Core Viewpoint - The Nasdaq 100 index has risen over 20% this year, with semiconductors and network infrastructure leading, while telecom and payment software lag behind. Looking ahead to 2026, the index is expected to show steady returns, particularly in the first half, due to AI spending concerns creating low expectation opportunities [1] Group 1: Key Issues Influencing Tech Stocks in 2026 - The debate on AI is shifting focus from computing power to physical AI (robotics, autonomous driving) and how regulation and return on invested capital (ROIC) will evolve [1] - Valuation recovery in the software industry will be influenced by the end of seat pricing models, the rise of intelligent agents, and competition in large language model (LLM) commercialization [1] - Apple's strategic positioning will be questioned: whether it will act as a defensive growth stock or an AI narrative vehicle, and if foldable smartphones can catalyze growth [1] - The supercycle in commodities, particularly in DRAM, memory chips, and precious metals like gold and copper, will raise questions about who can absorb cost pressures [1] - The paradox of generative AI efficiency, driven by layoffs enhancing productivity, may exacerbate non-farm employment data pressures [1] - Meta and similar companies will be seen as having the most investment value in the debate over profit margins and competition [1] - The potential for cyclical industries such as housing, commercial real estate, analog chips, or automotive sectors to hit bottom will be analyzed [1] - Conditions for hardware stocks to lead will depend on discussions around gross margins and visibility of spending potentially suppressing semiconductor market trends [1] - The development path of LLMs in the race for artificial general intelligence (AGI) will focus on whether US and Chinese models will move towards productization or primitive intelligence competition [1] - Potential blind spots include whether the return of agency businesses or SaaS stocks will become a consensus in 2026 [1]

Core Insights - JD.com is testing a new independent app called "JD AI Shopping," designed as a life service assistant that is "thoughtful, capable of shopping, and caring" [1] Group 1: Product Features - The app integrates JD.com's self-developed language model, Yansai, and differs from traditional Q&A formats by proactively pushing shopping guides, discount information, and lifestyle service inspirations to users [1] - Users can express simple intentions like "I want" in the input box to summon a personal shopping assistant for consumer decision-making and product matching [1] - The "Love Shopping Channel" features an "online shopping street" mode, allowing users to browse while asking questions and quickly access core product information such as parameters and reviews [1] Group 2: Additional Functionalities - The app includes features focused on exclusive discounts, such as "Daily Deals," and supports virtual try-ons through the "AI Outfit" function [1] - It also integrates local life services like food delivery, enabling users to complete the ordering and payment process through simple commands [1]

估值超200亿的AI独角兽，正站在资金链断裂的悬崖边——账面现金只够再烧9个月，即便算上89亿银行授信，撑过两年也是极限？这是智谱冲击"中国AI 第一股"时，招股书中暴露的残酷现实。 但事情远不止"没钱了"那么简单。 首先，钱越烧越快。上半年亏损17.5亿、月均3亿的消耗速度远超营收增速，而平均回款周期却从21天暴增至112天，大客户年年更换、多为一次性交易， 甚至出现了向客户采购额高于销售额的采销倒挂现象，这背后是一个远比现金流更凶险的信号：智谱的商业模式，是否正在丧失造血能力？ 其次，算力越买越贵。三年间算力服务费飙升70倍、吞噬七成研发费用，且全部费用成为"消耗品"，智谱陷入"不烧钱买算力等于技术退场，持续投入则 亏损指数级扩大"的死亡螺旋。即便如此，其GLM系列模型仍被《自然》期刊点赞、被OpenAI列为全球竞争对手，技术成果与商业价值之间的巨大落差， 成为资本市场最矛盾的拷问。 最后，生意越做越重。85%收入依赖本地化部署的"项目制外包"，想转云端SaaS却发现毛利率已跌至-0.4%，阿里、腾讯等巨头正用免费API和迁移补贴席 卷市场。智谱标榜的"独立可控"在G端是优势，在B端和C端却成枷锁——它必 ...

