台积电，持续领跑先进制程 台积电先进制程的进展，放眼市场，持续上演「一个人的武林」。台积电先前已表示，2nm制程今年下半年量产，预期量 产曲线将与N3制程相似。台积电最先进的A14制程技术开发按计划进行且进展良好，预计2028年量产，且后续针对A14制 程持续强化，包括2029年推出采用超级电轨的方案。 公众号记得加星标⭐️，第一时间看推送不会错过。 来源：内容来自综合 。 据报道，美国财政部长贝森特警告，台积电斥资400亿美元在亚利桑那州兴建的大型晶圆厂，可能只能生产7%美国所需的 半导体，此问题突显台积电面临在地监管与繁文缛节的阻碍。 报道称，贝森特在《All In Podcast》节目中说，「台积电希望在亚利桑那州兴建一座巨型晶圆厂系统，但我认为，它也 许能生产最多7%美国所需的芯片」，他指出，「他们正应付当地的建管人员」。 贝森特批评监管障碍拖慢这座400亿美元的晶圆厂区建设，他表示，「显然这些芯片厂行动太迅速，你一再变更计划，就 有某人来说，『你说管子在那里，但却没在那里，我们要你停工』」。 他补充，当他从一介公民转变成公仆后，他经历许多意想不到的事，其中之一就是「在美国，我们让建设变得非常困 难」 ...

Group 1 - The Chinese semiconductor industry is facing significant challenges due to U.S. sanctions, particularly the inability to obtain EUV equipment, leading to reliance on DUV technology for process improvements [1][3] - Recent reports indicate that domestic 5nm chips have achieved mass production with yield rates improving from 35% to 60%-70%, nearing TSMC's initial SF3 yield levels [1][3] - There are doubts regarding the feasibility of achieving high yields with 5nm technology using DUV equipment, with estimates suggesting a 50% increase in costs [1][3] Group 2 - Engineers have utilized quadruple patterning technology (SAQP) to enhance the resolution of DUV equipment, allowing for more precise chip designs despite inherent limitations [3] - Currently, there are no commercially available products utilizing the 5nm process, with the latest Kirin X90 chip still based on 7nm technology [3] - If the rumors about the 5nm process are confirmed, it would represent a significant positive development for all domestic chip manufacturers [3] Group 3 - The global semiconductor industry is closely monitoring China's technological advancements, with reports suggesting that China is testing its own EUV equipment, potentially starting production in the third quarter of this year [3][4] - Successful domestic production of EUV equipment would signify a breakthrough in China's semiconductor industry, enabling further advancements in process technology and challenging major players like Intel, Samsung, and TSMC [4]

Core Viewpoint - Intel's new CEO, Pat Gelsinger, is planning a significant strategic shift in the contract manufacturing business, focusing on advanced process technologies to secure major customer orders, contrasting sharply with the previous CEO's approach [1][2] Group 1: Strategic Adjustments - The new strategy may expose Intel to billions in asset write-down risks due to challenges in promoting the 18A process technology, which has faced market resistance [1] - Since taking over in March, the new CEO has initiated multiple cost-cutting measures, with a consensus forming by June that the heavy investment in the 18A process may not yield expected returns [1][2] - Intel is shifting its focus to the more advanced 14A process, which is seen as a potential competitor to TSMC's current offerings, and is preparing to approach major clients like Apple and Nvidia [2] Group 2: Financial Implications - Intel is projected to incur a net loss of $18.8 billion in 2024, marking a significant financial downturn for the company, which has already experienced its first annual loss since 1986 [2] - The company has committed to producing specific 18A process chips for Amazon and Microsoft, indicating that some orders will remain unaffected by the strategic changes [3] Group 3: Internal Decision-Making - The board is currently deliberating on whether to completely abandon the external sales of the 18A process, with a decision expected to be submitted for review soon [2] - The management team is focused on strengthening the roadmap, rebuilding customer trust, and improving financial performance, indicating a proactive approach to the challenges faced [2]

