本报讯 （记者 俞圣彤） 2月24日，江阴市举行新型工业化暨"人工智能+"行动推进大会，全面掀起"项 目建设攻坚年"奋战热潮，聚力打造产业强市，加快建设具有全国影响力的先进制造业科创中心。现 场，江阴市人工智能赋能中心揭牌。 江阴将通过"四维发力"构筑工业升级新引擎：增存量潜力，江阴支持龙头企业裂变扩张、上市公司并购 重组，推动外资利润再投资，确保200个以上重点增资扩产项目在2026年内完成投资超200亿元；强优势 产业领域，聚焦特钢新材料、高端纺织等四大特色产业，通过"链主企业+产业园区"模式推动集群化发 展；补关键技术环节，依托长电科技（600584）、盛合晶微等企业突破HBM内存、Chiplet异构集成等 先进AI芯片封装技术，形成"高校研发—园区转化—企业落地"创新闭环；育未来产业板块，瞄准具身机 器人、垂类大模型等赛道，以"长期主义"培育新质生产力。 江阴还将从核心赛道到全域场景构建立体攻坚体系。巩固AI芯片封装领域优势，加快布局AI产业，力 争实现全年产业规模增速超20%，产业规模突破15亿元，同时推进算力中心建设，建设AI赋能中心，让 人工智能更好地赋能实体经济、赋能制造业。此外，聚焦科技创新 ...

Core Insights - The semiconductor industry is projected to reach a record sales figure of $975 billion by 2026, driven primarily by the growth of artificial intelligence infrastructure [2] - The industry faces a paradox where strong demand from AI is pushing revenues to unprecedented heights, but there are significant risks associated with over-reliance on AI [1][5] - By 2026, AI chips are expected to account for nearly 50% of total industry revenue, yet their production volume remains low, highlighting a structural disparity [5][6] Market Conditions - The semiconductor industry's growth rate is expected to accelerate from 22% in 2025 to 26% in 2026, with long-term projections indicating sales could reach $2 trillion by 2036 [5] - The total market capitalization of the top ten semiconductor companies reached $9.5 trillion by December 2025, a 46% increase from the previous year [5] - The revenue from generative AI chips is forecasted to approach $500 billion by 2026, representing about half of global chip sales [5] Supply Chain Dynamics - The average selling price of chips is projected to be $0.74, with total chip sales expected to reach 1.05 trillion units by 2025 [6] - Memory revenue is anticipated to reach approximately $200 billion in 2026, constituting 25% of total semiconductor revenue [6] - The semiconductor industry is experiencing a supply-demand imbalance, particularly in memory products, leading to significant price increases [7][25] Strategic Considerations - The industry must address potential declines in AI chip demand post-2026 while maintaining high cash levels and low debt [13] - There is a need for strategic partnerships and investments to build ecosystems around semiconductor manufacturing and AI chip platforms [20][22] - The rise of vertical integration among semiconductor and AI infrastructure providers indicates a shift in capital allocation strategies [20][29] Future Outlook - The semiconductor industry is expected to face capacity constraints in 2026, impacting the production of advanced logic processes and memory chips [23][24] - Geopolitical factors and material supply limitations may disrupt procurement and manufacturing processes [19][26] - The transition of AI workloads from training to inference may challenge existing market leaders in AI GPU, CPU, and memory sectors [28]

Core Insights - The key bottleneck in AI development has shifted from computing power to memory, as highlighted by Intel's CEO Chen Lifeng [3][6][7] - The training and inference of large models require the exchange of tens of terabytes of data per second between GPUs and memory, likened to a high-performance car constrained by a narrow fuel line [3][6] - Only three companies globally can mass-produce the high-bandwidth memory (HBM) necessary for advanced AI, with their production capacity booked until 2028, leading to a significant imbalance between computing power and memory expansion [3][6] Short-term Implications - The current situation results in low utilization of computing power and increased costs for model training and inference [3][6] Mid-term Challenges - The semiconductor manufacturing capacity and the limits of the semiconductor industry will be tested [3][6] Long-term Considerations - There will be challenges related to power, heat dissipation, materials, and infrastructure support capabilities [3][6] - The competition in AI has evolved beyond algorithms to a national-level endurance contest encompassing computing power, memory manufacturing, energy, and infrastructure [3][6] Future Outlook - The future potential of AI will not be determined by sporadic model breakthroughs but by the robust industrial foundation that supports the stable operation of complex systems [3][6][7] - The ultimate form of AI is envisioned as a stable, operating intelligent industrial machine rather than merely a singular smart model [3][6][7]

