大族数控在港股敲钟上市，跻身"A+H"行列。（资料图片） 深圳商报首席记者 谢惠茜 2月6日上午，随着卓正医疗、大族数控在港股敲钟上市，宣告深圳迎来了里程碑时刻：全市上市公司总量突 破600家大关，其中，境内上市公司426家，境外上市公司174家。截至2026年1月末，深圳上市公司总市值超 19万亿元，稳居全国大中城市第二位。 这不仅是一个数字的跨越，更是一座城市实体经济活力、创新动能与金融生态深度融合的生动映照。从新中 国第一张股票在宝安破土而出，到今天600个火炬熠熠生辉，深圳企业的上市史，几乎就是半部中国现代企业 制度的演进史。 拓荒奠基 从"第一股"到全球销冠 回望来路，那些"拓荒者"的名字镌刻于中国资本市场的史册。 1983年，宝安县联合投资公司向社会公开发行"原始股"，催生了改革开放后深圳第一家股份制企业——深宝 安，也诞生了新中国第一张股票。1991年6月25日，深宝安登陆深交所，成为中国资本市场第一批上市企业； 1991年发行的"南玻B"，则成为中国内地第一支B股。 更早的1987年，招商银行作为全国第一家由企业法人持股的商业银行在深圳成立，中国平安也在蛇口呱呱坠 地，日后成长为世界级的综合金 ...

Core Insights - Anthropic has completed a significant funding round of $30 billion, raising its post-money valuation to $380 billion, more than doubling its valuation from the previous round in September last year, making it the second-largest private tech financing in history [1] - The funding round was led by Coatue and GIC, with participation from several notable investors including Microsoft and NVIDIA, who have committed up to $50 billion and $100 billion respectively [1] Group 1: Funding and Valuation - Anthropic's funding round reflects the increasing costs associated with AI model development and training, particularly the reliance on NVIDIA GPUs and competition with tech giants like Google [2] - The company reported an annual revenue of $14 billion, up from approximately $10 billion last year, indicating strong growth driven by enterprise demand for its AI tools [2][3] Group 2: Business Model and Clientele - Approximately 80% of Anthropic's revenue comes from enterprise clients, significantly boosted by its AI programming tool, Claude Code, which has seen a fourfold increase in enterprise subscriptions this year [3] - The popularity of Claude Code and advancements in other productivity products like Claude Cowork have triggered market reactions, leading to significant sell-offs in the software sector [3] Group 3: Competitive Landscape - OpenAI is accelerating its product and funding initiatives in response to competitive pressures, including the promotion of its AI programming tool Codex and discussions for a new funding round potentially nearing $100 billion [3] - Anthropic's latest model, Claude Opus 4.6, has shown improvements in coding capabilities and high-quality content generation, reflecting strong market demand for its products in critical work scenarios [4]

Core Insights - DeepSeek has initiated a gray testing phase for its flagship model, allowing for a context length of up to 1 million tokens, significantly expanding from the previous 128K tokens in version V3.1 released in August last year [1] - Users have reported mixed reactions to the recent updates, with some expressing dissatisfaction over the model's change in tone and interaction style, leading to a trending topic on social media regarding its perceived coldness [1][4] Group 1: Model Updates and Features - The latest version of DeepSeek supports the processing of extremely long texts, as demonstrated by its ability to handle a document with over 240,000 tokens [1] - The upcoming DeepSeek V4 model is expected to be released in mid-February 2026, with the current version being a speed-optimized variant that sacrifices some quality for performance testing [6] - DeepSeek's V series models are designed for optimal performance, with V3 marking a significant milestone due to its efficient MoE architecture [6] Group 2: User Feedback and Reactions - Users have criticized the new version for its impersonal approach, referring to users as "users" instead of personalized nicknames, which has led to a perception of the model being less engaging [4] - Some users have described the updated model as overly simplistic and lacking emotional depth, comparing its output unfavorably to older literary styles [4] - Conversely, a segment of users appreciates the model's newfound objectivity and rationality, noting that it appears more attuned to the psychological state of the questioner [5] Group 3: Technical Innovations - DeepSeek has introduced two innovative architectures: mHC for optimizing information flow in deep Transformers, enhancing stability and scalability without increasing computational load, and Engram for decoupling static knowledge from dynamic computation [7] - These innovations aim to significantly reduce the cost of long-context reasoning while maintaining performance [7]

