Starting Point:

Survival in the Cracks – The Beginning of Domestic Storage

In the storage industry, high technical barriers, large capital investments, and long R&D cycles have always been the three core thresholds restricting new players.Since the 1980s, the memory chip market has gradually formed a “tripolar pattern”: Samsung, SK Hynix, and Micron have occupied the main share, firmly controlling the core technology and market pricing power of memory chip manufacturing. Samsung’s 3D NAND, SK Hynix’s LPDDR, and Micron’s enterprise-level SSD solutions have monopolized the upstream and downstream of the consumer and enterprise markets in China.Faced with the monopoly of international giants, the difficulties faced by domestic storage in the initial stage can be summarized as “lacking technology, lacking talent, lacking funds, and lacking trust”.

Lacking Technology: In the field of memory chip design and manufacturing, domestic enterprises started late, and in the early stage, they were mostly concentrated in the low-end market, lacking advanced process and product design capabilities. The lack of technological accumulation made domestic manufacturers unable to compete with international giants like Samsung in the mainstream market.

Lacking Talent: The core technology R&D of the memory chip industry requires a large number of high-end talents, and at that time, China’s technical reserves and educational resources in the semiconductor field were still unable to meet the demand.

Lacking Funds: The R&D and mass production of memory chips require huge capital support, and the R&D investment of international giants is often dozens or even hundreds of times that of domestic enterprises. Samsung’s R&D investment in the storage field exceeds 10 billion US dollars every year, while the R&D budget of domestic manufacturers is only a fraction of it. At the same time, the memory chip industry has a significant economies of scale effect, and domestic enterprises with insufficient production capacity face great challenges in cost control and market competitiveness.

Lacking Trust: In the early market promotion, domestic storage brands were often regarded as low-end substitutes by consumers. Due to limited technical capabilities, the performance, stability, and durability of early products were significantly different from international brands. This brand trust crisis further restricted the breakthrough of domestic manufacturers in the high-end market.

Most early domestic SSDs relied on imported chips for packaging and production, with representative manufacturers such as Galaxy, Team, and Maxsun. Their products were often labeled as “low-end substitutes”. In early user feedback, domestic SSDs were “full of problems”.

Short Life: The early versions of flash memory particles had unstable life, often damaged after two or three years of use.

Poor Performance: Limited by the optimization capabilities of controllers and firmware, the continuous writing speed and random read-write performance of early SSDs were far inferior to international giants.

Uneven Quality: Some products even had large-scale repairs due to non-uniform production standards.

In 2015, a domestic storage company tried to launch its own branded SSD, but due to the use of outdated process technology, the product speed could not even catch up with Samsung’s mid-range model three years ago.

An engineer recalled: “Before the release, we were full of hope, but after the release, the reputation was negative, and even the advertising slogan became a joke in the industry.” Such failure cases were not uncommon at the time.However, domestic brands did not give up because of this. They gradually accumulated experience in the early trials and errors, laying the foundation for subsequent development.

Lurking: Policy Promotion and Technological Accumulation

At the same time, the “invisible hand” began to layout.

In 2014, the State Council issued the “National Integrated Circuit Industry Development Promotion Outline”, proposing to achieve autonomous and controllable goals in key areas of integrated circuits by 2030, with memory chips listed as one of the priority support directions. Under this policy background, the domestic storage industry ushered in unprecedented development opportunities.

Establishment of National Fund

To promote the development of the semiconductor industry, the National Integrated Circuit Industry Investment Fund (referred to as the “Big Fund”) was established in 2014, with a total fundraising of 130 billion yuan. The Big Fund provided key capital support for domestic storage manufacturers, especially the continuous investment in Yangtze Memory, which laid the foundation for its technology research and development and capacity expansion.

Local Government Support

In addition to national policies, local governments also support the development of local storage industries by setting up special funds, providing tax incentives, and talent introduction plans. For example, Yangtze Memory’s headquarters is located in Wuhan, and the Hubei provincial government has provided a number of policy supports for its project construction, helping it to start quickly in terms of funds and resources.

Technical Blockade Stimulates Innovation

The intensification of technology competition across the ocean has made Chinese enterprises face stricter technical blockades. This external pressure has instead stimulated the independent innovation motivation of domestic manufacturers, making them have to take a different technical path from international giants.

Establishment and Technical Path Selection of Yangtze Memory

As the leading enterprise in the domestic storage industry, Yangtze Memory has established a strategic direction of “starting from basic technology research” since its establishment in 2016.

The early goal of Yangtze Memory was not high, but focused on the technical verification of a 32-layer 3D NAND chip.Although this product’s performance is hard to compare with Samsung’s 64-layer chip at the same period, its significance lies in that it is the first time for Chinese storage enterprises to achieve autonomous control from design to manufacturing.

In 2018, Yangtze Memory released the world’s first Xtacking architecture technology. This technology not only improves the performance of memory chips but also significantly reduces manufacturing complexity and cost by separating the storage units and peripheral circuits and integrating them with vertical interconnection technology. This technological breakthrough has made Yangtze Memory shine in the global storage market.However, at the beginning of technology research and development, the challenges were huge.

In 2017, when the first generation of Xtacking chips was trial-produced, the yield was very low, and the product was almost difficult to mass-produce.

An engineer mentioned: “At that time, the lights of the whole building were often on until two or three o’clock in the morning, and many people even worried whether this technical direction was feasible.”However, after three years of continuous optimization and attempts, Yangtze Memory launched its first self-developed 64-layer 3D NAND flash memory chip in 2019 and quickly applied it to the consumer SSD market.

