IPO on KnightLi Blog

China's Two Memory Champions Head Toward IPO: Why CXMT and YMTC Are Moving to the Capital Market

Thu, 21 May 2026 08:38:23 +0800

China’s domestic memory industry has recently shown a signal worth watching: CXMT and YMTC, the two companies representing China’s DRAM and 3D NAND Flash efforts, are both moving toward the capital market.

According to public information, CXMT’s operating entity, Changxin Technology Group Co., Ltd., has resumed progress in its STAR Market IPO review and has been scheduled for a listing committee review. YMTC Holdings Co., Ltd. has completed IPO tutoring filing, officially starting its listing tutoring process. The two companies are at different stages, but together they point to the same shift: China’s memory industry is moving from an industrial catch-up phase into a phase of capitalization, scale expansion, and more visible long-term investment.

This article is not investment advice. It focuses on a more important question: why memory, why now, and what these two listings could mean for China’s semiconductor supply chain.

Two Main Lines

The so-called “two domestic memory champions” correspond to two different but equally important technology lines.

CXMT is focused on DRAM. DRAM is the core volatile memory used in servers, PCs, phones, and AI computing devices. It determines runtime throughput and capacity limits. As AI servers, cloud computing, and high-performance computing grow, demand for DRAM capacity, bandwidth, power efficiency, and stability also rises.

YMTC is focused on NAND Flash, especially 3D NAND. NAND is the core storage medium for SSDs, phone storage, enterprise storage, and data center tiering. It determines whether data can be stored with higher density, lower cost, and better stability over the long term.

In simple terms, DRAM lets systems “run,” while NAND lets data “stay.” AI data centers, smart devices, domestic servers, information technology localization, and edge devices all rely on these two types of chips.

That is why CXMT and YMTC are not ordinary semiconductor companies. They are part of the basic foundation of China’s computing infrastructure.

CXMT: A Key Review Stage for the STAR Market IPO

For CXMT, the current keywords are “review resumed” and “listing committee meeting.”

Public reports show that in May 2026, Changxin Technology Group Co., Ltd.’s STAR Market IPO review resumed and moved into the “inquiry” stage. The Shanghai Stock Exchange listing review committee is expected to review its IPO application on May 27, 2026. Earlier reports also mentioned a planned fundraising size of about RMB 29.5 billion, mainly for memory wafer manufacturing upgrades, DRAM technology iteration, and forward-looking R&D.

These details point to a clear reality: DRAM is a heavy-asset, long-cycle, highly volatile business. It requires continuous spending on process iteration, capacity ramp-up, yield improvement, equipment renewal, and customer validation. Short-term financing or industrial funds alone are unlikely to support a memory company through multiple cycles.

For CXMT, an IPO is therefore not just “raising money.” It is also a way for a long-term capital-intensive business to obtain a more stable funding channel and market pricing mechanism.

The challenges are also very real:

DRAM is highly cyclical, and prices can swing significantly with supply and demand.
International giants still have deep advantages in technology, yield, customer structure, and scale.
Advanced equipment, materials, EDA, packaging/testing, and supply-chain coordination remain affected by external conditions.
AI demand brings growth, but it also raises expectations for HBM, DDR5, LPDDR, server memory, and other product lines.

For CXMT, listing is only a stage marker, not the finish line. The real test remains technology roadmap, cost curve, customer validation, and cycle management.

YMTC: The NAND Line Begins IPO Tutoring

For YMTC, the current keyword is “tutoring filing.”

According to CITIC Securities’ announcement and public reports, on May 19, 2026, CITIC Securities signed a tutoring agreement with YMTC and submitted the tutoring filing application to the Hubei office of the China Securities Regulatory Commission. On the same day, the Hubei regulator accepted the filing. Public reports also show that YMTC’s IPO tutoring institutions include CITIC Securities and CSC Financial.

This means YMTC’s listing process has entered a more formal capital-market preparation stage. Unlike CXMT, which is already in the STAR Market review process, YMTC is still in the tutoring stage. It still needs to go through tutoring, application, acceptance, inquiry, listing committee review, registration, and other steps. The timetable remains uncertain.

YMTC’s industrial significance lies in 3D NAND. Competition in NAND is not about a single chip. It is about layer stacking, process capability, controller ecosystem, enterprise SSD reliability, customer certification cycles, and long-term supply capability across data centers, consumer electronics, and embedded markets.

Over the past few years, YMTC has become one of China’s most representative NAND companies. If its IPO proceeds smoothly, it can help supplement the capital needed for intensive R&D and expansion. It will also put the operating data, asset structure, customer mix, and R&D investment of China’s NAND line under more public scrutiny.

That scrutiny matters. Semiconductor progress cannot rely only on slogans or valuation stories. Real industrial progress eventually has to show up in products, financials, customers, cash flow, and sustained investment capability.

Why Both Are Moving Toward the Capital Market Now

This is not a coincidence.

First, AI is pulling memory demand upward again. When people talk about AI, they often focus on GPUs, inference chips, and large-model training. But in real data centers, storage and memory are also bottlenecks. Model parameters, vector databases, training data, logs, caches, retrieval augmentation, enterprise SSDs, and server DRAM all grow as AI applications expand.