Group 1 - US President Donald Trump anticipates negotiations with Iran to conclude within a month, aiming for a diplomatic agreement to curb Iran's nuclear ambitions [4][6][19] Group 2 - US existing home sales experienced the largest decline in nearly four years, dropping by 8.4% in January to an annualized rate of 3.91 million units, influenced by severe winter storms [8][22] - The South region, the largest housing market in the US, saw a 9% decrease in sales, falling to an annualized rate of 1.81 million units [8][22] Group 3 - Anthropic completed a $30 billion financing round, raising its post-money valuation to $380 billion, making it one of the highest-valued private companies globally [10][25] - The funding round was led by Singapore's government investment company and Coatue Management, with participation from notable investors including Nvidia and Microsoft [10][25] Group 4 - Denmark's second-largest pension fund, ATP, is considering reducing its exposure to the US private market due to potential investment risks, with a current portfolio valued at approximately $18 billion [11][27] - The CEO of ATP indicated a comprehensive evaluation of the US political system is underway, questioning the sustainability of past performance [11][27] Group 5 - Hedge fund manager David Einhorn is betting on significant interest rate cuts by the Federal Reserve, suggesting that the number of cuts could exceed current market expectations [13][29] - Einhorn has purchased SOFR futures, anticipating that aggressive rate cuts will lead to an increase in the value of these contracts [13][29] Group 6 - The AI landscape is shifting, with companies like Alphabet and Anthropic making significant impacts, while OpenAI's presence has diminished in discussions [15][31] - Analysts suggest that OpenAI may soon release a new model to regain market attention, which could positively affect related stocks [15][32]

Anthropic已完成一轮300亿美元融资交易，融资后估值达3800亿美元（含本轮资金），这为这家在与竞 争对手OpenAI竞争中逐步取得进展的人工智能公司提供了进一步支持。Anthropic周四表示，新加坡主 权财富基金GIC和Coatue Management领投本轮融资，其他投资者包括英伟达和微软等大型科技公司。 本轮融资使Anthropic的估值较此前几乎翻倍，使其成为全球最有价值的私营公司之一。Anthropic还确 认，将允许员工按本轮融资相同估值出售公司股份。 ...

价值50亿的红包正在向你涌来！ 春节倒计时，大厂红包大战的氛围越来越浓。文心、元宝、千问、豆包，这些在科技圈耳熟能详的AI产品，正试图借助一年一度的春节流量洪峰从极客玩具 升级成大众工具。 熟悉的红包营销手段，却有着截然不同的运营目标：AI的全民化破圈。 当大洋彼岸的OpenAI用GPT-5重塑一切，当Google的Gemini悄然融入数十亿用户的搜索框时，中国的科技大厂们坐不住了。这场被称为"首届AI春节红包 大战"的混战背后，是2025年中国互联网关于AI的集体焦虑，或者说，FOMO（Fear of Missing Out，害怕错过）情绪的集中爆发。 无需多言，红包大战一定会成为AI技术普及的催化剂。十一年前的春节，微信红包引爆移动支付，让全球用户跑步进入无现金支付时代。但现在，一个问题 却不容回避：当前AI产品成熟度到了移动支付当时的状态吗？如果没有，声势浩大的AI红包大战，还能复刻当年移动支付大战的成功叙事吗？ AI红包大战能否复刻移动支付大战？ 回望移动互联网时代那几场经典的红包战役，移动支付、网约车、短视频、外卖都有一个共同特征：加速的是成熟度极高的应用级产品。比如2014年红包 大战爆发时，扫码 ...