Although there is still a gap in performance and reliability compared with international brands such as Samsung, its cost performance advantage and local supply chain strategy have won a certain market share for it, and also made domestic storage have the foundation to compete with international giants for the first time.

Breaking Through: Technological Breakthroughs and Market Expansion

In the development process of Yangtze Memory, the Xtacking architecture is the most important technological milestone. The introduction of this technology not only marks Yangtze Memory’s transition from technological catch-up to independent innovation but also changes the global NAND flash memory industry’s technological competition landscape.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

The Xtacking architecture adopts a separate design for storage cells and peripheral circuits, efficiently integrating the two parts through vertical interconnection. This technological path brings many advantages.

Performance Improvement: Xtacking optimizes signal transmission speed, achieving higher random read-write speeds and lower latency. Compared with traditional designs, its performance improvement can reach more than 20%.

Manufacturing Efficiency: This architecture simplifies the production process, allowing wafer manufacturing and packaging testing to be carried out simultaneously, significantly shortening the production cycle.

Design Flexibility: Storage cells and peripheral circuits can be optimized separately, supporting higher-density storage cell stacking and increasing chip storage capacity.
With the support of Xtacking technology, Yangtze Memory quickly launched multiple generations of products.

In 2019, the first 64-layer 3D NAND chip was released, laying the foundation for the mass application of domestic NAND products.In 2020, the 128-layer 3D NAND was launched, reaching a global advanced level.

In 2022, the 176-layer 3D NAND was successfully developed, beginning to enter the high-end storage market and directly competing with Samsung and other international giants.These technological breakthroughs have enabled Yangtze Memory to have real competitiveness in the global market for the first time and have created favorable conditions for the rise of its retail brand, Zhitai.

Unlike early domestic SSDs, Zhitai has achieved a qualitative leap in quality, performance, and after-sales service.When the Zhitai SSD product TiPlus5000 was released, the market response was positive. Although it is still not as fast and performant as Samsung, its price is more attractive, and the cost-performance advantage is obvious.

In 2023, Zhitai released the TiPlus7100 series. This product is equipped with Yangtze Memory’s 128-layer 3D NAND chip, with continuous read-write speeds reaching 7100MB/s and 6500MB/s, not only catching up with Samsung’s flagship products but also performing excellently in terms of life and stability.

Once released, this product caused a warm response in the domestic market and became a star product on the JD platform.

In 2024, the performance of Zhitai during Double 11 became an important symbolic event in the development of the domestic storage industry:online SSD category Double 11 promotion total transaction amount (GMV) and sales double champion!

Zhitai’s JD transaction amount increased by 40% year-on-year, and total sales increased by 15% year-on-year. Among them, Zhitai TiPlus 7100 became one of the most popular explosive SSD products on the JD platform.

Specifically, the main reasons for Zhitai to surpass Samsung during Double 11 slaes promotion can be

Attributed to the following key factors:

Significant Price Advantage: During the 2024 Double 11 period, Zhitai attracted a large number of consumers through direct discounts and full reduction activities. During Double 11, many international first-line brands did not have price reduction measures, and even a certain brand had the operation of raising prices first and then returning to the original price.

Technology Empowers Promotion: Through social media, live streaming, and other forms, the advantages of Xtacking technology are transformed into selling points that consumers can perceive (such as faster transmission speeds, longer service life).

Service and Channels: Through JD’s self-operated and official after-sales support, the problem of low consumer trust in early domestic brands was solved.

Behind these data is a comprehensive breakthrough of Yangtze Memory and Zhitai in technology, brand, and market.

Rise: The Logic Behind and Industry Significance

The fundamental reason for Yangtze Memory’s rapid rise is its persistence and breakthrough in technological innovation.

The release of the Xtacking architecture has broken the traditional design thinking of the storage industry, raising the performance and manufacturing process of NAND flash memory to a new level. In the global storage industry, the core of NAND flash memory technology is dominated by companies such as Samsung, Micron, and Intel, and domestic storage has always been in a technological catch-up situation.

However, Yangtze Memory has successfully filled the technological gap through independent research and development and technological breakthroughs, achieving the goal of “autonomous control.” From the successful development of 64-layer, 128-layer, to 176-layer 3D NAND flash memory, Yangtze Memory has not only solved the technical shortcomings of domestic storage but also broken the technological blockade of international giants, proving that China has the ability to compete with the world’s top enterprises in the field of semiconductor storage.In addition, its technology accumulation and rapid iteration strategy are also crucial.

In just a few years, Yangtze Memory has continuously optimized the Xtacking architecture, introduced new products, and launched multiple generations of products. This ability to rapidly iterate technology allows it to quickly adapt to market demand changes, ensuring product competitiveness. At the same time, Yangtze Memory’s large-scale production has also helped it maintain a technological lead, ensuring domestic market supply and competitiveness.

Policy Support and Industry Environment

The “invisible hand” has been paying increasing attention to the semiconductor industry. From the establishment of the “Big Fund” to the local government’s support for the storage industry, national policies have provided great financial and policy support for domestic storage manufacturers. These supports not only help enterprises with technology research and development but also provide a more relaxed market environment for them.In addition to national policy support, local governments have also played an active role in promoting the development of the semiconductor industry.

As a key project supported by Hubei Province and Wuhan City, Yangtze Memory has received full support from local governments in terms of funds and talent introduction. This close cooperation between government and enterprises has prompted the domestic storage industry to complete the leap from technological catch-up to leadership in a short period.Moreover, the collaboration of the upstream and downstream of the industry chain has also provided a solid foundation for the development of domestic storage. The continuous improvement of materials, equipment, and packaging testing links has greatly enhanced the competitiveness of the domestic storage industry.

