memory

Market research firm Counterpoint released data indicating that in the second quarter of 2026, the NAND flash memory market by shipment volume saw Samsung, SK Hynix, and Yangtze Memory Technologies holding shares of 25%, 22%, and 14% respectively, marking the first time Yangtze Memory Technologies has entered the top three.However, based on revenue, Yangtze Memory Technologies ranks fifth in the industry, trailing behind Micron and Kioxia. The reason is that its product mix is concentrated on consumer applications, with a relatively low proportion of high-priced enterprise-grade solid-state drives for data centers.

According to the "Science and Technology Innovation Board Daily," as the original factory prices further rise in the third quarter, the price of server DDR5 has increased by 15% to 23% this month, with 32GB / 64GB / 96GB DDR5 RDIMM rising to 900/1590/2650 USD respectively. In the spot market for server RDIMM, the trade price for 32GB D5 is between 1800~2100 USD, 64GB D5 is between 3000~3200 USD, and 96GB D5 is between 3800~4100 USD. The price difference between contract and spot DDR5, especially for 96GB, shows a trend of narrowing. The supply of original factory DDR5 remains insufficient to meet demand; however, as contract prices rise seasonally, large server clients are becoming more sensitive to the increase in memory costs. Additionally, with long-term agreements being established, the supply of D5 for large clients is somewhat guaranteed, and the increase in DDR5 prices is expected to be more moderate.

According to the latest research by TrendForce, driven by rising prices of components such as memory, the material cost of the 256GB iPhone 18 Pro is expected to increase by about 38% year-on-year, making a retail price increase unavoidable. Apple may ease the extent of the price increase by sacrificing some gross margin to maintain shipment volume and market share. Analysis shows that the proportion of memory in the BOM cost of the last two generations of Pro models has risen from about 10% a year ago to about 34% in Q3 2026, and it is expected to exceed 40% in the first half of 2027. Memory has become the largest single cost item for the iPhone, replacing the traditional dominance of application processors and displays.TrendForce points out that Android manufacturers are facing greater profit pressure, especially mid-range and low-end models, with some brands potentially forced to significantly raise prices or discontinue loss-making product lines. Since early 2025, memory prices have increased by 5 to 7 times, and it is expected that global smartphone production will continue to be under pressure from the second half of 2026 to 2027. Apple may also adjust the pricing of older models in sync with the launch of new products to partially offset the impact of rising memory costs.

According to UBS's analysis of the Vera Rubin superchip Bill of Materials (BOM), the share of memory semiconductors in the total cost of the platform has risen to about 62%, a significant increase from the 53% of the previous generation Blackwell Ultra (GB300). The total cost of the Vera Rubin superchip (which includes 1 Vera CPU and 2 Rubin GPUs) is approximately $38,902, with memory-related costs reaching $24,297. The cost of HBM4 is about $4,943 (accounting for 12.7% of the total cost), while the cost of the SOCAMM2 connecting the CPU is about $19,355 (accounting for 49.8% of the total cost), making it the core driver of rising memory costs.SOCAMM2 is a server memory module based on low-power DRAM LPDDR5X, supporting a maximum capacity of 1.5TB per CPU. The Vera Rubin NVL72 cabinet (which includes 72 Rubin GPUs and 36 Vera CPUs) is equipped with a total of 74.7TB DRAM, of which HBM4 accounts for 20.7TB, and LPDDR5X for the CPUs reaches 54TB, with the LPDDR usage in a single cabinet equivalent to about 4,500 high-end smartphones. However, due to uncertainties in the supply of LPDDR5X and HBM4E, NVIDIA is considering reducing the HBM configuration for the next generation Rubin Ultra.

SK Hynix stated in a recent article that the competition in AI is shifting from the performance of individual chips to the design and operation of the entire infrastructure architecture. The competitiveness of AI data centers no longer depends on individual components, but on whether the five key elements of computing, memory, storage, networking, and power cooling can be seamlessly integrated.The article points out that the continuous expansion of AI model scales has led to a surge in demand for computing power and data movement. Training requires repeatedly reading massive datasets, while inference relies on quickly retrieving user request information, both of which place higher demands on the system architecture of data centers. At the memory level, HBM, server DRAM, and others have formed a hierarchical system, each undertaking different bandwidth and capacity tasks. At the networking level, as large-scale training and inference rely on multi-server parallel processing, networking has become a key factor determining the scalability of data centers. System design is shifting from single-server to whole rack and cluster-level expansion.According to Omdia's forecast, the AI data center chip market will grow from $123 billion in 2024 to $207 billion in 2025, reaching $286 billion by 2030. SK Hynix also mentioned that Microsoft's Fairwater data center in Wisconsin is about the length of five football fields, indicating that infrastructure is being deployed on a larger scale. SK Hynix emphasizes that memory is becoming a key layer connecting computing and data.

