通信 周跟踪（20260202-20260208） 领先大市-A(维持) ——北美资本开支增长并未见顶，Scaleup 是光通信重要增量 2026 年 2 月 13 日 行业研究/行业周报 通信行业近一年市场表现 行业动向： 1）北美主要 CSP 陆续发布财报，AI 飞轮高速运转，2026 资本开支指引超 预期。近期，北美主要 CSP 完成了 2025 年度财报的公布，普遍向外界传达的信 号是 AI 已从成本中心转为收入引擎，继续扩大投资急迫且真实。 资料来源：常闻 相关报告： 【山证通信】——moltbot 爆火背后 1.1 云业务通过 AI 算力与工具实现高增。例如，AWS 云销售额同比增长 20% 达到 1287 亿美元，其 Bedrock 上托管的 20+个模型得到客户来自项目编程、推理 和代理工作流程的全面需求。谷歌云 2025 年末年化营收突破 700 亿美元成为增 长最快的核心业务，主要得益于企业 AI 基础设施、AI 解决方案及核心云服务的 强劲客户需求。微软 Azure 的年收入已超过 750 亿美元，同比增长 34%，主要来 自 OpenAI 等大客户的长期订单、Azure AI Foun ...

Investment Rating - The report maintains an "Outperform" investment rating for the communication industry [4][7]. Core Insights - In January 2026, the communication industry index increased by 5.47%, outperforming the Shanghai Composite Index (+3.76%), CSI 300 Index (+1.65%), Shenzhen Component Index (+5.03%), and ChiNext Index (+4.47%) [3][13]. - The capital expenditure of the four major North American cloud providers is expected to grow significantly, with a projected increase of over 60% in 2026 [6][24]. - The procurement results for special optical cables by China Mobile indicate that eight manufacturers, including Tongding Interconnection and Yangtze Optical Fibre and Cable, have been selected, highlighting the focus on key sectors such as industrial manufacturing and digital government [6][7]. Summary by Sections Industry Performance - The communication industry index showed a strong performance in January 2026, with a 5.47% increase, outperforming major indices [3][13]. - Sub-sectors within the communication industry saw varied performance, with cable, other communication equipment, and system equipment rising by 19.70%, 10.72%, and 7.85% respectively [16]. Telecommunications Sector - In 2025, the telecommunications business revenue reached CNY 1.75 trillion, a year-on-year increase of 0.7% [6][45]. - By December 2025, 5G mobile phone users accounted for 65.9% of total mobile phone users, with a monthly data usage (DOU) of 23.04GB per user, reflecting a 17.0% year-on-year increase [6][45]. Cloud Infrastructure Investment - The combined capital expenditure of the four major North American cloud providers in Q4 2025 was USD 126 billion, marking a 62.0% year-on-year increase [24][25]. - For 2026, the total capital expenditure guidance for these providers exceeds USD 660 billion, with an expected growth of 61.0% [24][25]. AI and Smartphone Market - The retail sales of communication equipment in China increased by 20.9% year-on-year in December 2025, driven by the demand for smartphones [44]. - The global smartphone shipment is projected to grow by 2% in 2025, reaching 1.25 billion units, with AI smartphones expected to penetrate 34% of the market by 2025 [6][44]. Investment Recommendations - The report suggests focusing on companies involved in optical devices, optical chips, optical modules, and AI smartphones, including Tianfu Communication, ZTE, and China Mobile [7][6].

Core Insights - The article highlights the impressive performance of Gemini 3 Deep Think, which has achieved significant milestones in various benchmark tests, showcasing its advanced reasoning capabilities and potential applications in scientific and engineering fields [3][15][19]. Group 1: Performance Metrics - Gemini 3 Deep Think scored an unprecedented 84.6% on the ARC-AGI-2 benchmark, surpassing previous models that scored between 60%-70% [3][19]. - In the Humanity's Last Exam (HLE), it achieved a new state-of-the-art (SOTA) score of 48.4% [3][15]. - The model also reached a remarkable Elo score of 3455 on Codeforces, ranking it as the 8th best globally, with only 7 individuals scoring higher [1][15]. Group 2: Cost Efficiency - The upgrade of Gemini 3 Deep Think has led to an 82% reduction in reasoning costs, decreasing from $77.16 to $13.62 per task [21][15]. Group 3: Applications and Innovations - Gemini 3 Deep Think has demonstrated capabilities in analyzing sketches, modeling complex shapes, and generating files for 3D printing, indicating its utility in engineering tasks [7][15]. - The model successfully identified a subtle logical flaw in a complex mathematical paper, which had been overlooked in prior peer reviews, showcasing its potential in academic research [9][15]. - It optimized a method for growing complex crystals, achieving a precision previously unattainable, which could lead to new semiconductor materials [10][15]. Group 4: Team and Development - The development team of Gemini 3 Deep Think includes notable figures such as Yi Tay and Shunyu Yao, both of whom have significant backgrounds in AI and physics [27][28].