From the initial technological lag and market downturn to breaking through the encirclement through policy support, technological innovation, and brand promotion, domestic storage enterprises have gradually become an important force in the global storage industry. The success of Yangtze Memory and Zhitai marks the technological breakthrough of the domestic storage industry, but future competition is still full of unknowns.

On November 27th, Samsung Electronics announced routine personnel changes for the 2025 class of presidents, totaling 9 individuals, with 2 being promoted to president and 7 experiencing role changes.

Samsung, in crisis, has implemented a reorganization of its presidency, focusing on the memory semiconductor business.

Moving forward, the memory business will be directly managed by the CEO and vice chairman, without a separate president position.

In the regular 2025 president reshuffle, Samsung Electronics appointed Vice Chairman and head of the Device Solutions (DS) division, Jeon Young-hyun, as CEO, aiming to revive the semiconductor business.

The Device Experiences (DX) division remains under the leadership of CEO and Vice Chairman Jong-hee Han, with Vice Chairman and head of the Business Support Task Force (TF) Hyun-ho Jeong retaining his position, solidifying the vice chairman system.

The personnel changes include: transforming the memory division into a system directly under the CEO’s jurisdiction, replacing the foundry (contract manufacturing) business leader, and appointing an experienced and mature CEO to manage new businesses. To overcome uncertain internal and external business environments and achieve new leaps, a personnel reform has been announced, including the allocation and mining of tasks.

Additionally, to strengthen semiconductor technology competitiveness and renew the organizational atmosphere, we have established a president-level CTO position in the foundry division and a president-level management strategy position directly under the DS division, empowering senior presidents with authority. Challenges such as brand and consumer experience innovation have been introduced to improve the company. The company announced a focus on enhancing its medium and long-term value.

Among them, Jinman Han, Vice President of the DS Americas (DSA) division responsible for the semiconductor business, has been appointed as the president of the foundry business division. Jinman Han has served as the head of DRAM and flash memory design teams, SSD development teams, and the Strategic Marketing Office, and was appointed as the head of the Americas region at the end of 2022, leading the semiconductor business at the forefront. With his combination of technical expertise and business acumen, along with extensive experience dealing with global customers, it is expected that he will enhance the competitiveness of the foundry business through process technology innovation and strengthening networks with key customers.

Kim Yong-gwan, a member of the Business Support Special Task Force (TF) and vice president, has been promoted to the position of President of Management Strategy for the DS division. Kim Yong-gwan, after working in semiconductor planning and finance, and in the Strategy Group and Management Diagnostics Group of the Future Strategy Office, moved to the Business Support TF in May, responsible for semiconductor support, and is expected to play a leading role in the early recovery. It is reported that Samsung has indicated an intention to enhance semiconductor competitiveness through this forward deployment.

Vice Chairman Jeon Young-hyun, who is also the head of the DS division, serves as Samsung Electronics’ CEO and head of the DS division, memory business division, and SAIT President.

The CEO of Samsung Electronics has traditionally been a vice chairman. In particular, the DS division is generally led by executives such as Chang Kyu Lee and Ki Nam Kim.

Under the leadership of former DS division head Chang Kyu Lee, he unusually held a president-level position, but in May of this year, Vice Chairman Jeon Young-hyun took over the organization again and elevated the status of the president.

Samsung Electronics is expected to shift from a single CEO system led by the DX division head (Vice Chairman) Han Jong-hee to a dual CEO system, including the former vice chairman, thereby enhancing the competitiveness of the semiconductor business. Samsung Electronics explained that the reason for restoring the two-person system of CEO and vice chairman is “to establish a business responsibility system for each division.”

Nam Seok-woo, CEO of the Global Manufacturing and Infrastructure Headquarters of the DS division, has been transferred to the position of Chief Technology Officer (CTO) of the foundry division.

It is noteworthy that a direct system with the CEO also serving as the head of the business division has been established.

Most importantly, as Samsung Electronics’ memory division has ceded leadership in the high-bandwidth memory (HBM) and other artificial intelligence (AI) memory markets to competitors, a seasoned senior leader seems to have issued a special command to directly lead the division.

Jong-hee Han, Vice Chairman and head of the DX division and Home Appliances (HA) division, will also serve as the chair of the newly established Quality Innovation Committee.

This is seen as a determination to fundamentally prevent quality disputes over Samsung Electronics’ products, in light of recent quality controversies surrounding the Galaxy Buds.

Lee Young-hee, head of the Global Marketing Office and Global Brand Center of the DX division, has been appointed as the chair of the Brand Strategy Committee.

Wonjin Lee has been appointed as the head of the Global Marketing Office of the DX division. Wonjin Lee, a Google advertising and services business expert consultant, resigned from his position as the head of the service business team in the Mobile Experience (MX) division at the end of last year and will return to the management frontline after a year, serving as the head of the global marketing department, overseeing marketing, branding, and online business.

Koh Han-seung, CEO of Samsung Bioepis, has been transferred to the position of head of Samsung Electronics’ Future Business Planning division.

Park Hak-gyu, head of the Management Support Office of the DX division, has been transferred to the position of president, responsible for Samsung Electronics’ business support TF.

Samsung Electronics usually appoints presidents in early December, but this year, like last year, it was advanced by about a week.Previously, Samsung Electronics Chairman Lee Jae-yong stated in his final remarks at the second trial on November 25th, “I am well aware that there has been a lot of concern about the future of Samsung recently, and the reality we are facing now is more difficult than ever, but this is a difficult situation.