Core Insights - The storage chip market has experienced a significant price surge since the second half of last year, with consumer-grade memory prices increasing by over 600%, making it a highly valuable investment product [1] - The demand for AI infrastructure has led to a depletion of production capacity among major storage chip manufacturers, including Samsung, Micron, and SK Hynix, with all their HBM production capacity booked until 2026 [1][11] - The rising prices of storage chips have disproportionately affected low-end smartphones, particularly brands like Transsion, which saw a 22.6% revenue increase but an 11% drop in net profit in Q3 last year [3][4] Market Dynamics - The high profit margins associated with HBM memory used in AI chips have prioritized its production over consumer-grade products, leading to supply constraints for low-end smartphones [3] - Various smartphone manufacturers are shifting focus from low-priced product lines to higher-margin high-end models due to the impact of rising storage costs [4][12] - Transsion's market share dropped from fourth to "others" category, indicating a significant decline in its competitive position [3][4] Pricing and Cost Structure - The cost structure of low-end smartphones is heavily impacted by the rising prices of storage chips, which are considered a rigid cost that cannot be easily adjusted [5][8] - The BOM (Bill of Materials) cost for a 12GB+256GB iPhone Pro Max indicates that storage costs account for approximately 10% of total costs, highlighting the sensitivity of low-end models to price increases [8][9] - The average selling price (ASP) of smartphones is projected to exceed $400 for the first time, reflecting a shift towards higher-priced models as low-end market segments shrink [13][16] Supply Chain Challenges - The demand for AI-related products has created a seller's market for storage chips, leading to unpredictable cost structures for smartphone manufacturers, especially in the low-end segment [12] - Major companies like Apple are facing significant supply pressures, with only partial agreements in place for NAND Flash and DRAM supplies for the upcoming quarters [11] - Samsung's internal conflicts regarding storage pricing have further complicated supply chain management, as the mobile division struggles to secure favorable terms from the semiconductor division [12]

Core Viewpoint - Intel's stock price increased by 2.43% on February 11, closing at $48.28, driven by advancements in AI memory technology and positive market sentiment [1] Group 1: Stock Price Movement - Intel's stock price rose due to the unveiling of Z-Angle Memory (ZAM) technology, which features a vertical stacking architecture with a maximum single-chip capacity of 512GB and a power consumption reduction of 40%-50% [2] - The semiconductor sector saw a 2.19% increase on February 11, benefiting Intel as funds shifted towards companies with expected technological breakthroughs [3] Group 2: Market Activity - Following a 6% drop on February 10, Intel's stock showed signs of technical recovery, with a trading volume of $4.67 billion and a turnover rate of 1.95% on February 11, indicating high market activity [4] - The stock experienced a daily fluctuation of 5.71%, with a low of $46.86 and a high of $49.55, reflecting intense buying and selling pressure [4] Group 3: Company Fundamentals - Citigroup's report indicates that Intel's capital expenditure is expected to stabilize between $15 billion and $16 billion by 2026, with improvements in the foundry customer pipeline [5] - The enhancement of Intel's 18A process capacity and progress in collaborations for the 14A process have alleviated market concerns regarding process technology delays [5]

