnings.
XIV. After the Offer Price Is Set, First Clarify the Earnings Basis
ChangXin’s IPO is likely to see the widest valuation disagreement because different earnings bases produce completely different P/E ratios. The company was already profitable in 2025, but net profit attributable to shareholders was only RMB1.875 billion, while recurring net profit attributable to shareholders was RMB5.316 billion. In 1H26, expected profit suddenly jumps to more than RMB50 billion. Using 2025, the profit base still looks low; using annualized 1H26, earnings look very substantial. Both bases need to be considered. Looking at either one in isolation can easily be misleading.
Pre-IPO total share capital is 60.193 billion shares. Post-IPO total share capital is 66.881 billion shares, and 67.884 billion shares after full exercise of the greenshoe. On a rough post-IPO share-count basis, 1H26 net profit attributable to shareholders of RMB50-57 billion implies 1H EPS of about RMB0.75-0.85; simply annualized, that implies RMB1.49-1.70. After full exercise of the greenshoe, 1H EPS is about RMB0.74-0.84, or roughly RMB1.47-1.68 annualized.
This calculation should not be treated as a full-year forecast, but it helps frame market pricing once the offer price is released.
Once the offer price is released, it can be read this way:
First, if the offer price implies a very high P/E on 2025 earnings, that does not necessarily mean the pricing is unreasonable, because 2025 does not yet fully reflect 2026 pricing and DDR5 volume ramp.
Second, if the offer price implies a very low P/E on annualized 1H26 earnings, that does not necessarily mean it is cheap, because 1H earnings themselves sit within a strong upcycle window.
Third, mid-cycle earnings are the most useful reference point. The prospectus does not provide this figure directly. It needs to be calibrated later using 3Q26 and 4Q26 gross margin, pricing, and inventory impairment. ChangXin’s first post-listing re-rating will very likely revolve around whether 2026 earnings can be partially capitalized.
Fourth, A-share scarcity will increase market attention. ChangXin is not an ordinary chip design stock, nor a traditional consumer electronics component company. It is a rare DRAM IDM leader in the A-share market. Scarcity can support a valuation premium, but if the offer price has already fully priced in the strong 1H profit, subsequent earnings validation will become more demanding.
XV. ChangXin Post-Listing Tracking Checklist
After ChangXin lists, investors should not just watch the stock price every day. More importantly, they need to track whether fundamentals are continuing along the strong-cycle path outlined in the prospectus.
First, watch the offer price and post-IPO market cap. The offer price determines which earnings basis the market is using to value ChangXin. If the market cap already fully reflects annualized 1H26 earnings, stronger 3Q and 4Q data will be needed to support it.
Second, watch 3Q revenue and gross margin. The 1H26 profit range has already been provided, and 3Q will test whether elevated DRAM pricing is continuing.
Third, watch the revenue mix from DDR5 and servers. The real improvement in ChangXin’s earnings quality comes from volume growth in DDR5, servers, and higher-generation products. Mobile LPDDR is the base; server DDR5 is the upside.
Fourth, watch inventory impairment losses. If DRAM prices fall, inventory write-downs may re-enter the income statement. The lesson from 2023 should not be forgotten.
Fifth, watch the progress of fundraising investment projects. If technology upgrades and forward-looking R&D; proceed smoothly, they will support the next round of cost reduction and product upgrades. If progress is slow, depreciation and expenses will hit the financial statements first.
Sixth, watch global peer capex. The expansion pace of Samsung, SK hynix, and Micron will determine industry supply-demand after 2027. If global supply again overwhelms demand, ChangXin will also find it difficult to stay immune.
SK hynix Deep-Dive Update: Is It Still Expensive? DDR5 Takes Over Price Increases, Korean Exports Surge, and Samsung Catch-Up Risk
Samsung Electronics Deep-Dive Update: 2Q26 Core Profit Beats Expectations, HBM4 Exceeds USD1 Billion, and Valuation Reset After the First Mobile Loss
Micron Deep Dive: Can Peak Earnings Be Capitalized, and How AI Memory LTAs Are Rewriting the Cycle Discount
XVI. Conclusion: The Key Question for ChangXin Technology Is How Much of the High Profit Can Be Retained
ChangXin Technology’s prospectus already provides a clear picture of the company. It is the DRAM leader in mainland China, with three 12-inch DRAM wafer fabs and products covering DDR5 and LPDDR5/5X. Its global share had already reached 7.67% in 2025. The story is no longer simply “can China make domestic DRAM,” but rather “can China’s domestic DRAM leader generate stable profits in a global memory upcycle.”
Expected 1H26 net profit attributable to shareholders of RMB50-57 billion is the strongest financial anchor for this company during the IPO stage. This figure will attract the market, and it will also create debate. Optimists will focus on domestic scarcity, DDR5 volume ramp, AI servers, and tight supply; more cautious investors will focus on the DRAM cycle, elevated pricing, depreciation pressure, and historical losses.
The most prudent judgment is this: ChangXin has already proved that it can earn substantial profits during a strong upcycle. The next step is to prove that it can retain sufficient profit when the cycle rolls over. Before the offer price is released, there is no need to rush to conclude whether it is cheap or expensive. Only when the offer price, listed market cap, 3Q pricing, and server revenue mix are all available will ChangXin’s first round of market pricing have a real anchor.
This company is worth writing about and worth tracking. It is not a target that can be explained clearly in a generic domestic substitution article. After ChangXin Technology lists, the A-share market will, for the first time, have a DRAM leader of real scale as a sample. Every subsequent quarter will become a key window for observing whether China’s memory industry can move from capacity catch-up to profit realization.