I will definitely overcome this and take a step forward.” He continued, “Please give us the opportunity to overcome this difficult situation and become a Samsung loved by the people.”

Therefore, following the personnel changes of the chairman, there is a high possibility of executive personnel and organizational restructuring in the near future.

Due to unfavorable internal and external situations, the scale of executive promotions is expected to be reduced compared to previous years.Samsung Electronics plans to complete personnel and organizational restructuring and hold a global strategy meeting in mid-December to discuss next year’s business plans.

SK Hynix’s global NAND flash memory market share is expected to surpass 20% for the first time this year. SK Group Chairman Chey Tae-won has decided to take personal charge of the NAND business, which is expected to accelerate the process of narrowing the gap with market leader Samsung Electronics.Industry insiders stated on November 18th that SK Hynix’s global NAND market share, calculated by sales (according to IDC statistics), has increased by more than 10 percentage points over the past four years, growing from 11.7% in 2020 to 22.5% in the second quarter of this year. Although SK Hynix acquired Intel’s NAND division and launched Solidigm in 2021, the expansion of its market share was not as rapid as anticipated, rising from 13.7% in 2021 to 19% in 2022, but almost stagnating at 19.2% last year. The semiconductor economic downturn had a significant impact on Solidigm’s ongoing deficits.

Solidigm has steadily reduced its deficits and successfully turned a profit in the second quarter of this year (net profit of 7.86 billion won). Due to increased investment in AI data centers, the demand for high-capacity enterprise SSDs (eSSDs) has surged, and Solidigm’s competitiveness in this market has been rewarded. Only Samsung Electronics and Solidigm possess quad-level cell (QLC) NAND technology. NAND is categorized into SLC, MLC, TLC, and QLC based on data storage methods. Compared to SLC with the same cell, QLC can store four times more data, easily achieving high capacity and improving production cost efficiency.Due to sluggish demand in PCs and mobile devices, SK Hynix’s NAND shipments in the third quarter decreased compared to the previous quarter. However, profitability was maintained at a level similar to the second quarter because the average selling price (ASP) of high-value eSSDs increased by about 10% from the previous quarter. eSSDs accounted for more than 60% of SK Hynix’s NAND sales in the third quarter.Thanks to the continued strong demand for eSSDs, SK Hynix’s annual NAND market share this year is expected to exceed 20% for the first time. During a conference call to announce the third-quarter results, SK Hynix stated, “Although the NAND market share calculated by volume may decrease this year, we anticipate an increase in the market share calculated by sales compared to last year.”SK Hynix is poised to accelerate the reduction of the market share gap with Samsung Electronics, whose NAND market share is in the mid-30% range. SK Hynix aims to increase its market share with a focus on high-value products such as high-capacity eSSDs. The company recently launched a new eSSD product capable of achieving 122TB, the largest capacity for NAND solutions. There are also predictions that with Chairman Chey Tae-won recently taking over as the chairman of the Solidigm board, the NAND business will be further strengthened.

Xtreme 200E SSD，Today’s topic is about Huawei’s new solid-state drive (SSD). We have obtained a prototype of Huawei’s SSD that has not yet been released (it will be officially launched at the end of November). However, there is essentially no difference between this engineering sample and the final version, except for the serial number. It’s a 1TB single-sided chip SSD. Let me tell you, getting this SSD was quite an effort. Let’s see what secrets are hidden under this label. It’s just a regular sticker and does not serve any cooling function.

For the 1TB capacity, there are two chips and one controller, which is a standard DRAM-less solution. It’s not easy to distinguish the manufacturer of the chips from the numbering. Without further ado, let’s compare it with the Zhitai Ti600, which is also a popular SSD sold online. Through comparison, we can confirm that Huawei’s new SSD definitely does not use the Lian Yun controller, and it is even less likely to use controllers from Samsung, Phison, or Silicon Motion. So, there should be only one answer in our minds.

Warranty

Nice, in terms of warranty, Huawei offers a 400TBW warranty for the new SSD, coincidentally, the Zhitai Ti600 also offers a 400TBW warranty. It is also unlikely that Huawei would use storage chips from Micron, Samsung, or Kioxia. So, the answer in front of us might be the one we all have in mind. In terms of price, I estimate that it should not be too expensive.

Testing

The SSDs tested together are: Kioxia RC20 1TB and Western Digital SN850 1TB. First, we clarify that this test will not do Huawei’s SSD any injustice. Both SSDs tested alongside are flagship SSDs with cache. Looking at the basic parameters, the PCIe-4.0 x 4 full-speed is flawless, with sequential read speeds of 7100 and sequential write speeds of 6300, which is a pretty good performance. Digital novices like to focus on this parameter, and the commonly used 4k91mb and 286mb performance is basically enough for daily use. This data is also basically consistent with Zhitai’s Ti 600, and we need to look at the 200GB+ mixed data write test.

SSD Cache

Why test 280GB? Because I initially thought it would have a 200GB cache, like the Zhitai Ti600. But after testing, the speed did not drop, and the cache was quite large. We directly increased it to 280GB to test the cache. Huawei’s SSD has a full-disk simulated cache of 250GB, and data exceeding 250GB will experience a decrease in transfer speed, and it will experience a speed drop. A 250GB simulated cache in an empty state is indeed very large, and it is completely sufficient for our daily home use. However, it is still a QLC chip after all.

After the Huawei SSD is half full, the mixed test shows that the 200E SSD only lasted 20GB before the cache was released, which is a common issue with QLC DRAM-less SSDs. That’s just how it is. Don’t understand? Let’s look at the actual performance. When we control the disk and copy 200GB of data, the read speed is relatively stable, basically around 3.6GB, and the entire copy time is around 55 seconds. The 200GB is copied, very fast, no problem at all.