Core Viewpoint - The storage chip market is experiencing significant price increases driven by AI demand, but there are warnings from companies like SMIC about potential overestimation of demand and supply chain disruptions in traditional sectors [1][3][9] Group 1: Market Dynamics - DDR5 memory prices surged nearly 70% in a month, with SSD prices in Shenzhen doubling, leading to unusual trading practices like "buying memory with motherboards" [1] - Major semiconductor companies, including Samsung and SK Hynix, are projected to achieve record revenues in their storage businesses by 2025, while domestic module manufacturer Jiangbolong reported a 1994% increase in net profit [3] - SMIC's capacity utilization reached 95.8%, forcing delays in mobile phone orders to prioritize AI chip deliveries [3] Group 2: Supply Chain Issues - The price surge is attributed to unexpected capacity cuts by manufacturers like Micron in the consumer storage segment and structural shortages in HBM memory due to AI demand [3][5] - There is a panic-driven cycle of stockpiling within the industry, with DRAM modules being allocated through a "distribution system," marking a first in the industry [3][5] - Smartphone manufacturers like Xiaomi and OPPO are facing DRAM inventory shortages of less than three weeks, leading to price increases of 100-400 yuan for new models [3][5] Group 3: Future Outlook - The current situation is expected to lead to a reduction in orders for PMICs and display drivers, which are key products for SMIC, as smartphone manufacturers struggle with storage shortages [5] - TrendForce has revised down its 2026 smartphone shipment forecasts, indicating that the high prices are negatively impacting the end market [5] - SMIC predicts that the current high-price environment will not last, warning of potential overcapacity as demand from end-users is suppressed [7][9]

Core Viewpoint - The hardware industry is facing significant challenges in 2025 and 2026, with major delays in the release of the RTX 60 series graphics cards, as NVIDIA shifts its focus to AI products, leaving the gaming market in a state of stagnation [1][3]. Group 1: Graphics Card Market - NVIDIA has postponed the release of the RTX 60 series graphics cards to the end of 2027, prioritizing AI products over gaming hardware [1]. - The graphics card market is expected to experience a substantial update void over the next two years due to this shift [3]. - NVIDIA's introduction of DLSS 4 and 4.5 updates aims to extend the lifespan of older graphics cards, but this is seen as a temporary fix rather than a genuine hardware upgrade [4][7]. Group 2: Memory Market Dynamics - Intel CEO Lip-Bu Tan predicts that the current memory price increase cycle will last until at least 2028, with no signs of relief in sight [10]. - The demand for high-bandwidth memory (HBM) driven by AI applications is causing a significant increase in costs for consumer electronics, leading to a "stagflation" scenario where prices rise while performance declines [12][14]. - Major memory manufacturers are reallocating resources to HBM production, which is squeezing the supply of consumer-grade DDR4/5 and LPDDR memory, further escalating costs for manufacturers [14][16]. Group 3: Semiconductor Manufacturing Challenges - The semiconductor industry is facing limitations in process technology, with TSMC's N2 process showing minimal performance improvements compared to the more established 3nm process, leading to higher costs for marginal gains [17]. - Manufacturers are likely to adopt conservative strategies, focusing on refining existing 3nm products rather than pursuing new advancements, resulting in a stagnation of technological progress [19][21]. - The industry is expected to see more products that are merely name upgrades with downgraded specifications, reflecting a broader trend of stagnation in consumer electronics [21][23]. Group 4: Consumer Electronics Outlook - The consumer electronics market is entering a phase of technological stagnation, where new products may not represent significant performance improvements, but rather increased prices for minimal enhancements [23][25]. - Consumers are advised to adjust their expectations regarding technological advancements, recognizing that the current environment is characterized by rising costs and stagnant performance [28][31].

Core Insights - Memory prices have increased by 80% to 90% in 2026 compared to Q4 2025, with DRAM, NAND, and HBM prices reaching record highs [1] - OEMs are adjusting their memory configurations and prioritizing high-end products with LPDDR5 to manage cost pressures [1] - DRAM operating profit margins reached approximately 60% in Q4 2025, surpassing HBM for the first time, with expectations for further increases in Q1 2026 [2] Group 1: Market Trends - The memory shortage is more severe than initially anticipated, driven by demand from AI-driven data centers and cloud service providers, leading to expected price increases of 90% to 95% for DRAM and 55% to 60% for NAND in Q1 2026 [3][4] - OEMs typically bulk purchase memory a year in advance, which has led to discrepancies in pricing between pre-installed systems and independent memory kits [3] - The demand for LPDDR memory is expected to rise significantly, with prices projected to increase by approximately 90%, marking the largest increase ever [4] Group 2: AI Impact - The transition from training to inference in AI infrastructure requires more DRAM and storage space, particularly for storing key-value caches during model inference [5] - The need for high-performance storage has surged due to the growth of AI inference applications, leading to increased procurement by major communication service providers [4][5] - Despite expectations for DRAM prices to peak later this year, a return to normal levels is projected to take several years, with high prices expected to persist until 2028 [5]

