Group 1 - In 2025, global investors experienced a significant market rally, with global stock markets rising approximately 30% and the US stock market rebounding over 50% since April [1] - Gold prices surged over 60% in 2025, marking the second-highest annual increase since 1979, driven by heightened demand for safe-haven assets amid high global debt and shifting monetary policies [1] - The strong performance of global assets in 2025 is attributed to the dual drivers of AI and renewable energy, which have become core growth engines in capital markets [1] Group 2 - A report from BCA Research indicates that the biggest threat to the US economy in 2026 may stem from the financial markets themselves, rather than an economic recession [5] - The report highlights that approximately two-thirds of global stock market capitalization is concentrated in US stocks, with 40% of that in just ten stocks, all heavily reliant on the success of the generative AI narrative [5][6] - Recent market trends show a divergence among leading tech stocks, suggesting that not all tech stocks are viewed as a single entity, allowing for value investors to validate certain companies [5][6] Group 3 - Morgan Stanley's chief economist in China suggests a new asset allocation strategy of 60% stocks, 20% gold, and 20% bonds, reflecting a reassessment of strategic asset values amid declining fiat currency credibility [9] - Analysts predict that gold prices could reach $4,500 by mid-2026, driven by structural demand from central banks, which currently account for 20% of the annual gold production of 5,000 tons [13] - The report emphasizes that the performance of gold is closely tied to the depreciation of the US dollar, with approximately 70% of gold's price increase attributed to changes in the dollar's value [10] Group 4 - The AI sector is identified as a key driver of the stock market's rise in 2025, with significant trends emerging in AI, electricity, and longevity economics [12] - The expectation of further interest rate cuts by the Federal Reserve in 2026 is anticipated to exert downward pressure on the US dollar, with predictions of the dollar to RMB exchange rate reaching around 6.9 by mid-2026 [12] - The shift towards AI-driven productivity is expected to transform business operations, moving from traditional models to agile teams that leverage digital agents for task execution [16]

Investment Rating - The report suggests a "Hold" rating for the computer industry, indicating a relative performance within -5% to +5% compared to the benchmark index [44]. Core Insights - The report emphasizes the importance of three core elements for AI applications: data, models, and interfaces, which are crucial for evaluating the potential of AI applications [14][18]. - The competition for AI entry points is intensifying, with major players like OpenAI launching new products such as the ChatGPT Atlas browser, which integrates AI capabilities directly into the browsing experience [22][27]. - The demand for computing power is surging, highlighted by Anthropic's partnership with Google to deploy 1 million TPU chips, aiming for over 1GW of computing capacity by 2026 [30][28]. Summary by Sections Custom Agent Insights - The report discusses the evolution of AI agents, noting that successful applications require specialized data, enhanced model capabilities, and open interfaces for interaction with external systems [14][18]. - Examples include the integration of real-time data in travel agents like Fliggy, which improves the accuracy and timeliness of recommendations [16][17]. AI Browser Developments - OpenAI's ChatGPT Atlas browser allows users to interact with AI directly within the browser, providing personalized memory capabilities and the ability to perform tasks based on user input [22][26]. - This development signifies a strategic move to establish a core operating system for internet navigation, potentially replacing traditional browsers [27]. Computing Power Demand - The report highlights the increasing need for computing resources driven by advancements in AI, with Anthropic's collaboration with Google representing a significant investment in cloud computing [30][28]. - The introduction of the Doubao-Seed-3D-1.0 model by ByteDance is expected to further elevate computing demands by enabling high-quality 3D model generation from single images [33][31]. Companies to Watch - The report identifies several companies to monitor in the computing power sector, including Cambrian, Haiguang Information, and others, as well as AI application developers like Alibaba and Tencent [34].

Group 1 - DeepSeek has announced the update of its official App, web version, and mini-program to DeepSeek-V3.2-Exp, resulting in a significant reduction in API costs by over 50% for developers [1] - The cost of AI inference computing power has been decreasing due to advancements in AI large models and improvements in the performance and cost-effectiveness of inference chips, with hardware costs dropping approximately 30% annually and energy efficiency improving by about 40% [1] - The continuous decline in costs for large models, represented by DeepSeek, supports the commercialization of AI applications and enhances the efficiency of distilled models [1] Group 2 - The rapid iteration of large models and enhanced inference capabilities are creating opportunities for customized Agent applications, allowing users to tailor agents based on personal data and needs [2] - Companies like Cambricon and Huawei Ascend have announced their compatibility with DeepSeek-V3.2-Exp and have open-sourced the vLLM-MLU inference engine [2] - Companies such as Fanwei Network, Kingsoft Office, and Dingjie Smart are involved in the development of Agent and AI applications [3] Group 3 - Huawei Ascend has achieved software and hardware adaptation with companies like Softcom and Changshan Beiming [4] - Jiuqi Software plans to upgrade its Nüwa GPT in early 2025, integrating deeply with mainstream large models and launching various intelligent applications [4] - Jiuqi Software's AI distillation technology is similar to that of DeepSeek, indicating a trend in the industry towards efficient model optimization [4]

Investment Rating - The report maintains an "Accumulate" rating for the industry [3] Core Insights - The emergence of ChatGPT Pulse marks a significant shift towards proactive and personalized AI interactions, transforming ChatGPT from a passive responder to an active assistant [1][18] - Alibaba Cloud is positioning itself as a full-stack AI service provider, aiming to develop a custom agent platform that supports the evolution towards Artificial Super Intelligence (ASI) [2][19] - The report highlights the potential for custom agents to serve as personal work and life assistants, particularly in the financial sector, enhancing efficiency and decision-making [6][30] Summary by Sections ChatGPT Pulse Release - ChatGPT Pulse, released on September 25, 2025, acts as a super personal assistant, providing users with tailored content based on their interactions and preferences [1][14] - Users can manage the research content and connect their Gmail and Google Calendar for more contextual updates, enhancing the personalization of the service [14][18] Alibaba Cloud's ASI Development Path - Alibaba's strategy involves three phases towards ASI: emergence of intelligence, autonomous action, and self-iteration [19][20] - The company is investing heavily in AI infrastructure, with a plan to increase global data center energy consumption by ten times by 2032 compared to 2022 [24] Custom Agent Outlook - The report outlines various applications for custom agents in the financial sector, including personal research knowledge bases, intelligent mass messaging assistants, and personalized work assistants [6][32] - The potential for users to create and monetize their custom agents through platforms like MuleRun is also discussed, indicating a shift towards AI as a tradable digital asset [43][30] Industry Trends - The report notes a significant growth trend in the computer industry, with projections indicating a 70% increase from September 2024 to September 2025 [4] - The rapid iteration of large models and their increasing reasoning and tool usage capabilities are expected to drive the adoption of custom agents [6][30]