Second, the memory industry is entering a new upcycle. DRAM and NAND are both highly cyclical. Prices, inventory, and capital expenditure influence each other. When the industry moves out of a trough and demand improves again, leading companies are more likely to attract attention from the capital market.

Third, localization has moved from “can we make it” to “can we scale it reliably.” In the early stage, the key question was whether domestic chips could be produced at all. Now the more important question is whether they can ship continuously, supply reliably, enter mainstream customers, reduce cost, and cross multiple technology generations.

Fourth, the capital market also needs hard-tech benchmarks. Compared with consumer internet or asset-light software, memory chips better represent long-cycle hard-tech investment. If CXMT and YMTC progress smoothly, they will have a strong demonstration effect for China’s semiconductor supply chain.

What It Means for the Supply Chain

If CXMT and YMTC continue their listing progress, the impact will go beyond the two companies.

Upstream equipment, materials, components, cleanroom engineering, specialty gases, targets, photoresists, packaging/testing, controllers, module makers, server vendors, and data center customers could all be pulled along. Memory chips sit inside an extremely long supply chain. Whenever a leading company expands investment, it creates validation opportunities and orders for upstream suppliers.

More importantly, listing forces the industrial story to become more transparent. The market will keep asking:

How much is actually spent on R&D?
How far have advanced processes and stacking technologies progressed?
What are capacity utilization and yield like?
Who are the main customers, and is customer concentration high?
Can gross margin survive cycles?
Can the balance sheet handle the next round of expansion?
Where are the remaining gaps versus global leaders?

These questions may sound sharp, but they are exactly the questions China’s semiconductor industry must face as it matures.

Avoid Excessive Optimism

The IPO push by China’s memory leaders is a positive signal, but it should not be simplified into “localization is complete.”

Memory chips are among the harshest semiconductor markets in the world. The sector requires massive capital expenditure, fast technology iteration, strong price cycles, volatile inventory, and intense competition from global giants. Even if a company makes steady technical progress, it may face profit pressure during a downturn. Even if demand is strong, it may still face uncertainty in equipment, materials, customer certification, and the international environment.

Ordinary investors should especially avoid translating “domestic memory” directly into guaranteed returns. What matters is the prospectus: revenue structure, profit quality, cash flow, inventory, capacity, customers, related-party transactions, R&D investment, and risk disclosures.

The industry is worth watching, but investment decisions need restraint.

Conclusion

CXMT and YMTC moving toward the capital market at the same time is a symbolic moment for China’s semiconductor industry.

CXMT represents China’s DRAM line, and YMTC represents China’s 3D NAND line. One is tied to runtime memory for computing, while the other is tied to long-term storage for massive data. Together, they form part of the most basic storage foundation for the AI era, the data center era, and China’s domestic computing system.

In the coming years, the key question is not “who lists first.” It is whether these companies can build a long-term positive cycle among capital, technology, capacity, customers, and supply chains.

If the keyword of China’s memory industry over the past decade was “breakthrough,” the keyword of the next decade may be “scale” and “cycle resilience.” Listing is only the knock on the door; the real exam comes afterward.

This article is based only on public information and is not investment advice. IPO progress, issuance arrangements, valuation, use of proceeds, and financial data should be based on official exchange documents, CSRC filings, company prospectuses, and formal announcements.

References

Behind Cerebras' IPO Surge: Can Wafer-Scale AI Chips Challenge Nvidia?

Mon, 18 May 2026 00:19:51 +0800

Cerebras Systems has finally entered the public market.

The company, known for its “wafer-scale AI chips”, began trading on Nasdaq on May 14, 2026 under the ticker CBRS. According to Cerebras’ official announcement, the IPO price was $185 per share, with 34.5 million shares of Class A common stock offered, including the underwriters’ full exercise of a 4.5 million share over-allotment option.

On its first trading day, Cerebras opened sharply higher and briefly approached $386. Based on the IPO price, the company raised more than $5.5 billion, making it one of the most closely watched AI hardware IPOs in the U.S. market in 2026.

That is why many media outlets call it an “Nvidia challenger”. But it is not accurate to simply describe Cerebras as “the next Nvidia”. What makes it unusual is that it has chosen a technical path very different from traditional GPUs.

Cerebras Is Not Building a Normal GPU

Cerebras’ core product is WSE, short for Wafer-Scale Engine.

Traditional chip manufacturing cuts a whole wafer into many small chips, then packages, tests, and ships them. Cerebras takes the opposite approach: it tries to turn an entire wafer directly into one giant chip.

The advantages of this route are straightforward:

Larger chip area.
More on-chip compute units.
On-chip SRAM closer to compute cores.
Shorter data movement inside the chip.
Better fit for certain AI inference and training workloads.

In AI computing, moving data is often harder to optimize than raw computation. Cerebras’ idea is to keep compute and storage on the same piece of silicon as much as possible, reducing the latency and energy cost caused by data repeatedly leaving the chip.