Core Viewpoint - The introduction of Algorhythm Holdings' new AI tool, SemiCab, has led to significant declines in the stock prices of several trucking and logistics companies, raising investor concerns about reduced service demand in the freight sector due to the elimination of inefficiencies [1][2] Group 1: Market Reaction - Major trucking and logistics stocks such as C.H. Robinson and RXO fell over 20% during Thursday's trading session [1][2] - J.B. Hunt Transport Services experienced a decline of approximately 9%, while XPO's stock dropped nearly 7.9% [1][2] - Expeditors International of Washington saw a decline of nearly 16.5% [1][2] Group 2: AI Tool Impact - Algorhythm's stock, previously a penny stock, surged approximately 31% following the announcement of the SemiCab tool [1][2] - The market is increasingly scrutinizing traditional companies that may struggle to keep pace with the rapid advancements in AI technology [1][2] Group 3: Analyst Insights - Analyst Daniel Moore from Baird noted an emerging debate around open-source automation agents like Molt Bot, which could enhance the automation of routine backend tasks and level the technological playing field for smaller operators [1][2]

谷歌的深度思考模型Gemini 3 Deep Think迎来重大升级，将其专业推理能力从抽象理论推进至实际应用场景。这一升级聚焦于解决现代科学研究与工程领域 的复杂挑战，标志着谷歌在企业级AI市场的战略性押注。 美东时间12日周四，谷歌官宣Gemini 3 Deep Think升级，称升级后的模型在多项行业基准测试中取得突破性成绩，包括在"人类的最后考试"Humanity's Last Exam（HLE）基准测试、ARC-AGI-2测试中获得84.6%的成绩，经ARC Prize基金会验证；在竞技编程平台Codeforces上，Gemini 3 Deep Think获得3455的Elo 评分。 升级后的深度思考模式即日起面向Google AI Ultra订阅用户开放，同时通过Gemini API向部分研究人员、工程师及企业用户提供早期访问权限。谷歌表示， 该模型已在实际研究中展现应用价值，从发现研究论文中的逻辑漏洞到优化半导体材料生长工艺。 这一发布使谷歌在AI推理模型竞争中与OpenAI的o1系列和Anthropic的Claude正面交锋。随着通用AI能力日趋商品化，专业推理能力成为企业级市场的新战 场，而 ...

经济观察网泰勒斯公司(TU.N)于2026年2月12日盘前发布了2025财年第四季度及全年财报，AI业务收入 已接近8亿美元的年度目标，公司计划到2028年将AI收入提升至约20亿美元。 业绩经营情况 机构对2025年第四季度营收预期为54.08亿加元，每股收益预期0.211加元。 业务进展情况 财务状况 季度股息上调4%至0.4184加元，并计划自2026年起逐步取消折价股息再投资计划。当前杠杆率为3.5 倍，公司目标在2027年底降至3倍。 股票近期走势 近7个交易日股价区间振幅2.73%，累计跌幅0.14%，同期电信服务板块上涨1.99%。 机构观点 加拿大首个主权AI工厂于2025年9月投入运营，作为英伟达云合作伙伴，其客户拓展进展成为焦点。 2026年2月数据显示，买入或增持观点占比53%，目标价均值为15.41美元，较2月11日收盘价存在上行 空间。 以上内容基于公开资料整理，不构成投资建议。 ...

Group 1 - The core focus for investment in the upcoming year includes three main areas: domestic AI ecosystem, innovative pharmaceuticals, and commercial aerospace [1][2] - The domestic AI industry is expected to replicate the growth trajectory seen overseas, with significant advancements in large model technology and semiconductor manufacturing [1] - The AI industry chain is anticipated to strengthen progressively, with downstream sectors centered around large models and cloud computing, and upstream sectors including AI chips, storage, semiconductor manufacturing, equipment, and materials [1] Group 2 - The innovative pharmaceutical sector has shown explosive growth over the past 2-3 years, despite a recent pullback due to valuation pressures; companies with international capabilities are seen as good investment opportunities [2] - The commercial aerospace industry is expected to accelerate commercialization over the next 5-10 years, driven by breakthroughs in rocket recovery and low-orbit satellite technologies, making it a sector worth close attention [2]