Group 1 - The software sector is experiencing the most intense short-selling attack in over a decade, with Morgan Stanley reporting that the short-selling volume in the software and SaaS sectors reached one of the highest levels since 2010 [1][7] - Hedge funds have quickly resumed short-selling strategies after a brief period of covering, with new short positions in the software sector exceeding levels seen at the end of January [2][7] - Concerns about AI replacing human jobs are spreading across more sub-industries, significantly impacting market sentiment, as evidenced by a notable drop in the stock price of CH Robinson [4][8] Group 2 - Defensive sectors are outperforming cyclical stocks, with Goldman Sachs noting the worst two-day performance for cyclical stocks since the "Liberation Day," with a cumulative drop of over 350 basis points [6][8] - The 10-year U.S. Treasury yield has fallen to around a three-month low of 4.08%, indicating increased market risk aversion, while the VIX index closed above 20 [6][8] - The market is becoming increasingly sensitive to the potential disruptive impacts of AI, with major tech companies experiencing valuation declines [9] Group 3 - Goldman Sachs described the current trading environment as one of the most volatile seen, with many tech and growth stocks appearing oversold as the Russell Tech Index falls to its 200-day moving average [6][10] - There is a discussion on whether the valuation adjustments in the market have been excessive, with some companies not returning to pre-earnings release levels despite the rapid pricing in of perceived technological changes [10] - The large tech stocks, referred to as "Mag 7," have underperformed the market by approximately 7.5 percentage points over the past few months, indicating a significant pullback [11]

听雨 发自 凹非寺 量子位 | 公众号 QbitAI 面对Claude Opus 4.6和GPT Codex 5.3的猛烈攻势，谷歌反手就是一个 Gemini 3 Deep Think 的重大升级。 | | Who | # | = | | --- | --- | --- | --- | | 1 | _ Benq | IEE | 3792 | | 2 | 롤 ecnerwala | 214 | 3715 | | 3 | jiangly | 192 | 3664 | | 4 | VivaciousAubergine | 65 | 3646 | | 5 | Kevin114514 | 104 | 3604 | | б | tourist | 297 | 3592 | | 7 | strapple | ୧୧ | 3486 | | 8 | dXqwq | 83 | 3436 | | ਰੇ | · maroonrk | 208 | 3423 | | 10 | Otomachi_Una | 62 | 3413 | | 11 | မြ ksun48 | 313 | 3408 | | 12 | heuristica | ୧୫ ...

谷歌DeepMind掌门人Demis Hassabis在达沃斯论坛上描绘了AI的终极图景：不仅要在2026年将AI设计的药物送入临床试验，更预言未来10到15年人类将迎 来"发现的黄金时代"。 在瑞士达沃斯举行的世界经济论坛期间，谷歌DeepMind首席执行官德米斯·哈萨比斯（Demis Hassabis）接受了《财富》杂志的专访。 在这场关于人工智能未来的深度对话中，哈萨比斯不仅披露了谷歌大模型Gemini的最新进展，更重点阐述了AI在生物医药领域的颠覆性潜力，以及智能体 （Agents）和机器人技术的爆发时间点。 剑指"解决所有疾病"：2026年或成转折点 作为AlphaFold（蛋白质结构预测模型）的缔造者，哈萨比斯正试图通过其创立的初创公司Isomorphic Labs将这一技术转化为实际的药物疗法。 他在采访中透露了一个关键的时间节点：2026年。哈萨比斯表示： "我们希望到2026年初，首个药物能进入临床试验阶段。" Isomorphic Labs目前正在研究癌症、心血管疾病和免疫学等多个领域，并已与礼来公司（Eli Lilly）及全球顶级制药巨头建立了合作伙伴关系。目前，该公 司已有17个药物项 ...

刚刚，谷歌Gemini 3 Deep Think原地进化！在Codeforces比赛中拿下全球第七，击败了人类选手。短短三个月，全方位刷爆SOTA。 猝不及防，谷歌DeepMind深夜又放大招了！ 今天，Gemini 3 Deep Think重磅升级，几乎刷爆全领域的SOTA，标志着AI推理能力进入了全新维度。 离职Anthropic入职谷歌的华人学者姚顺宇参与了Gemini 3 Deep Think 这一次，在科学研究和硬核工程领域，Deep Think堪称一个「最强大脑」。 它可以将草图渲染成一个高保真、实用的3D笔记本电脑支架图，并直接将其打印出来。 谷歌VP晒出这个副项目，最终的成品是这样子的。 9 ll l 6 | > Top rated | | | | --- | --- | --- | | # | User | Rating | | 1 | Benq | 3792 | | 2 | ecnerwala | 3715 | | 3 | jiangly | 3664 | | 4 | VivaciousAubergine | 3646 | | 5 | Kevin114514 | 3604 | | 6 | ...