Application Scenarios

If your SSD is already half full, and you copy files again, you will find that it is very easy to release the cache at this time, and the speed will be similar to your home mechanical hard drive when the cache is released. Therefore, the Huawei 200E SSD, because it is equipped with QLC chips, although its controller is strong, the storage chips are too ordinary. The most suitable group of people is ordinary home users who do not store too many things, and the 1TB capacity can be used for several years. So, this SSD is no problem at all, so it feels like the storage chips are holding back the controller.

Self-developed Chips

However, considering the current market environment and what tricks Trump might pull after he comes to power across the sea, no one knows what kind of environment it will be after Trump comes to power. We should encourage any brand or group in China that wants to do something from 0 to 1. Products need to be developed one by one, and the market needs to be developed bit by bit. We also hope that Huawei will release some PLC or higher-performance SSDs in the future, after all, the step from 0 to 1 is the most critical. This is what I mentioned at the beginning of the article, that the price of this SSD should not be too expensive. So, above is the prospective review of Huawei’s new 200E SSD. If you have any opinions on Huawei’s new SSD, you can also leave a comment. I am Teacher Alex, see you next time.

In 1995, Nvidia’s first chip, the NV1, faced numerous setbacks, leaving the company with only 30 days of working capital. By 2024, Nvidia’s Blackwell GPU was in such high demand that the company’s market value had surpassed $3.6 trillion. Also in 1995, Intel held over 75% of the global PC processor market, establishing its leadership in the personal computer market. Fast forward to 2024, Intel’s stock price has plummeted nearly 60% since the beginning of the year, with its market value dropping to over $80 billion, marking the first time in three decades that it has fallen below the $100 billion threshold. The changes in the market values of these two semiconductor giants are a microcosm of the current semiconductor landscape.There is a saying that goes, “People tend to overestimate the changes that happen in a year, but underestimate the changes that happen in five years.” Looking back, the changes in the semiconductor industry in 2024 have been greater than the sum of the changes over the previous five years. 2024 will be a turning point for the semiconductor industry. Why do we say that? Because, as an industry, the semiconductor sector is inseparable from the basic characteristics that define it: division of labor, products, and services. And products like computers and chips cannot be separated from the science behind them. The industry and science are interdependent, and even before the 1970s, the industry was clearly dependent on science. (The controversy over the Silliman affair was a watershed in the relationship between science and industry.) The significant changes in the semiconductor industry this year are the result of a three-way tug-of-war between technology (science), policy, and the market.

01 The Transformation of Semiconductor Companies

In 2024, there have been many news items that have stimulated the nerves of industry insiders. The most prominent is still the surging market value of Nvidia. Microsoft and Apple have been the two companies with the highest market value in the world since 2010. Although the two have been “entangled” in the ranking of “U.S. stock market leader” for more than 10 years, no one else has replaced them. On June 18, 2024, local time, Nvidia’s stock price rose by 3.51%, closing at $135.58, with a total market value of $3.335 trillion, surpassing both Microsoft and Apple to become the company with the highest market value in the world. This is a change that affects the whole world, such as in August 2011 when Apple first defeated ExxonMobil to win the title of the world’s most valuable company, which was also a symbolic moment for technology companies to defeat traditional oil companies. And Nvidia becoming the world’s most valuable company represents a shift in the focus of technology industry development. Traditional fields represented by Apple, such as consumer electronics and software, have led an era in the past, while Nvidia’s first place means that the focus of technology industry development is accelerating towards the AI field. In addition, on November 8th of this year, Nvidia replaced Intel as a component of the Dow Jones Industrial Average. S&P Global stated that the adjustment of components is to ensure that the index has a more representative semiconductor exposure. In fact, this is also proof: AI has truly entered the commercial track, and the impact of AI on the semiconductor industry is growing. In the semiconductor industry in 2024, in addition to the continuously increasing upper limits, there are also constantly refreshed lower limits. There are two giants in the semiconductor industry, which are well-known to everyone: Intel and Samsung. Since 1992, Intel has been the world’s largest semiconductor manufacturer and has been leading for 25 years. Although Samsung’s chip business surpassed Intel for the first time in 2017, it is still a constant struggle for the leading position. This year’s two chip giants seem to be stuck in a “swamp,” constantly struggling. As mentioned earlier, Intel’s market value has fallen due to its continuously losing performance. In the second quarter of this year, Intel’s loss reached $1.6 billion, far higher than the loss of $437 million in the previous quarter. And the latest third-quarter financial report announced its largest quarterly loss in its 56-year history – $16.6 billion (about 118.2 billion yuan). Samsung is also in a difficult situation in 2024. In October, Samsung announced that its third-quarter operating profit was about 91 trillion won, lower than the market expectation of 115 trillion won. Jeon Young-hyun, head of Samsung’s electronic device solutions department, also issued an apology, stating that the company’s performance did not meet market expectations, causing people’s worries about the company’s basic technical competitiveness and the future. As the senior management leading the company’s operations, they will take responsibility. (It seems that South Korea has also learned the apology system from Japan.) Similarly, the stock market reflects the company’s situation. Samsung’s stock price fell to 51,700 won on November 14th, local time, a record low since June 24, 2020. If this trend continues, Samsung’s stock price for the whole year will be the worst performance in over 20 years.

02 Europe’s “Decadence”

The changes in 2024 are not only the obvious corporate rankings but also regional changes. Amid global government attention to the semiconductor industry, there is a region where semiconductors have been declining throughout the year – Europe. Many people may not have noticed, but the data from WSTS cannot be deceived.

