Core Viewpoint - Tianshu Zhixin (09903) experienced a significant stock price increase of over 13%, closing at 193 HKD with a trading volume of 1.81 million HKD, indicating strong market interest in the company’s developments [1] Company Overview - Tianshu Zhixin provides general GPU products and AI computing solutions tailored for various industries. The product portfolio includes general GPU chips and accelerator cards, as well as customized AI computing solutions, which encompass general GPU servers and clusters [1] Product Development - The company recently released a roadmap for its fourth-generation chip architecture, with expectations to surpass NVIDIA's Rubin architecture by 2027. Additionally, Tianshu Zhixin unveiled its new edge computing product series, "Tongyang," claiming superior performance in multiple real-world scenarios compared to NVIDIA's AGX Orin [1]

Core Viewpoint - The stock price of Cambricon, a leading domestic AI chip company, experienced a significant drop of 9.18% on February 3, closing at 1128 yuan per share, largely due to market rumors regarding its revenue projections for 2026 being significantly lower than expectations [1][3]. Group 1: Stock Performance and Market Reaction - Cambricon's stock price fell sharply, with an intraday decline of up to 14%, resulting in a market capitalization loss of over 48 billion yuan in one day [1][3]. - The company issued a statement denying the rumors about its revenue guidance and emphasized that it has not organized any small-scale communications recently [1][3]. - Despite a slight recovery in stock price after the statement, negative market sentiment persisted [1]. Group 2: Revenue Projections and Financial Outlook - Cambricon projected its 2025 annual revenue to be between 6 billion to 7 billion yuan, representing a year-on-year growth of 410.87% to 496.02%, with a net profit forecast of 1.85 billion to 2.15 billion yuan [3]. - The expected net profit for Q4 2025 is estimated to be between 245 million to 545 million yuan, indicating a significant decline compared to analyst expectations of 613 million yuan [3]. Group 3: Industry Context and Competitive Landscape - The adjustment in the AI chip sector reflects both domestic and international market pressures, with concerns about the saturation of AI infrastructure driven by large language models [2]. - The domestic market is seeing an influx of new AI chip companies, which may dilute Cambricon's market position and raise questions about its valuation and future performance [2]. - Other companies in the sector, such as Tenstorrent and Mozi Technology, are actively launching new products and providing detailed updates on their technological advancements, contrasting with Cambricon's limited communication [4].

Core Insights - Domestic GPU companies are seizing the industry window period with intensive initiatives, exemplified by TianShu ZhiXin's release of its fourth-generation architecture roadmap, aiming to surpass NVIDIA's Rubin architecture by 2027 [1][2] Group 1: Product Development and Architecture - TianShu ZhiXin has introduced a roadmap for its four-generation GPU architecture, with specific timelines: "TianShu TianShu" in 2025 to surpass NVIDIA's Hopper architecture, "TianShu TianXuan" in 2026 to match Blackwell, and "TianShu TianQuan" in 2027 to exceed Rubin [3] - The company emphasizes energy efficiency and improved parallel computing resource utilization in its architecture design, achieving over 90% effective utilization in AI-related calculations [3][4] - The efficiency of the "TianShu TianShu" architecture is reported to be 60% higher than the industry average, with a 20% performance increase over NVIDIA's Hopper architecture in specific scenarios [4] Group 2: Product Launches and Market Positioning - TianShu ZhiXin has launched the "TongYang" series of products aimed at edge computing, which includes four models catering to various computational needs [5][6] - The "TongYang" series has demonstrated superior performance compared to NVIDIA's AGX Orin in multiple real-world scenarios, with the goal of becoming the leading domestic solution for edge computing [6][8] Group 3: Market Dynamics and Competitive Landscape - As of now, TianShu ZhiXin's products have served over 300 clients and completed more than 1,000 deployments across various sectors, indicating a growing commercial application [8] - The domestic GPU market is witnessing increased activity, with several companies like MoEr Thread, MuXi Co., and BiRan Technology completing IPOs, reflecting the ongoing interest in hard technology [8][9] - Despite the enthusiasm in the capital market, domestic GPU companies still hold a small market share compared to NVIDIA, which commands a significant global market presence [9]