Core Insights - The global semiconductor foundry market is undergoing significant changes, with TSMC maintaining a dominant position while SMIC is rapidly gaining market share at the expense of Samsung Electronics [2][3][4] Market Performance - In Q1 2025, the total revenue of the top ten foundries decreased by 5.4% to $36.43 billion, with TSMC's revenue falling by 5% to $25.517 billion, yet its market share increased by 0.5 percentage points to 67.6% [3][4] - Samsung's revenue dropped significantly by 11.3% to $2.893 billion, leading to a decline in market share from 8.1% to 7.7% [3][4] - SMIC was the only company among the top three to achieve revenue growth, with Q1 2025 revenue reaching 16.301 billion yuan, a year-on-year increase of 29.4% [4][5] Market Share Dynamics - SMIC's market share rose from 5.5% to 6%, narrowing the gap with Samsung from 3.2 percentage points at the end of 2024 to 1.7 percentage points [4][5] - SMIC has surpassed GlobalFoundries and UMC to become the third-largest foundry [4][5] Factors Influencing Growth - SMIC's growth is attributed to strategic inventory buildup in response to U.S. tariffs and favorable domestic policies, which allowed it to thrive amid a market slowdown [2][4] - The decline in Samsung's market share is linked to delivery issues leading to customer loss and reduced orders, compounded by challenges in technology development and production efficiency [5][6] Competitive Landscape - TSMC's market share is significantly ahead of Samsung, with a gap of 59.9 percentage points as of Q1 2025, highlighting TSMC's stronghold in advanced process technology [7][8] - TSMC's 2nm process technology is progressing well, with a yield rate exceeding 60%, while Samsung's yield is around 40%, indicating a widening technological gap [7][8] Future Outlook - If SMIC continues its growth trajectory and optimizes its production processes, it may potentially surpass Samsung in the future, although significant technological advancements are still required [6][8] - The competition in the 2nm node is critical, with TSMC currently leading in technology and customer relationships, making it challenging for Samsung to catch up without substantial improvements [8]

Group 1 - TSMC's 2nm process yield for memory products has surpassed 90%, indicating significant manufacturing efficiency [1] - TSMC is targeting orders from major US tech companies such as Nvidia, Apple, Qualcomm, AMD, and Broadcom, with expectations of reaching full capacity utilization soon [1] - The Arizona facility is currently producing N4 chips, corresponding to 5nm and 4nm technologies, with Nvidia's AI chips undergoing process validation and expected to start production by the end of this year [1] Group 2 - TSMC's 2nm wafer production is projected to reach four times the volume of the previous year's 5nm production, reflecting strong demand [2] - Major clients like Apple, MediaTek, Qualcomm, and AMD are planning to utilize TSMC's 2nm process for their next-generation chips, which is expected to drive revenue growth [2] - Analysts predict TSMC's revenue growth of 26.3% and 14.1% for 2025 and 2026, respectively, with advanced process nodes below 7nm increasing from 68.8% in 2024 to 79.3% in 2026 [2] Group 3 - In advanced packaging, TSMC's CoWoS shipments are expected to increase from 585,000 units in 2026 to 923,000 units, with capacity rising from 660,000 to 1,000,000 units, representing year-on-year growth of 52% and 58% [3] - Despite currency fluctuations affecting profit expectations for 2025-2027, TSMC's revenue is still anticipated to grow by approximately 25% in 2025 [3]

