Report Industry Investment Rating No information provided regarding the report industry investment rating. Core Viewpoints of the Report - The newly - introduced "Interim Measures for the Regulatory Rating of Wealth Management Companies" aims to implement differentiated supervision on wealth management companies of different risk levels, guiding the industry towards standardized transformation and stable development [3][8]. - The implementation of the regulatory rating will accelerate the differentiation of the industry pattern, forcing institutions to return to the origin of net - value operation, optimize business models, and promote the unification and standardization of the large asset management industry's regulatory standards [3][37]. - The wealth management market has experienced a cycle of "transformation pain, scale expansion, short - term callback, and stable growth". The proportion of fixed - income products has increased, and the industry is facing challenges in product diversification [3][25][32]. Summary by Directory I. Core Content of the "Measures" and the Evolution of Wealth Management Industry Supervision (1) Core Framework: Six - Dimension Scoring and Differentiated Graded Supervision - The "Interim Measures for the Regulatory Rating of Wealth Management Companies" constructs a rating system focusing on risk management and asset management capabilities. As of December 2025, 32 wealth management companies' outstanding wealth management products accounted for 92% of the market [8]. - The rating system includes six elements: corporate governance, asset management ability, risk management, information disclosure, investor rights protection, and information technology. The weights of asset management ability and risk management are both 25%, with a total of 50% [9]. - The regulatory rating uses both positive incentives and negative constraints. Higher - rated companies receive more support and are encouraged to innovate, while lower - rated companies face stricter supervision [12]. (2) Regulatory System Iteration: From Framework Building to Refined Supervision - Since 2018, the wealth management regulatory system has taken the "New Asset Management Regulations" as the top - level design, forming a closed - loop regulatory framework from products to institutions and from business to entities [14]. - Key policies include the "New Asset Management Regulations" in 2018, the "Measures for the Supervision of Commercial Bank Wealth Management Business" in 2018, and the "Interim Measures for the Regulatory Rating of Wealth Management Companies" in 2026 [15][18][19]. (3) Wealth Management Rating: Inheriting Common Features and Implementing Precise Policies According to Industry Attributes - The regulatory rating of wealth management companies is highly consistent with the rating logic of industries such as trusts, insurance asset management, and banks, and is an important complement to the unified financial regulatory framework [21]. - Different industries' ratings have different orientations. For example, trust company ratings focus on risk resolution and transformation, insurance asset management ratings focus on asset - liability matching, and commercial bank ratings focus on core indicators such as capital, assets, and liquidity [22]. II. Re - shaping of the Wealth Management Market Pattern and Outlook on the Impact of Rating Policies (1) Market Pattern Evolution: Stable Scale Growth and Concentration towards Fixed - Income Products - The wealth management market has experienced a cycle of "transformation pain, scale expansion, short - term callback, and stable growth". The number and proportion of broken - net products of bank wealth management subsidiaries have fluctuated cyclically since 2018, reaching a peak in 2022 and then declining [25][26]. - The number of wealth management products decreased from 2018 to 2021 and then stabilized, while the market scale bottomed out in 2019 and then increased. As of March 17, 2026, the total market scale reached about 30.9 trillion yuan, with 47,000 products [26][29]. - Fixed - income products have become the dominant force in the wealth management market, with their proportion rising from about 60% in 2018 to about 97% in 2026. The industry is facing challenges in product diversification [32]. (2) Industry Pattern Outlook: Coexistence of Opportunities and Challenges, Promoting the Survival of the Fittest - The implementation of the regulatory rating will accelerate the differentiation of the industry pattern. Leading bank - affiliated wealth management companies are likely to obtain high - level ratings and gain more advantages, while small and medium - sized companies may face restrictions [37]. - The era of small - scale product marketing and valuation smoothing mechanisms is over. The rating system will force wealth management companies to return to the origin of net - value operation, optimize business models, and improve information disclosure [39]. - The rating will guide the optimal allocation of funds, standardize the order of the large asset management market, and promote the unification of regulatory standards among different asset management industries [40].

香港交易及結算所有限公司及香港聯合交易所有限公司對本公告的內容概不負責，對其準確性 或完整性亦不發表任何聲明，並明確表示，概不對因本公告全部或任何部份內容而產生或因倚 賴該等內容而引致的任何損失承擔任何責任。 GF SECURITIES CO., LTD. 廣發証券股份有限公司 （於中華人民共和國註冊成立的股份有限公司） （股份代號：1776） 证券代码：524714 证券简称：26 广发 Y2 广发证券股份有限公司 2026 年面向专业投资者公开发行永续次级债券（第二期） 票面利率公告 海外監管公告 本公告乃根據香港聯合交易所有限公司證券上市規則第13.10B條刊發。 根據中華人民共和國的有關法例規定，廣發証券股份有限公司（「本公司」）在深圳 證券交易所網站( http://www.szse.cn )刊發的《廣發証券股份有限公司2026年面向專 業投資者公開發行永續次級債券（第二期）票面利率公告》。茲載列如下，僅供參 閱。 承董事會命 廣發証券股份有限公司 林傳輝 董事長 中國，廣州 2026年3月19日 於本公告日期，本公司董事會成員包括執行董事林傳輝先生、秦力先生、孫曉燕 女士及肖雪生先生；非執行董事李 ...