从定位上看，Codex-Spark 并不是为了替代现有的 Codex，而是补齐其在"即时交互"场景中的短板：在过去，Codex 更擅长长时间运行的复杂任务，而 Codex-Spark 的目标则非常明确——把人与模型之间的交互延迟压缩到接近"无感"的程度。 这一发布同时也是 OpenAI 与 芯片初创企业 Cerebras 合作的重要阶段性成果。为了减少对英伟达芯片的依赖，上个月 OpenAI 签署了一项金额超过 100 亿 美元的协议，使用 Cerebras 的硬件以提升其模型的响应速度，而 Codex-Spark 被视为这项合作落地的第一个技术里程碑。 为实时而生：Codex-Spark 的核心是"速度" OpenAI 发布新模型，专为实时编码而生 昨晚，OpenAI正式发布了GPT-5.3-Codex-Spark的研究预览版本。这是一款从 GPT-5.3-Codex 主模型中"裁剪"而来的精简版本，同时也是 OpenAI 首个专 门围绕实时编码（real-time coding）场景设计的模型。 在官方定义中，Codex-Spark 是一个"专为实时使用 Codex 而设计的模型"，它支持进行针对性编辑、 ...

Core Viewpoint - Amazon has officially entered a technical bear market after its stock price dropped 21.4% from recent highs, while Microsoft became the first member of the Mag7 to enter a bear market. Investor confidence in the return on investment from AI spending by these companies is low, with Amazon planning capital expenditures of up to $200 billion by 2026 [1][3]. Group 1: Amazon and Microsoft Performance - Amazon's stock closed at $199.60 after eight consecutive days of decline, marking its entry into a technical bear market as the second company in the Mag7 to do so [3]. - Microsoft’s stock price fell 25.9% from recent highs after its Azure cloud business growth failed to meet investor expectations, leading to its entry into a bear market on January 29 [3]. Group 2: AI Spending and Investor Sentiment - The total projected capital expenditure for AI by Amazon, Microsoft, Meta, and Alphabet is expected to reach $650 billion by 2026, with Amazon having the highest planned expenditure [3]. - Meta's stock is close to entering a bear market, just 2.3% away from the threshold, despite exceeding Wall Street expectations for Q4 revenue and earnings, as increased AI spending and profit margin pressures have affected investor confidence [3]. Group 3: Market Dynamics and Cash Flow Concerns - There is a noticeable rotation among Mag7 members, with investors moving away from Microsoft, Nvidia, and Oracle-related OpenAI trades towards Alphabet and Broadcom ecosystems [5]. - Alphabet's vertically integrated technology stack has somewhat mitigated concerns over excessive spending, allowing it to avoid the worst impacts of the tech stock sell-off, with its stock down 9.2% from recent highs [5]. - Increased capital expenditure levels for Amazon may lead to negative free cash flow this year, prompting the company to seek additional capital through debt markets [5].

Summary of Key Points from Conference Call Records Industry and Company Involvement - **Rare Earth Industry**: Emphasis on the rational development of rare earth resources and innovation in rare earth technology [1] - **Tungsten Industry**: Significant price increase in tungsten raw materials, with current prices reaching five times the level at the beginning of 2025 [2] - **Nickel Mining**: Indonesia's substantial reduction in nickel mining output, with expected production between 260 million to 270 million tons, down from 379 million tons in 2025 [3][4] - **Glass Fiber Industry**: Taiwanese manufacturers shifting production from E-glass to low-DK glass fabric due to increased demand [5] - **Cloud Computing and AI**: Meta's investment in a data center and predictions of massive growth in AI token consumption in China [6][7] - **Chemical Industry**: Price adjustments in the pesticide sector and incidents of illegal chemical production leading to regulatory actions [8] - **Dye Industry**: Price increases in disperse dyes due to low inventory levels and market consolidation [9] - **Battery Technology**: Development of solid-state battery standards and strategic partnerships for next-generation battery materials [10] Core Insights and Arguments - **Rare Earth Development**: The need for innovation in rare earth technology to enhance resource utilization [1] - **Tungsten Price Surge**: A 50% increase in tungsten prices within a month, indicating strong demand and potential supply constraints [2] - **Nickel Production Cuts**: Indonesia's drastic cuts in nickel production quotas, with a 70% reduction in the largest nickel mine's output [3][4] - **Shift in Glass Fabric Production**: Taiwanese companies adjusting production lines to meet the rising demand for low-DK glass fabric [5] - **AI and Cloud Market Growth**: Significant investments in AI infrastructure and predictions of a 370-fold increase in AI token consumption by 2030 [6][7] - **Chemical Pricing and Regulation**: The average cost of chlorantraniliprole and regulatory measures against illegal chemical production [8] - **Dye Price Increases**: Anticipated further price hikes in disperse dyes due to low inventory and market dynamics [9] - **Solid-State Battery Development**: Collaboration between companies to innovate in solid-state battery technology [10] Other Important but Potentially Overlooked Content - **Meta's Data Center Investment**: A $10 billion investment in a new data center in Indiana, expected to be operational by late 2027 or early 2028 [6] - **AI Model Developments**: The launch of new AI models and tools aimed at enhancing programming capabilities [7] - **Regulatory Actions in Chemical Sector**: Increased scrutiny and actions against illegal chemical production following a fatal incident [8] - **Market Dynamics in Dye Industry**: The consolidation of the dye industry leading to fewer manufacturers and potential for price increases [9] - **Battery Technology Partnerships**: Strategic collaborations aimed at advancing solid-state battery technology [10]