Core Viewpoint - The article highlights the emergence of AI companies from the "SenseTime system," emphasizing their rapid growth and potential to become unicorns, particularly focusing on MiniMax and Vivix AI as key players in the AI landscape [4][10]. Group 1: Company Performance - MiniMax has achieved significant revenue growth, with a reported income of $53.44 million for the first nine months of 2025, surpassing its total revenue of $30.52 million for 2024 [7][9]. - The company is nearing breakeven for its consumer products, indicating a strong commercial viability [7]. - Vivix AI, founded by a former executive from SenseTime, reached a valuation of $1.32 billion within just ten months of its establishment [10]. Group 2: Market Position and Strategy - The "SenseTime system" has produced several successful AI startups, with each major sector in AI featuring companies founded by former SenseTime employees [10][11]. - MiniMax is recognized for its forward-looking strategies, having launched innovative AI applications and models ahead of industry trends, such as the MoE model [20][21]. - The company has a diverse product matrix, which has helped it remain resilient during market fluctuations [21]. Group 3: Talent and Experience - The success of the "SenseTime system" is attributed to the technical expertise and practical experience of its founders, many of whom have a strong background in AI technology and product development [12][18]. - The article notes that the unique combination of technical skills and project experience among these entrepreneurs has made them attractive to investors [15][18]. - The ability to replicate successful strategies and learn from past experiences is emphasized as a key factor in the growth of these companies [26].

Core Insights - Beijing Zhiyu Huazhang Technology Co., Ltd. and MiniMax have recently passed the hearing for their IPOs, marking a significant step for China's AI startup ecosystem, particularly in the large model sector [1][3] - The listings of Zhiyu and MiniMax are seen as a pivotal moment in establishing a closed loop of "technology research and development - capital financing - commercial realization" for large model enterprises, potentially setting a valuation benchmark for future entrants [3] - The entry of Chinese large model companies into the capital market signifies a new phase in the US-China AI competition, extending beyond technological rivalry to capital market dynamics [3] Company Overview - Zhiyu was founded in 2019, focusing on cognitive intelligence large model development, and is a partner for standardized innovation in AI chips for small and medium-sized enterprises in Beijing [4] - The company launched China's first proprietary pre-trained large model framework, GLM, in 2021, and has since developed a model-as-a-service (MaaS) platform [5] - As of June 30, 2025, Zhiyu's models have supported over 8,000 institutional clients and approximately 80 million devices [5] Financial Performance - Zhiyu's revenue has shown significant growth, with figures of 57.4 million, 124.5 million, and 312.4 million for 2022, 2023, and 2024 respectively, reflecting a compound annual growth rate of over 130% [12] - Despite revenue growth, the company has experienced increasing net losses, with figures of 143.7 million, 788.0 million, and 2,958.0 million for 2022, 2023, and 2024 respectively [12][14] - Research and development expenses have surged, constituting a significant percentage of total revenue, indicating a strong focus on innovation [14] Market Dynamics - The large model market is rapidly evolving, with projections indicating that the global market could exceed $300 billion by 2030 [15] - The Chinese large language model market reached 5.3 billion in 2024, with expectations to grow to 101.1 billion by 2030, driven primarily by institutional clients [17] - Independent providers like Zhiyu face different competitive dynamics compared to non-independent providers, as they rely on their native technology and business models [17] Technological Advancements - The AI landscape is undergoing a significant transformation, with large models emerging as a key technology paradigm shift, addressing the need for general-purpose AI [18] - The frequency of model updates among leading companies is accelerating, with major advancements in capabilities and performance [19] - Open-source developments are fostering rapid iteration among closed-source developers, enhancing the competitive landscape [20]

当公募基金资产净值突破36万亿元之际，数亿投资者的财富期盼与实体经济的融资需求在此交汇，高质 量发展已不再是口号，而是基金销售机构必须扛起的时代责任。腾安基金作为腾讯集团旗下的基金销售 机构，自成立之初，便秉持"用户为本、科技向善"的理念，践行金融为民的指导思想。腾安基金深刻认 识到，告别规模至上的旧逻辑、构建以投资者利益为核心的新生态，既是响应《推动公募基金高质量发 展行动方案》号召的应有之义，更是公司行稳致远的根本所在。 腾安基金将行业高质量发展要求转化为长期发展动力，并以金融科技能力和微信生态为依托，通过"让 利于民、优化机制、创新体验"三维实践，书写"以投资者为本"的时代新答卷。 ● 腾安基金销售（深圳）有限公司 产品赋能 精准匹配每一分投资需求 腾安基金始终坚持"选出优秀产品、卖给合适的人、助力投资者拿得住"的理念，并为此构建了以"长期 业绩稳定性+投研能力可持续性+合规风控有效性"为核心的三维评估体系。 公司摒弃短期排名崇拜，更关注基金是否持续跑赢业绩比较基准、能否在市场波动环境下为投资者创造 超额收益；通过精选历经市场牛熊考验的"长青基金"，为投资者构筑坚实防线，链接管理人和投资者长 期利益。 ...