We have calculated the chip sales announced by WSTS since January of this year, and so far, the chip sales in Europe for the whole year have been negative year-on-year. Of course, there is also Japan, but Japan’s year-on-year growth has slightly turned positive in August and September. The European semiconductor industry fell into a weak state in 2024. The “three giants” of the European semiconductor industry: Infineon, STMicroelectronics, and NXP, all have poor revenue. Infineon’s revenue in the third quarter reached 3.702 billion euros, a year-on-year decrease of 9%. Currently, Infineon has lowered its performance outlook for 2024, adjusting this year’s revenue expectation to 15.5 billion to 16.5 billion euros, lower than the previous expectation of 16.5 billion to 17.5 billion euros. STMicroelectronics’ net revenue in the third quarter was 3.25 billion US dollars, a year-on-year decrease of 26.6%. NXP’s revenue decreased by 5.4% to 3.25 billion US dollars, slightly lower than the analysts’ expected 3.26 billion US dollars. Nowadays, Europe’s automotive and industrial markets have already shown negative growth.
From a more specific level of observation, the performance of optoelectronics and discrete devices is not satisfactory, the MCU market shows a shrinking trend and has shown negative growth, and the analog market has also experienced a long period of decline.

03 Tracing Back to the Source

The only constant in the world is change itself. This year, there have been so many changes in the semiconductor industry. Let’s trace back to the root causes and examine policies, markets, and technology.

The impact of policies on the industry is self-evident. At the policy level, the United States, Europe, Japan, South Korea, and Vietnam have all released semiconductor-related policies in the past two years. However, looking at the current situation, the implementation of the semiconductor policies in the United States and Europe has not been significant.

So far, more than half of the $52 billion subsidy in the U.S. Chip Act has been allocated, with over $35 billion granted to about 26 projects. Yet, the funds seem to be slow in coming, with Intel constantly complaining, “It’s been two years, and we haven’t seen a penny.”

Policy

The U.S. Chip Act is even more precarious after Trump’s potential return to office next year, as he is known to be opposed to the act. In interviews, Trump has bluntly stated, “It’s terrible.” The European Chip Act has not made any waves either. Proposed in 2022, there are not many factories being built in Europe, and Intel’s project has not even started. However, it is important to note that the Chip Act has played a role in attracting investment without actual capital. Companies like Intel, TSMC, Samsung, Micron, GlobalFoundries, and Amkor have already begun constructing factories in various regions of the United States due to the subsidies from the Chip Act. Even if the timeline for financial assistance is extended or the subsidies are canceled, it is unlikely that the partially constructed fabs will become abandoned projects. It can be said that for semiconductor companies, building factories in the United States is already an inevitable action. Although the Chip Act has not been fully implemented and Trump denies its effectiveness, there is no doubt that he supports the repatriation of advanced manufacturing to the United States. This is why, when looking at Intel, we also believe that Intel will not be as bad as it seems. Despite Intel being constantly criticized, we cannot ignore its significant role on the geopolitical stage and as one of the U.S. government’s chips in the semiconductor industry and foreign policy strategies. When considering Intel’s prospects, we must not overlook the elements of U.S. strategy and geopolitics. The semiconductor industry now occupies an important strategic position, which is common knowledge. Regardless of whether it can reach the extreme 3nm process, globally, there are only three companies capable of manufacturing advanced process chips: TSMC, Intel, and Samsung. A significant part of the U.S. Chip Act’s plan is directed towards Intel’s construction of two large factories in Arizona and the building of two more factories in Ohio. Therefore, Intel’s future is still full of Highlights. It is not appropriate to directly assume that it will “decline to the end.”

Market

Why has the term “involution” started to be used domestically in recent years? It is because the pace of technological innovation has slowed down, and with the impact of the three-year pandemic, China’s economic development speed has also slowed down under various pressures, leading to “involution.” The “second curve” theory believes that industrial development has a life cycle, and any growth curve is a parabola that rises and then falls. The secret to sustained growth is to start a new growth curve before the inflection point appears, thus forming a continuous improvement and development trend with the new and old momentum in sequence. We are now in the depression period of the last round of technological and industrial revolutions. To make the economy prosperous again, a new round of technological and industrial revolutions must be ignited.This is the inflection point brought by the cycle. The new round of explosion is undoubtedly in AI, which is also a consensus among many industry insiders.

Tech

AI brings new growth points to the market, with many technologies hidden under AI. Above AI, we need to see the development of technology. The outbreak of AI this time is undoubtedly driven by technology. The emergence of GenAI has been brewing for a long time. And this outbreak of AI has driven other technologies that were brewing. The first representative is: TSMC’s advanced packaging process, CoWoS. The two most concerned products in the semiconductor industry: GPU, HBM, both of these products rely on advanced processes to be produced. And TSMC’s CoWoS packaging plays a very key role, so that some people have begun to jokingly call “TSMC will be the world’s first packaging and testing factory.” Here, the packaging and testing, which originally belonged to the back-end process, has shown a clear trend of front-end. And in the process of HBM suppliers continuously improving the stacking density, the interconnection between the fifth-generation HBM (i.e., HBM4) and the logic layer in the future will be jointly completed by storage giants and foundries. According to SK Hynix, HBM4 will greatly change the industry’s perception of DRAM as a “universal chip,” turning it into a customized storage special process chip. In this process, the related technical know-how will be jointly undertaken by foundries and storage factories, which is also the meaning of TSMC’s “Foundry 2.0” plan this year. The second representative is: the architecture competition, that is, the market competition between x86 and Arm, RISC-V. This year, it is obvious that the market share of Arm architecture chips is increasing. The x86 alliance has made it very clear not long ago: face to face. Although it seems to be a change in the market among the three architectures at first glance, it is actually a thirty-year-long competition between CISC (Complex Instruction Set) and RISC (Reduced Instruction Set), these two chip design philosophies. This competition is essentially an extension of the development of computer architecture as a “science and technology” in the “industry” field. The emergence of AI has given these two design philosophies more application scenarios. From the current point of view, the architecture under the RISC architecture (Arm, RISC-V) indeed has an innate advantage because it has the low power consumption and high energy efficiency that AI naturally needs. This also means that Arm may one day achieve a higher market share, which is driven by both technology and the market.