Core Insights - The article discusses the significant advancements made by the Chinese AI computing company, Tensu Zhixin, which has recently launched a roadmap to surpass international giants like Nvidia's Hopper, Blackwell, and Rubin by 2025 to 2027 [2][4][5] - Tensu Zhixin's new architecture, the Tensu Tian Shu, has already demonstrated a performance improvement of approximately 20% over Nvidia's Hopper in key model scenarios, marking a substantial leap for domestic solutions [4][5][32] - The company aims to redefine the narrative of domestic GPU industry by moving away from a "benchmarking" approach to a self-defined leadership in AI computing [4][33] Group 1: Evolution of Computing in China - The roadmap released by Tensu Zhixin outlines a clear timeline and quantifiable breakthroughs, marking a departure from the traditional "follow and catch up" strategy adopted by many domestic companies [5][35] - The architecture is set to achieve a practical utilization rate of over 90% in executing attention mechanisms by 2025, showcasing the company's commitment to high efficiency [5][35] - The upcoming architectures, Tian Xuan and Tian Ji, are expected to further enhance performance and address industry-specific computational needs by 2026 [7][37] Group 2: Technological Innovations - Three core technological innovations underpin the aggressive roadmap: TPC BroadCast, Instruction Co-Exec, and Dynamic Warp Scheduling, which collectively enhance computational efficiency and resource utilization [10][39] - The company has adopted a problem-oriented research and development approach, addressing common industry pain points such as FP8 accumulation precision and matrix transposition overhead [11][40] - This focus on practical solutions has resulted in significant performance improvements, such as a 50% reduction in memory usage for model inference [14][44] Group 3: Redefining Value in Computing - Tensu Zhixin proposes a new value coordinate system for the computing industry, emphasizing high efficiency, predictability, and sustainability as key competitive factors [12][41] - The company aims to optimize total cost of ownership (TCO) by enhancing effective computing output per unit of power, thus reducing unnecessary costs for enterprises [14][43] - The design philosophy ensures that hardware remains adaptable to future algorithmic advancements, extending its lifecycle and value [16][44] Group 4: Market Positioning and Product Launch - The launch of the "Tongyang" series of edge computing products fills a gap in the domestic high-end edge computing market and establishes a comprehensive "cloud + edge + end" computing layout [18][46] - The product matrix covers a range of performance metrics from 100T to 300T, with specific models tailored for various applications [19][47] - Tensu Zhixin's products have already been deployed in over 300 customer environments, demonstrating their effectiveness in real-world applications [22][49] Group 5: Long-term Strategy in the GPU Industry - The competition in the GPU industry is fundamentally about building an open and collaborative ecosystem, which is essential for long-term success [52][52] - Tensu Zhixin is focused on creating a closed-loop ecosystem for domestic AI computing through hardware foundation, software adaptation, and partner collaboration [53][52] - The company aims to ensure that its products maintain value over a decade, positioning itself as a long-term player in the industry rather than merely a competitor to Nvidia [53][55]

这份路线图，或许正是那块厚实跳板 第一次发出的、清晰可闻的蓄力声。 当大部分同行还在用"对标"作为宣传话术时，这家公司直接把超越Hopper、Blackwell、Rubin的时间 点写进了2025-2027年的日历。底气何来？答案是：过去七年没走捷径。 在AI芯片发布会"人均对标英伟达"的语境下，一份真正敢把超越日期写进标题的架构路线图，本身就 构成了新闻。 1月26日，天数智芯公布了其2025至2027年的四代架构路线图。与寻常的未来可期不同，它像一份产 品手册般精确： 2025年天数天枢超Hopper，2026年天数天璇架构对标Blackwell，同样在2026年，天 数天玑架构超越Blackwell，2027年天数天权超Rubin ，2027年之后将转向突破性计算芯片架构设计。 未来3年，天数智芯将基于此次发布的四代架构，陆续发布多款产品，持续提升计算性能。 更不同寻常的是，这份这份未来宣言的第一行，已被验证。在天数智芯路线图中的最近一代架构—— 2025年推出的天枢，在关键的大模型场景——DeepSeek V3上，平均性能已比英伟达Hopper架构高出 约20%。 "他们不是在预测未来，他们是在汇报进 ...

将投中网设为"星标⭐"，第一时间收获最新推送 作者丨 簪竹 来源丨 投中网 登陆港股18天后，天数智芯再次吸引众人目光。 1月26日，这家AI算力企业开了场发布会，不仅公布了四代架构路线图，同时还发布了边端算力产品"彤央"系列。 这家AI算力企业已经打通商业落地闭环，进入放量增长阶段。 从天数智芯透露的信息来看，他们2025年推出的天数天枢架构已超越英伟达目前主力的Hopper架构，在DeepSeek R3场景中实现性能超越Hooper 20%。此次发布会上，天数智芯正式公布了天枢之后的三代架构：2026年天璇对标Blackwell、天玑超越Blackwell，2027年天权超越英伟达目前最牛 的架构Rubin。 同样值得注意的是，天数智芯新发布的边端算力产品"彤央TY1000"，实测性能已超越英伟达AGX Orin。 随着"彤央"系列正式发布，再加上已推出 的"天垓"系列训练GPU和"智铠"系列推理GPU，标志着他们完成了"云+边+端"全场景算力布局。 通过互联网、科研、金融、医疗等领域众多客户1000多次实战部署，天数智芯的数千卡集群已稳定运行超过1000天。这也意味着，这家AI算力企业已 经打通商业落地 ...