Core Insights - The ongoing IPO process of Zhiyu AI and MiniMax has drawn market attention, highlighting the competitive race for the title of "the first global large model stock" [1] - Despite rapid revenue growth, the company faces significant losses due to substantial R&D investments, indicating a harsh reality of a "burning money competition" in the industry [1][2] Revenue and Losses - Zhiyu AI's revenue for the years 2022 to 2025 is projected to be 57.4 million yuan, 125 million yuan, 312 million yuan, and 191 million yuan, respectively, with a gross margin exceeding 50% [1] - The company's losses during the same period are expected to be 144 million yuan, 788 million yuan, 2.958 billion yuan, and 2.358 billion yuan [1][2] R&D Investments - The company attributes its expanding losses to significant R&D investments, with expenditures of 84.4 million yuan, 528 million yuan, 2.195 billion yuan, and 1.594 billion yuan during the reporting period [2] - In 2024, the company's computing costs are projected to reach 1.553 billion yuan, accounting for 70.7% of total R&D spending [2] Market Position and Strategy - Zhiyu AI ranks first among independent general large model developers in China based on projected revenue for 2024 [1] - The company primarily targets B-end and G-end markets, offering model-as-a-service (MaaS) through localized and cloud deployment options [3][4] Revenue Structure - The revenue structure has shifted from a heavy reliance on localized deployment (97.6% in 2022) to a more diversified approach, with localized deployment accounting for 69.4% and cloud deployment for 30.6% in the first half of 2025 [4] Competitive Landscape - Despite leading in revenue, Zhiyu AI faces competition from major internet companies and emerging players like DeepSeek, which emphasizes the need for cost efficiency in marketing and customer acquisition [5] - The gross margin for the cloud deployment segment has dropped to -0.4% in the first half of 2025, indicating challenges in profitability [5] Industry Outlook - The Chinese large language model market is projected to reach 5.3 billion yuan in 2024, with expectations to grow to 101.1 billion yuan by 2030 [6] - The company's successful IPO process is seen as a milestone for the AI industry, potentially influencing the narrative from "technology storytelling" to "commercial value realization" [6]