According to The Wall Street Journal, Apple is testing memory chips from the Chinese DRAM manufacturer Changxin Storage (CXMT), planning to apply them to product lines such as the iPhone and MacBook. The background of this move is the AI demand driving up global memory prices; Apple has previously lobbied the U.S. side, stating that the current pricing from Korean and American suppliers is "unsustainable." The report states that Apple is seeking approval from the White House to conduct business with Chinese chip suppliers. Currently, the U.S. allows the purchase of commercial off-the-shelf chips but restricts customized products.Changxin Storage has recently rapidly expanded its market share, with second-quarter revenue growing more than 8 times year-on-year, and its global DRAM market share rising to 7%, completing its IPO on the Science and Technology Innovation Board last month. Meanwhile, PC manufacturers such as HP and Acer have begun using Changxin Storage memory chips in some devices sold in non-U.S. markets to alleviate supply pressure. If Apple ultimately reaches a procurement agreement, it may face political resistance from Washington.

Micron Technology stated to Deutsche Bank at the FMS 2026 conference that AI is reshaping the memory industry landscape, with the proportion of memory in the total value of systems rising from about 10% thirty years ago to nearly 50% currently, and this trend is accelerating. Currently, both DRAM and NAND are in a state of supply shortage. Due to the need for physical disassembly of systems for memory upgrades, the demand for memory in AI systems is more rigid, and price elasticity is lower than market expectations. The Strategic Customer Agreement (SCA) is expected to cover about 40% of sales, providing a price stability mechanism for both parties.Micron's management described the demand for CPU-side AI Agents as "pre-season warm-up," viewing it as a new pillar of growth outside the GPU ecosystem. For emerging technologies like SRAM, Micron believes their scalability is limited and cannot challenge the core position of HBM, with only about 5% of systems currently running such specific workloads. Deutsche Bank maintains a "Buy" rating, believing Micron has a unique advantage of "growth without sacrificing profits." During FMS, Sandisk and SK Hynix jointly released the first HBF technical specifications, but Micron believes that its actual value and commercialization prospects remain controversial.

According to The Information, Nvidia is weighing a significant adjustment measure, considering using less high-bandwidth memory for the Rubin Ultra GPU to address the shortage of advanced high-bandwidth memory chips.

Ahmed Shihab, Chief Product Officer of Western Digital, stated that the key to AI storage competition is not simply pursuing the fastest medium, but whether it can continuously scale capacity at an affordable cost. He believes that many architectures run well in the early stages, but when the data scale grows from several PB to hundreds of PB or even EB levels, issues may arise due to uncontrolled costs.Flash storage is suitable for high-performance, low-latency scenarios such as model weights, GPU overflow, KV caches, and session contexts; HDDs are more suitable for large-scale, long-term storage, and cost-sensitive data such as training corpora, logs, checkpoints, compliance records, synthetic data, and inference outputs. He summarized: "Flash handles the present, HDD handles the entire lifecycle."At the scale of AI, storage costs themselves will become an architectural issue. Long-term storage of large volumes of data on high-performance media will crowd out budgets for computing, networking, power, and personnel. For many bulk storage workloads, the issue is not whether flash can store, but whether customers can afford the total cost of using flash in the long term.In the future, AI storage will not be dominated by a single technology, but should be designed in layers based on performance, cost, power consumption, density, reliability, and data lifecycle. He emphasized that this is not a competition between flash and HDD, but rather a need to match the appropriate medium for different workloads from the very beginning; otherwise, the architecture may impact business sustainability as it scales.

According to official news, Marvell Technology announced the launch of a new generation of memory solution portfolio for AI infrastructure, covering server-level AI storage, rack-level CXL memory expansion and pooling, as well as multi-rack optical interconnect shared memory, aimed at addressing the growing memory capacity and bandwidth bottlenecks in Agentic AI inference processes.Marvell stated that as AI model sizes increase, context windows extend, and KV Cache demand grows, traditional tightly coupled architectures of computing and memory are limiting AI inference efficiency. Through memory disaggregation, memory resources can be made more independent of computing resource expansion, improving GPU utilization and reducing data movement latency. The products released include:Bravera SC6 PCIe 6.0 SSD controller: Designed for AI inference storage scenarios, it helps cloud service providers migrate more KV Cache to high-performance SSDs, enhancing infrastructure efficiency. This product features an architecture compatible with multi-vendor NAND and is expected to begin sampling in the fourth quarter of 2026.Marvell Structera X memory expansion solution: Based on CXL technology, it supports rack-level memory expansion and resource pooling, helping data centers share and allocate memory resources more flexibly, reducing AI infrastructure costs.Marvell Photonic Fabric optical interconnect memory solution: Constructs a shared memory architecture across multiple racks using optical interconnect technology, supporting up to 32TB warm KV Cache offloading and helping AI inference clusters enhance throughput capacity.Marvell stated that the Photonic Fabric solution can achieve a 2 to 3 times increase in Token throughput under existing data center space and power consumption constraints, supporting larger scale models and longer context AI applications.Marvell executive Will Chu stated that AI infrastructure is transitioning from a single server architecture to a system where computing, memory, and connectivity operate in synergy, and in the future, memory needs to expand more independently to enhance resource utilization and Token efficiency.As the demand for AI Agents and large model inference continues to grow, memory capacity, bandwidth, and data transfer efficiency are becoming new focal points of competition in AI infrastructure, following computing power.