如果您希望可以时常见面，欢迎标星收藏哦~ 苹果、联发科、高通等科技行业的知名企业已瞄准台积电的2纳米制程，据称后者已于4月1 日开始接受订单。每片晶圆3万美元的成本，对于下一代制程节点而言，这已然是难以接受的 难题，但这些公司不惜斥资数十亿美元来获取竞争优势或保持领先地位。 然而，接下来的路只会越来越难走，因为最近的一项预测显示，在2纳米制程之后，1.4纳 米"埃"制程将紧随其后，但其成本可能进一步攀升。 1.4nm，太贵了 在今年四月的北美技术研讨会上，台积电发布了其 A14（1.4 纳米级）制造技术，并承诺该技术将 在性能、功耗和晶体管密度方面显著优于其 N2（2 纳米）工艺。但据台媒中国时报报道，台积电 A14工艺每片晶圆的成本可能高达4.5万美元。与2纳米节点相比，这相当于价格上涨了50%。 那么这个晶圆贵在哪里？ 台积电透露，新节点将采用其第二代环栅 (GAA) 纳米片晶体管，并将通过 NanoFlex Pro 技术提 供更大的灵活性。台积电预计A14 将于 2028 年投入量产，但不支持背面供电。据介绍，A14 是 台积电全节点的下一代先进硅技术。从速度来看，与 N2 相比，其速度提高了 15%， ...

Core Viewpoint - Xiaomi's decision to adopt the second-generation 3nm process technology for its new chip, the玄戒O1, marks a significant advancement in its chip development journey, positioning the company as a key player in the semiconductor industry [3][6][10]. Group 1: Chip Development Journey - Xiaomi has committed to a long-term investment plan of at least 500 billion yuan over the next decade to enhance its chip development capabilities [6][7]. - The company is now the fourth globally to release a self-developed 3nm chip, following Apple, Qualcomm, and MediaTek, highlighting its rapid progress in chip technology [7][8]. - The average design cost for a 3nm chip is approximately 1 billion USD, with Xiaomi having invested 13.5 billion yuan in R&D as of April this year [7][8]. Group 2: Strategic Importance of 3nm Technology - The adoption of 3nm technology is seen as a strategic move to enhance Xiaomi's bargaining power with chip suppliers and mitigate supply chain risks [3][11]. - Advanced process technology is crucial for improving smartphone chip performance, especially with the upcoming 6G technology [4][8]. - Xiaomi's choice to leap directly to 3nm avoids redundant efforts and positions the company for future optimizations [8][9]. Group 3: Market Position and Future Prospects - Despite entering the chip manufacturing space, Xiaomi's chips are expected to initially be used in its mid-range models, with high-end models likely continuing to rely on Qualcomm [10][11]. - The company aims to leverage its 3nm chip technology in various smart devices, including AR glasses and TVs, while also exploring applications in AI technology [10][11]. - Long-term, Xiaomi's chip team is expected to gain valuable experience, potentially positioning the company as a leading domestic chip design firm akin to Qualcomm in the future [12].

Core Viewpoint - Xiaomi has announced the development of its self-designed 3nm chip, the "Xuanjie O1," which is a significant advancement over previous industry expectations of 7nm and 4nm processes, marking a strategic move in its semiconductor journey [1][2]. Group 1: Investment and Development - Xiaomi plans to invest at least 50 billion yuan (approximately 7.1 billion USD) over the next decade in its chip development efforts [2][3]. - As of April 2023, Xiaomi has already invested 13.5 billion yuan (approximately 2 billion USD) in research and development for this chip, with an expected R&D expenditure of over 6 billion yuan (approximately 850 million USD) this year [3]. Group 2: Technological Significance - The transition from 28nm to 3nm technology is crucial for enhancing smartphone chip performance, with each generation of technology providing significant performance improvements [1][2]. - Xiaomi is the fourth company globally to release a self-developed 3nm mobile processor, following Apple, Qualcomm, and MediaTek [2]. Group 3: Strategic Implications - The adoption of 3nm technology enhances Xiaomi's negotiating power with chip suppliers and serves as a strategic measure against supply chain risks and the perception of dependency on imports [1][8]. - The company aims to use its self-developed chips primarily in its own devices, particularly in mid-range models, while high-end models may still rely on Qualcomm chips for the time being [6][8]. Group 4: Future Prospects - The experience gained from developing the 3nm chip could position Xiaomi as a leading chip design company in China in the long term, potentially rivaling companies like Qualcomm in the future [9]. - The 3nm chip may also find applications beyond smartphones, including in AR glasses and AI devices, although cost considerations will dictate its use in high-end products [7][8].