Group 1 - Nvidia has reached a non-exclusive licensing agreement with AI chip startup Groq, reportedly worth $20 billion, acquiring Groq's founder Jonathan Ross and engineering team [1] - Groq focuses on LPU chips for inference, achieving an output speed of 500 tokens per second per card, which is ten times faster than Nvidia's GPUs, utilizing a temporal instruction set architecture to mitigate HBM shortages and reduce costs [1] - This transaction represents a "technology licensing + talent acquisition" model, allowing Groq to continue its cloud business independently while Nvidia aims to enhance its inference computing capabilities targeting the Google TPU market [1] Group 2 - Tsinghua TSAIL Laboratory and Shengshu Technology have jointly open-sourced the TurboDiffusion video generation acceleration framework, reducing the processing time of a 1.3B-480P model on a single RTX 5090 from 184 seconds to 1.9 seconds, achieving a 97-fold acceleration [2] - The framework integrates four core technologies: SageAttention2++ quantization, SLA sparse linear attention, rCM step distillation, and W8A8 quantization, decreasing end-to-end latency from 900 seconds to 8 seconds [2] - SageAttention has been successfully integrated into NVIDIA TensorRT and deployed on platforms such as Huawei Ascend and Moole Technology, with major companies like Tencent, ByteDance, and Alibaba already applying it [2] Group 3 - Shanghai Municipal Planning and Resources Bureau and SenseTime have launched the first 600 billion parameter foundational model in the national planning and resources field, named "Yunyu Xingkong," which can answer questions, adjust maps, perform statistics, recognize images, and generate reports [3] - The model is trained on the Kunyu Jinglue corpus and is integrated with the government intranet's professional version and core business systems, achieving a 98% accuracy rate for specialized terms and a 95% approval rate for human Q&A [3] - It employs a "1+6" (base + vertical) model system and an intelligent scheduling engine, supporting natural language calls for 2D and 3D spatial data, exploring a new paradigm for data productization and service-oriented government models [3] Group 4 - Tencent Cloud and Anhui Yilu Weixing have launched the first AI assistant in the ETC field, named "Assistant Agent," based on Tencent's Mix Yuan model, which has served over one million users since its internal testing began in April [4] - The assistant integrates multimodal interaction technology, supporting both text and voice input, achieving a 95% accuracy rate in Q&A and a 90% problem-solving rate, capable of handling complex requests such as device inquiries, traffic record checks, and invoicing [4] - It deploys 105 state monitoring algorithms to collect real-time device operation data, enabling voice interaction and key status reporting for a "service find person" capability, allowing users to control devices via voice commands [4] Group 5 - Dexmal has proposed the GeoVLA framework, utilizing a dual-stream architecture to retain VLM semantic understanding while endowing robots with 3D geometric perception capabilities through point cloud embedding networks and spatial awareness action experts [6] - In the LIBERO-90 long-range multi-task test, it achieved a 97.7% success rate, surpassing OpenVLA-OFT, and reached an average success rate of 77% in ManiSkill2, with an overall average of 86.3% in real-world tasks [6] - It demonstrated outstanding performance in out-of-distribution scene robustness tests, maintaining a 60% success rate with varying basket heights and a 70% success rate with a 45° viewpoint shift, proving its understanding of true 3D spatial structures [6] Group 6 - The SciMaster team, composed of Shanghai Jiao Tong University's TSAIL Laboratory, Shanghai Algorithm Innovation Research Institute, and DeepSense Technology, has launched ML-Master 2.0, achieving a 56.44% medal rate in the MLE-bench, topping the leaderboard [7] - This system is designed for real machine learning engineering, introducing a hierarchical cognitive caching mechanism that models context as Experience, Knowledge, and Wisdom [7] - It employs a "generate-validate" protocol to achieve ultra-long-range autonomous capabilities, with applications already in theoretical computational physics and embodied intelligence, currently open for Waiting List applications via the SciMaster platform [7] Group 7 - Jim Fan, head of embodied intelligence at Nvidia, stated that Tesla's FSD v14 is the first AI to pass the physical Turing test, with Elon Musk noting that "perception is maturing," and the software has been launched in seven countries including the US [9] - Tesla has established 14 technical barriers, including a sensor freezing scheme for 4-6 years to accumulate data, an instant value judgment engine for intelligent data filtering, and a Neural Codec for processing raw Bayer data [9] - The end-to-end transformer facilitates the transition from photon input to motor torque output, with hardware-in-loop quantization training conducted on the Cortex supercomputer's vehicle chip, updating 12 versions within 77 days, although issues remain with lane switching and lane change decisions [9]

来源：toB老人家 责任编辑：何俊熹 来源：toB老人家 责任编辑：何俊熹 ...

Core Viewpoint - Doubao-Seed-1.8 model was officially released on December 18 but was temporarily "shelved" less than 10 days later due to user access issues [1][2]. Group 1: Product Release and Features - Doubao-Seed-1.8 was launched on December 18 and made available on Volcano Engine, targeting enterprises and developers with API access [2]. - The model's multimodal agent capabilities are reported to be comparable to top global models, receiving positive feedback for its agent tools, visual understanding, general language abilities, and deep thinking capabilities [2]. Group 2: User Access and Updates - Users reported issues accessing Doubao-Seed-1.8 starting December 22, leading to inquiries about its status [1]. - Volcano Engine's customer service indicated that adjustments would be made to the visual language model (VLM) capabilities, with an expected update completion by January 4 [2]. - During the public testing phase, only certified enterprise users are allowed to apply for access [2]. Group 3: Company Response - An inquiry was sent to ByteDance regarding the transition of Doubao-Seed-1.8 from official release to public testing status, but no response was received by the time of reporting [2].