According to Counterpoint Research data, in the second quarter of 2026, the global DRAM market saw Samsung Electronics regain the top position with a 39% market share, significantly rebounding compared to the same period last year; SK Hynix's share fell from 39% in the same period last year to 26%, despite a 214% year-on-year revenue growth; Micron followed closely with a 25% share, only 1 percentage point behind SK Hynix. Counterpoint pointed out that Samsung achieved a rebound driven by the growth in general DRAM demand and the expansion of its share in the HBM sector.General DRAM prices rose significantly month-on-month, while the average price of HBM declined year-on-year due to the drop in HBM3E prices and delays in HBM4 shipments. Changxin Storage saw a 716% year-on-year revenue growth, benefiting from strong domestic general DRAM demand and capacity expansion; Nanya Technology grew by 690% year-on-year, benefiting from the expansion of DDR/LPDDR4/5 supply. Counterpoint predicts that with the shipments of NVIDIA Vera Rubin and AMD Instinct MI455X, HBM4 shipments will significantly increase in the second half of the year, and the improvement in product mix is expected to support a rebound in HBM prices.

According to Reuters, informed sources say that Changxin Storage is considering building a second 12-inch DRAM wafer fab in Beijing's Yizhuang and is negotiating financing with the Beijing Economic and Technological Development Zone and several state-owned technology companies. The company is seeking at least 60 million yuan (approximately 8.9 million USD) in support, but negotiations are still in the early stages, and the scale and structure of the financing may be adjusted.The proposed factory will be located at the site of Changxin Storage's existing DRAM wafer fab in Beijing. The planned production capacity and total investment for the project have not yet been determined, while building a fab capable of producing advanced DRAM typically requires over 10 billion USD. Currently, Changxin Storage operates two 12-inch DRAM wafer fabs in Hefei and one in Beijing, each with a monthly production capacity of about 100,000 wafers. Changxin Storage is also building new factories in Shanghai and Hefei, and once the related projects are fully operational, the company's monthly production capacity could double to over 600,000 wafers.This expansion comes as demand for storage chips is driven by AI infrastructure, data centers, and consumer electronics entering an upward cycle. The company completed an 8.6 billion USD IPO last month, marking the largest fundraising for a semiconductor company listed in mainland China, and its stock price has risen by 13% since going public. Changxin Storage is currently the fourth largest DRAM manufacturer in the world, but Samsung Electronics, SK Hynix, and Micron together still account for nearly 90% of the global market share in the first quarter. Beijing and Shanghai are also providing funding and other support to Changxin Storage to gain economic and strategic benefits from its expansion.

Goldman Sachs published a research report stating that Samsung Electronics' operating profit for the second quarter met expectations. The firm raised its earnings per share forecast for 2026 to 2028 by about 2% and increased the 12-month target price from 480,000 won to 490,000 won. The firm expects that demand for memory will significantly exceed supply, with the gap widening in 2027 and supply remaining tight in 2028.Goldman Sachs pointed out that the company has signed long-term agreements with major data center customers, which will provide better profit visibility and downside protection.In addition, the company aims to become a leader in the AI memory market, with HBM4 expected to account for over 60% of HBM revenue in the second half of 2026, and anticipates significant upside potential for shareholder returns. The stock is currently trading at a forecasted price-to-earnings ratio of 3.1 times and a price-to-book ratio of 1.2 times, with a shareholder return rate close to 50%.

Roundhill Memory ETF (DRAM) has included CXMT (Changxin Technology) in its holdings, with a weight of 2.52%.The DRAM ETF focuses on memory chip companies. As of August 2, the top three holdings of this ETF are Samsung Electronics, Micron Technology, and SK Hynix, with weights of 26.39%, 24.54%, and 22.77%, respectively. Other major holdings include Seagate Technology, Western Digital, SanDisk, Kioxia, Nanya Technology, and Zhaoyi Innovation.

According to Jinshi reports, Apple CEO Tim Cook stated that the company is evaluating all options regarding memory supply to find alternative component suppliers.

According to Jinshi reports, Apple CEO Tim Cook stated that spending on memory in the third fiscal quarter was "significantly higher" than in the second fiscal quarter.

According to Gate market data, U.S. stocks closed higher on Thursday, with the Dow Jones Industrial Average initially up 1.19%, the S&P 500 index up 1.67%, and the Nasdaq up 2.78%. Micron Technology rose 18%, SanDisk rose 26%, SK Hynix rose 17%, and Nvidia rose over 2%.The Nasdaq Golden Dragon China Index rose 1.05%, and iQIYI rose 4%.