香港交易及結算所有限公司及香港聯合交易所有限公司對本公告的內容概不負責，對其準確性 或完整性亦不發表任何聲明，並明確表示，概不對因本公告全部或任何部份內容而產生或因倚 賴該等內容而引致的任何損失承擔任何責任。 董事長 中國，廣州 2026年3月19日 於本公告日期，本公司董事會成員包括執行董事林傳輝先生、秦力先生、孫曉燕 女士及肖雪生先生；非執行董事李秀林先生、尚書志先生及郭敬誼先生；獨立非 執行董事梁碩玲女士、黎文靖先生、張闖先生及王大樹先生。 （股份代號：1776） 海外監管公告 本公告乃根據香港聯合交易所有限公司證券上市規則第13.10B條刊發。 根據中華人民共和國的有關法例規定，廣發証券股份有限公司（「本公司」）在深圳 證券交易所網站( http://www.szse.cn )刊發的《關於延長廣發証券股份有限公司2026 年面向專業投資者公開發行永續次級債券（第二期）簿記建檔時間的公告》。茲載 列如下，僅供參閱。 承董事會命 廣發証券股份有限公司 林傳輝 GF SECURITIES CO., LTD. 廣發証券股份有限公司 （於中華人民共和國註冊成立的股份有限公司） 关于延长广发证券股份有限公司 20 ...

证券代码：524714 证券简称：26 广发 Y2 发行人将按上述票面利率于 2026 年 3 月 20 日至 2026 年 3 月 23 日面向专 业机构投资者网下发行。具体认购方法请参考 2026 年 3 月 18 日刊登在深圳证券 交易所网站（http://www.szse.cn）、巨潮资讯网（http://www.cninfo.com.cn）上的 《广发证券股份有限公司 2026 年面向专业投资者公开发行永续次级债券（第二 期）发行公告》。 特此公告。 广发证券股份有限公司 2026 年面向专业投资者公开发行永续次级债券（第二期） 票面利率公告 本公司及董事会全体成员保证信息披露的内容真实、准确、完整，没有虚假 记载、误导性陈述或重大遗漏。 广发证券股份有限公司（以下简称"发行人"）面向专业机构投资者公开发 行面值总额不超过 200 亿元（含）的永续次级公司债券已获得中国证券监督管理 委员会证监许可〔2026〕41 号文同意注册。广发证券股份有限公司 2026 年面向 专业投资者公开发行永续次级债券（第二期）（以下简称"本期债券"）为前述注 册批复项下的第二期发行，计划发行规模不超过 50 亿元（含） ...

Group 1 - The core viewpoint of the report emphasizes that the securities industry is experiencing a recovery in valuation, driven by favorable industry conditions and potential policy reforms [4][5][34] - The report highlights that since the "924" market rally in 2024, the A-share securities index has significantly outperformed the Shanghai Composite Index, achieving a peak excess return of 35.1% [5][10] - The report identifies three main factors contributing to the underperformance of the securities sector since 2025: sustainability concerns regarding high growth expectations, the impact of refinancing on valuation, and pressures from the funding environment [15][19][24] Group 2 - The report forecasts a robust growth of approximately 10% in the securities sector's net profit for 2026, building on a high base from 2025 [34][37] - It discusses the shift in the securities business model towards wealth management and large investment banking, indicating a transition from traditional brokerage and proprietary trading to more stable revenue sources [40][41] - The report outlines the expected impact of policy reforms aimed at enhancing the role of securities firms as core intermediaries in the capital market, with a focus on increasing direct financing [46][49]

关于延长广发证券股份有限公司 根据《广发证券股份有限公司 2026 年面向专业投资者公开发行永续次级债 券（第二期）发行公告》，发行人和主承销商于 2026 年 3 月 19 日 15:00-18:00 以 簿记建档的方式向网下专业机构投资者进行利率询价。 考虑到簿记建档当日市场情况，经发行人和簿记管理人协商一致，决定延长 本期债券发行时间，将簿记建档结束时间由 2026 年 3 月 19 日 18:00 延长至 2026 年 3 月 19 日 19:00。 特此公告。 （以下无正文） 2026 年面向专业投资者公开发行永续次级债券（第二期） 簿记建档时间的公告 广发证券股份有限公司（以下简称"发行人"）面向专业投资者公开发行面 值总额不超过 200 亿元（含）永续次级公司债券已获得中国证券监督管理委员会 证监许可〔2026〕41 号文注册。广发证券股份有限公司 2026 年面向专业投资者 公开发行永续次级债券（第二期）（以下简称"本期债券"）为前述注册批复项下 的第二期发行，计划发行规模不超过 50 亿元（含）。 （本页无正文，为《关于延长广发证券股份有限公司 2026 年面向专业投资者公 开发行永续次级债 ...