本报讯(记者张文湘)1月26日，上海天数智芯半导体股份有限公司(以下简称"天数智芯")，在"智启芯 程"合作伙伴大会现场发布四代架构路线图，提出以"高效率、可预期、可持续"为核心的"高质量算 力"设计目标，打造"AI++"算力系统新范式，同时发布边端算力产品"彤央"系列，并全面展示多行业深 度应用案例及开放生态建设成果，以布局云边端的全场景产品与解决方案，与产业链上下游的客户和生 态伙伴共绘算力赋能千行百业的发展蓝图。 本次发布中，天数智芯首次公开了公司产品和解决方案在互联网、金融、医疗、科研等领域规模化落地 成果，以及在具身、工业、商业、交通智能等领域的标杆案例，全面印证国产算力的成熟与可靠。 天数智芯针对行业面临的能效比偏低、创造力不足、实际使用困难等问题，提出了"高质量算力"的解决 方案。公司打造的"AI++"算力系统新范式，统一芯片内与芯片外构建算力系统，建立软件驱动算力系 统的全新模式，以底层库为基石，以模型与计算中间层为支柱，用软件的智慧，释放硬件的潜能，支撑 上层不断繁荣的应用生态。 "彤央"系列产品，完成"云+边+端"全场景算力布局。"彤央"系列承载着"赋能边端智慧，连接物理空 间"的核心愿景 ...

Core Viewpoint - The A-share market is experiencing fluctuations, but the sci-tech chip sector is showing resilience and strength, with significant capital inflow driven by "AI catalysis and independent innovation" [1][6]. Group 1: Market Performance - As of 11:01 on January 27, the sci-tech chip ETF Huatai (588750) rose by 2%, with over 220 million yuan net inflow for three consecutive days [1]. - Major component stocks of the sci-tech chip ETF saw significant gains, with Dongxin Co. hitting the daily limit, and other stocks like Chip Source Micro and Yuanjie Technology rising over 14% and 11% respectively [3][4]. Group 2: Industry Trends - The chip sector is receiving strong financial support, with the Shanghai three major leading industries mother fund announcing the selection of 17 sub-funds, including 4 for integrated circuits, totaling 22.89 billion yuan [6]. - The demand for storage chips is driven by AI, with a structural reconfiguration leading to price increases in high-end storage products like HBM and DDR5, expected to continue through 2026 [7]. - The demand for computing chips is surging as AI transitions from training to large-scale inference, resulting in a supply-demand imbalance across all categories of computing chips [8]. Group 3: Domestic Innovation and Market Opportunities - Domestic innovation and localization are accelerating in the chip industry, supported by top-level design and policy funding, creating a rapid growth window for domestic computing leaders [10]. - The domestic equipment procurement rate is projected to rise from approximately 18% in 2022 to around 30% by 2025, with potential growth to 60%-70% in the future [11]. - The sci-tech chip sector is positioned to benefit from both AI demand and domestic substitution, with the Huatai ETF focusing on core segments of the chip industry, showcasing higher growth potential and elasticity [12][14]. Group 4: Investment Insights - The Huatai sci-tech chip ETF is expected to show a net profit growth rate of 94% in the first three quarters of 2025, significantly outperforming peers [14]. - The ETF has demonstrated a maximum increase of 173% since September, indicating strong upward elasticity compared to other industry indices [15].

消息面上，据天数智芯官微消息，近日，在天数智芯"智启芯程"合作伙伴大会上，公司现场发布四代架 构路线图，提出以"高效率、可预期、可持续"为核心的"高质量算力"设计目标，打造AI++算力系统新范 式，预期于2027年超越英伟达Rubin架构，迈向更具创新的突破性架构设计；与此同时，公司发布边端 算力产品"彤央"系列，实测性能优于英伟达AGX Orin，并首次全面展示多行业深度应用案例及开放生 态建设成果。 天数智芯(09903)涨超4%，截至发稿，涨4.57%，报196.8港元，成交额2266.72万港元。 ...

消息面上，据天数智芯官微消息，近日，在天数智芯"智启芯程"合作伙伴大会上，公司现场发布四代架 构路线图，提出以"高效率、可预期、可持续"为核心的"高质量算力"设计目标，打造AI++算力系统新范 式，预期于2027年超越英伟达Rubin架构，迈向更具创新的突破性架构设计；与此同时，公司发布边端 算力产品"彤央"系列，实测性能优于英伟达AGX Orin，并首次全面展示多行业深度应用案例及开放生 态建设成果。 智通财经APP获悉，天数智芯(09903)涨超4%，截至发稿，涨4.57%，报196.8港元，成交额2266.72万港 元。 ...