Core Insights - The recent surge in memory prices has stabilized, with major players like Samsung, SK Hynix, and Micron initiating a "real demand survey" to assess the market and signal increased production [1][10][15] - The price of memory products, such as DDR5 RAM, has shown a downward trend, with prices dropping from a peak of 2000 yuan to 1699 yuan [1][3] - The memory price increase has been attributed to panic buying and supply chain disruptions rather than genuine demand growth, leading to a disconnect between consumer demand and price hikes [4][9] Market Dynamics - The price of DDR4 memory increased from 138 yuan to 899 yuan within a year, representing a more than sixfold increase, while DDR5 prices have also surged significantly [4] - The price hikes have affected graphics card prices, with some models seeing a 20% increase due to rising memory costs [4] - The phenomenon of "bullwhip effect" is evident, where small fluctuations in consumer demand lead to exaggerated responses in the supply chain, causing significant price volatility [8] Company Responses - Samsung, SK Hynix, and Micron's joint initiative to investigate real demand is a rare move aimed at cooling the overheated market and addressing the role of distributors in inflating prices [10][15] - Despite record profits reported by these companies, the disparity between their revenue growth and the price increases in the market raises concerns about sustainability [10][11] Future Outlook - The memory market is expected to see a gradual price decline, but significant drops may take 3-6 months due to production ramp-up times [16] - The ongoing AI market dynamics and potential investment disputes may further influence memory demand and pricing strategies [16][17] - The current price levels are unlikely to return to pre-2023 lows, as new demand from AI applications is expected to support higher price floors [17]

Core Insights - The semiconductor foundry and packaging sectors have shown strong performance, with software stocks generally outperforming hardware stocks [2][19] - The ASP (Average Selling Price) for mature processes is expected to improve due to strong downstream replenishment demand, with forecasts indicating a utilization rate (UTR) of over 75% for Taiwanese manufacturers [19][22] Semiconductor Performance - Foundry/packaging stocks increased by 2.1% yesterday, with a 5.2% rise over the past five days and a 124.4% increase over the last 180 days [3][20] - Semiconductor equipment stocks rose by 1.2% yesterday, with a 148.3% increase over the last 180 days [3][20] - Software large-cap stocks increased by 1.1% yesterday, but have seen a decline of 11.1% over the last 60 days [3][20] NAND Market Outlook - Bernstein predicts a significant upward cycle for NAND driven by AI demand and limited supply growth, with SanDisk expected to have substantial upside potential [22] - For FQ2'26, SanDisk's EPS is projected at $3.79, exceeding street expectations of $3.45, with a more optimistic FQ3'26 EPS forecast of $6.52 based on a 22% QoQ ASP increase [22] - JP Morgan anticipates a substantial increase in NAND TAM growth from a long-term average of 10% to over 30% in the next three years, primarily driven by AI-related eSSD demand [23] Taiwanese Foundries - Citigroup reports that VIS and UMC have outperformed the market by 48% and 22% respectively year-to-date, with UTR expected to rise to 75%+ this year [19][20] - UMC's 28/22nm utilization rate remains above 90%, with overall UTR improving from 60-70% to 75-80%, leading to upward adjustments in profitability forecasts for 2026/27 [19][20] Key Company Earnings Projections - SanDisk's base case for F3Q (March) ASP is projected to increase by 22% QoQ, corresponding to an EPS of $6.52, significantly higher than the expected $4.62 [19][22] - In a bullish scenario, if F3Q ASP increases by 40%, EPS could reach $9.06, with FY27 EPS potentially at $67.5 [19][22]