That is the most attractive part of the WSE approach. Instead of scaling along the same GPU path, it uses a much larger single chip to pursue higher on-chip bandwidth and lower data movement cost.

Why the Market Got Excited

The AI chip market is currently highly dependent on Nvidia. Whether companies are training large models, deploying inference services, or building AI data centers, Nvidia GPUs remain the mainstream choice.

That makes the market naturally interested in two kinds of companies:

Companies that can reduce dependence on Nvidia’s supply chain.
Companies that can offer higher performance or lower cost for certain AI workloads.

Cerebras fits both narratives.

It is not building a general-purpose CPU or an ordinary accelerator card. It designs systems directly around AI training and inference. The company has also repeatedly emphasized that its wafer-scale chips and cloud inference platform can deliver very high throughput in certain model inference scenarios.

This kind of story is easy for the market to amplify in 2026. AI infrastructure is still expanding, and enterprises, cloud providers, and model companies are all looking for more compute sources. If a chip company can prove that it is not just “another small GPU” in some scenarios, the market will pay attention.

The OpenAI Partnership Expands the Upside Story

Another reason Cerebras is closely watched is its relationship with OpenAI.

According to media reports, Cerebras signed a cooperation agreement with OpenAI worth more than $20 billion. The original Sohu article noted that, as of the end of 2025, the remaining performance obligations from that agreement reached $24.6 billion.

For a newly listed AI hardware company, such long-term agreements are important. They suggest that the company has not only a technical story, but also demand from major customers.

Still, long-term orders are not the same as realized revenue. AI data center deployment depends on manufacturing capacity, packaging, power supply, delivery schedules, customer budgets, and changes in model strategy. For chip companies, winning orders is only the first step. Delivering on time, scaling reliably, and building margins are harder.

Customer Concentration Remains a Major Risk

Cerebras also has an obvious risk: high customer concentration.

The Sohu article noted that G42 contributed 85% of Cerebras’ revenue in 2024, falling to 24% in 2025, while Mohamed bin Zayed University of Artificial Intelligence contributed 62% of revenue in 2025. This means that even after G42’s share declined, Cerebras’ revenue still depended heavily on a small number of large customers.

For AI infrastructure companies, customer concentration has two sides.

The benefit is that large customers can bring rapid growth, long-term contracts, and order visibility.

The risk is that if customers cut budgets, change technical direction, delay data center construction, or face regulatory changes, revenue volatility can be significant.

That is why Cerebras should not be judged only by its IPO pop. The first-day stock price reflects enthusiasm and expectations. Long-term valuation will still depend on revenue structure, delivery capability, margins, and customer diversification.

The Technical Limitation: Memory Capacity

WSE has clear strengths, but its limitations are also clear.

The Sohu article noted that the WSE-3 chip has 44GB of SRAM, while Nvidia’s B200 has 192GB of memory. Cerebras places a large amount of compute and SRAM on the same wafer, which reduces data movement, but also limits available memory capacity.

For large models, memory capacity directly affects context length, batch size, and deployment architecture. Context windows are getting longer, and flagship models are increasingly moving toward million-token context windows. In that trend, on-chip SRAM capacity becomes a real constraint.

Traditional GPUs can continue expanding memory through HBM stacking, packaging expansion, and multi-GPU interconnects. Cerebras’ wafer-scale approach is harder to expand in a simple way because the wafer area is already occupied by compute units and SRAM. Adding more SRAM may mean sacrificing compute area.

This does not mean the Cerebras architecture has failed. It means it is an architectural choice optimized for specific workloads. It may be very strong in certain inference scenarios, but it does not necessarily cover every AI training and inference need.

Can It Replace Nvidia?

In the short term, Cerebras is unlikely to replace Nvidia.

Nvidia’s advantage is not only GPU performance. It also includes the CUDA ecosystem, developer tools, system integration, networking, full-stack server solutions, cloud provider support, and customer migration costs. AI companies often choose Nvidia not because one chip wins on one metric, but because the entire ecosystem is the most stable.

Cerebras’ more realistic opportunity is to become a complementary option for specific AI workloads:

High-throughput inference.
Specific large-model services.
Tasks sensitive to latency and on-chip bandwidth.
Customers that want to reduce dependence on a single GPU supply chain.
Model companies willing to test new architectures for performance.

In other words, it is not an “Nvidia killer”. It is more like an aggressive alternative path in the AI compute market.

Summary

Cerebras’ IPO surge shows that capital markets are still willing to pay a high premium for AI infrastructure stories.

Its wafer-scale chip architecture is genuinely distinctive, separating it from ordinary AI accelerator companies. Together with major customer relationships such as OpenAI, Cerebras has a strong market narrative.

But the risks are just as real: customer concentration, delivery pressure, memory capacity limits, ecosystem barriers, and the system-level gap with Nvidia will all determine how far it can go.

For ordinary readers, the most interesting part of Cerebras is not how much the stock rose. It is that the company proves AI compute competition will not have only one GPU path. Future large-model infrastructure may include GPUs, wafer-scale chips, in-house accelerators, and cloud-based specialized inference platforms at the same time.