According to Kioxia (formerly Toshiba Memory), driven by the robust demand for data storage from artificial intelligence, NAND flash memory demand is expected to grow 2.7 times by 2028. for the future needs, Kioxia will introduce new process technologies and further expand production capacity in the coming years to meet the impending surge in NAND demand.

According to Kioxia (formerly Toshiba Memory), driven by the robust demand for data storage from artificial intelligence, NAND flash memory demand is expected to grow 2.7 times by 2028. for the future needs, Kioxia will introduce new process technologies and further expand production capacity in the coming years to meet the impending surge in NAND demand.

Japanese media reports: reports indicate that Kioxia is expanding its production capabilities in Japan to support future growth. In particular, Kioxia is advancing the construction of its Kitakami plant in Iwate Prefecture, aiming to commence production in the fall of next year (2025). Originally, the plant was scheduled to begin production last year, but due to a decline in the industry’s demand for flash memory, its production schedule has been repeatedly altered. The new plant’s capacity, combined with Kioxia’s production capacity in Yokkaichi, will provide Kioxia with sufficient capacity to meet future market demands. In October, due to the negative sentiment of potential investors, Kioxia abandoned its IPO plans. Therefore, this forecast may be aimed at bolstering investor confidence in 3D NAND flash memory (especially Kioxia’s products) and re-emphasizing the plan to start production at the second Kitakami plant. Earlier this year, Kioxia resumed full production at its Yokkaichi and Kitakami plants. Due to weak demand for 3D NAND flash memory used in smartphones, Kioxia had previously reduced production by more than 30% starting from October 2022.

With the decline in flash memory inventory and the recovery of the smartphone and PC markets (Japan was the most evident regional market for global PC recovery in Q3), the demand for 3D NAND also began to recover in the second half of 2023. The demand for storage chips in the terminal equipment market began to stabilize, while the demand for data centers surged with the AI boom. As the global market’s demand for AI servers and data center-grade storage devices continues to grow, Kioxia not only has the production capacity for 3D NAND but also offers enterprise-level SSDs, including flash memory controllers and firmware, to meet the robustly growing SSD demand. Other factors that may drive 3D NAND demand include AI experiences on devices, which also require high-capacity, high-performance local storage.

The Japanese government has provided Kioxia and its partner WD (Western Digital) with subsidies of up to $1.64 billion to expand production capacity by expanding the Yokkaichi and Kitakami plants—since Kioxia is considered a key player in the global 3D NAND market by the Japanese government and aligns with Japan’s strategic goals to revitalize the semiconductor industry.

NAND Chip Prices Plunge

In October 2024, in the memory semiconductor market, DRAM prices remained stable, while NAND flash memory prices experienced a significant drop close to 30%.

Analysts believe that the decline, primarily in commodity products, is due to weak demand in the PC and mobile sectors.

Data shows that the average fixed transaction price for storage cards and USB commodity NAND flash memory products in October was $3.07, a decrease of 29.18% month-on-month.

NAND flash memory prices had been on an upward trend for five consecutive months since last October, followed by six months of stability, before turning downward in September.

The sluggish demand for TLC NAND Flash chips has led to price drops for SLC and MLC NAND, with the expected decline narrowing in November after significant price drops in September and October.

The average fixed transaction price for PC DRAM commodity products in October was $1.7, unchanged from the previous month.

DRAM prices have been on an upward trend since October of last year, with stability in May to July, a turn downward in August, and a sharp drop of 17.07% in September. Since then, prices in October have remained stable with no changes.

Due to reduced purchasing volumes by buyers and the confirmation of fourth-quarter contract prices, PC DRAM has shown a stable trend.

Retail Investors Trapped by Samsung Electronics

In October, the Korea Exchange reported significant differences in the behavior of retail and foreign investors in the South Korean stock market for the entire month of October.

Retail investors favored Samsung Electronics, purchasing stocks worth 4.2 trillion won (approximately $3.09 billion), while foreign capital showed a strong preference for SK Hynix, purchasing 745 billion won.

On October 2nd, Samsung Electronics’ stock price was 61,300 Korean won, and SK Hynix’s stock price was 169,100 Korean won.

However, by October 28th, Samsung Electronics’ stock price plummeted over 9% to a low of 55,700 Korean won during the trading day, while SK Hynix’s stock price soared, increasing by 15.9% to 196,000 Korean won.

The stark contrast in stock performance reflects different investor sentiments and market dynamics.

From the beginning of the year to today, Samsung Electronics’ stock price has fallen by about 26%, while SK Hynix’s stock price has risen by 30%.

In the first six trading days of October, retail buying was concentrated on Samsung Electronics, with a net purchase amount of 2.15 trillion won. This trend continued throughout the month, with the purchase amount reaching 4.27 trillion won by October 31st.

On the other hand, due to SK Hynix’s outstanding performance, foreign capital has shown a preference for the company.

The rise in SK Hynix’s stock price is attributed to strong third-quarter earnings and a positive future outlook.

Researchers point out that SK Hynix raised its profits for 2024 and 2025 by 2.2% and 4.5%, respectively, in the third quarter, while most other companies, including Samsung Electronics, saw no changes.

Samsung Electronics announced an operating profit of 3.86 trillion won for its semiconductor division in the third quarter, which was below the expected 4 trillion won.

Poor performance led to foreign capital continuously selling off Samsung Electronics’ stocks throughout October.

In contrast, SK Hynix’s strong performance and dominance in the HBM market have bolstered investor confidence.

It is predicted that SK Hynix will continue to solidify its dominant position in the HBM market during periods of slowing demand.

Samsung Electronics’ Storage Profit Margin at Approximately 22%

Comparing the third-quarter performance of Samsung Electronics and SK Hynix, it is evident that there is a significant gap in profit margins for memory business between the two companies.

SK Hynix’s operating profit is around 40%, while Samsung Electronics, although not disclosed, is estimated to be around 22%.

This means that SK Hynix’s leading position is very solid, and its management efficiency is much higher.

The competition between the two around HBM will become increasingly fierce.

Institutions estimate that Samsung Electronics’ storage division’s third-quarter profit was 5.3 trillion won, with a profit margin of around 22%, a decrease of 7% from the second quarter’s 29%.

Samsung Electronics was unable to supply NVIDIA and saw a significant decline in profitability as shipments in the Chinese market continued to increase.

SK Hynix benefited from the high demand for HBM, with third-quarter profits of 7 trillion won and a profit margin of 40%.

In the past, it was common for Samsung to have a profit margin of over 40%, while SK Hynix’s profit margin was around 20%; now, the opposite has occurred.

Samsung Electronics’ overall profit for the DS division in the third quarter was 3.86 trillion won, with a profit margin of 13%.

It is estimated that the profit for other divisions is approximately 5.3 trillion won, with non-storage losses of 1.5 trillion won.

Clearly, Samsung Electronics’ non-storage divisions are suffering significant losses, and the outlook is not optimistic.

Samsung plans to invest a cumulative total of 47.9 trillion won in its semiconductor business by the end of this year.

The storage division will focus on facility investments, shifting towards high-value-added products such as HBM and DDR5, as HBM prices are at least four times that of commodity DRAM products, enabling higher profits.

SK Hynix will have a cumulative investment of 10 trillion won by the end of this year and is expected to have 10 trillion won in equipment investment next year.

Next year’s investment direction will be to transform products, reduce the output of commodity products, and increase sales of DDR5 and LPDDR5.

Equipment investment will increase, but slightly decrease compared to this year.

Samsung Electronics Aggressively Chasing SK Hynix

In order to catch up with SK Hynix in HBM, Samsung Electronics is taking aggressive actions in hopes of regaining the industry’s top position.

It is expected that personnel and organizational adjustments will be implemented as early as this month, leading to significant changes within the company.

Samsung Electronics is focusing on high-profit storage products like HBM and has announced a competitive strategy for its current contract manufacturing.

As an integrated semiconductor company specializing in design, storage, and wafer foundry, there were previous concerns about technology leaks and customer competition, making it difficult to gain the trust of more customers.

Now, Samsung is preparing to abandon its previously held strategy by offering to cooperate with other wafer foundry companies to attract customers to use their HBM, thereby quickly improving storage performance.

On November 1st, at the 55th-anniversary ceremony of Samsung Electronics, the storage division leader stated that without change, there can be no innovation or growth.

The upcoming significant organizational adjustments and personnel changes are highly anticipated.

NAND flash memory chips Market and investment analysis

The plummeting prices of NAND flash memory chips are a complex phenomenon influenced by multiple factors. Here are some potential reasons and impacts:

1. Low Demand**: As you mentioned, low demand in the PC and mobile sectors is one of the main reasons for the decline in NAND flash memory prices. When there is a reduction in new product releases in these areas or a decrease in consumer purchasing power, the demand for storage chips also decreases.

2. Oversupply**: If the supply of NAND chips exceeds the demand in the market, it can lead to a price drop. This could be due to overestimation of market demand in the past or technological advancements that increase production efficiency, thus increasing supply.

3. Technological Advancements**: As technology evolves, producing higher-performance and larger-capacity NAND chips becomes more cost-effective. This can lead to a decrease in the prices of older technology chips as they are replaced by newer technologies.

4. Market Competition**: With more manufacturers entering the market, intensified competition can lead to price wars, which in turn can drive down prices.

5. Economic Environment**: Changes in the global economic environment, such as inflation and trade policies, can also affect the prices of NAND chips.

6. Inventory Adjustments**: Companies may adjust their inventories based on market conditions, which can lead to short-term supply and demand imbalances, affecting prices.

7. Product Lifecycle**: As new products are introduced, older products may be discounted to clear inventory, making way for new products.

The significant drop in the average fixed transaction price of storage cards and USB general NAND flash memory products in October, as you mentioned, could be the result of a combination of these factors.

A substantial decrease in prices can impact the profit margins of manufacturers and may also stimulate demand by offering lower purchase costs for consumers. However, prolonged low prices can affect the healthy development of the industry, causing some manufacturers to exit the market or seek new growth opportunities.

For investors and businesses, it is crucial to closely monitor market dynamics and adjust strategies to adapt to market changes. For consumers, this may be a good time to purchase storage devices, but it is also necessary to consider the risks associated with price fluctuations