Core Viewpoint - Nvidia has raised its revenue guidance for the Blackwell and Rubin series chips at the GTC conference, extending the forecast to 2027, indicating a significant increase in AI demand visibility [1] Group 1: Nvidia's Revenue Guidance - Nvidia expects the Blackwell and Rubin series chips to generate $1 trillion in revenue by 2027, an increase from the previous guidance of $500 billion by the end of 2026 for data center equipment [1] - The revenue guidance extension to 2027 reflects a clear improvement in AI demand visibility [1] Group 2: LPX Architecture and Chip Production - The LPX architecture is set to begin shipping in the second half of 2026, featuring the Groq 3 LPX rack with 256 LPU processors, 128GB on-chip SRAM, and 640TB/s expansion bandwidth [2] - The combination of LPX with the Vera Rubin platform is expected to enhance inference throughput/power ratio by 35 times [2] - LPU chips will be manufactured by Samsung, with rack shipments anticipated to start in the latter half of this year [2] Group 3: Copper Foil Market Dynamics - The demand for high-frequency and high-speed copper foil is increasing due to severe signal attenuation, leading to higher performance requirements for copper-clad laminate materials [3] - Major players in the copper foil market, such as Mitsui and Taiwanese companies, are negotiating price increases for ultra-thin copper foil used in AI servers, with an average price increase of about 15% expected [3] - Domestic manufacturers are likely to follow suit with price increases, benefiting from the tight supply-demand situation in high-end electronic circuit copper foil [3] Group 4: Investment Recommendations - Recommended stocks include Defu Technology (301511.SZ), which is well-positioned to benefit from copper foil price increases; Cuprum Copper Foil (301217.SZ), with extensive experience in electronic circuit copper foil; and Jiyuan Technology (688388.SH), which collaborates with CATL and has acquired Endatong for optical modules [4] - Other notable mentions are Nord Shares (600110.SH), leading in lithium battery 4.5-micron products, and Zhongyi Technology (301150.SZ) [4]

Group 1 - Nvidia showcased the Vera Rubin POD platform at GTC, focusing on enhancing competitiveness in cluster computing and inference capabilities for Agent applications [1] - The Vera Rubin POD consists of two types of racks: MGXNVL rack for core GPU computing tasks and MGXETL rack for collaborative processing through direct interconnects [1] - A single Vera Rubin 1152 SuperPOD is composed of 16 Vera Rubin NVL72 racks, 2 Vera CPU racks, 10 Groq 3 LPX racks, 2 BlueField-4 STX storage racks, and 10 Spectrum-6 SPX network racks, highlighting a heterogeneous collaborative system architecture [1] Group 2 - The Groq3 LPX rack accelerates decoding with 256 LPU processors, 128 GB on-chip SRAM, and a bandwidth of 640 TB/s, enhancing the performance of the Vera Rubin NVL72 and LPX combination [2] - Under conditions of 400 TPS per user, the combination of Vera Rubin NVL72 and LPX can achieve up to 35 times the TPS improvement per megawatt compared to NVIDIA GB200 NVL72, making it suitable for low-latency, interactive Agent applications [2] Group 3 - The Vera CPU rack integrates 256 Vera CPUs with a high-density liquid cooling design, supporting over 22,500 concurrent reinforcement learning or agent sandbox environments for testing and validating outputs from Vera Rubin NVL72 and LPX [3]

Investment Rating - The report maintains a "Buy" rating for several brokerage stocks, including Dongfang Securities, Guotai Junan, and CITIC Securities, among others [11][29]. Core Insights - The brokerage sector has experienced a price decline despite stable earnings growth, attributed to factors such as capital pressure, policy stability, and changes in investor risk preferences [2][3]. - The report suggests that the brokerage industry is transitioning from volatile growth to stable growth, with improved earnings stability and a favorable environment for strategic allocation [3][5]. Summary by Sections Investment Rating - The report lists specific stocks with target prices and maintains a "Buy" rating for Dongfang Securities, Guotai Junan, and CITIC Securities, among others, indicating strong potential for price recovery [11][29]. Reasons for Decline in Brokerage Stocks - The brokerage index has dropped by 8% this year, primarily due to capital pressure from significant net redemptions in core ETFs, policy measures that compress market elasticity, and concerns over the sustainability of earnings growth [2][4]. Earnings Growth Sustainability - Historical data shows that brokerage earnings have been volatile, but recent market capacity expansion and diversified business lines have led to reduced volatility in earnings, suggesting a shift towards more stable growth [3][4]. Current Position of Brokerages - Current valuations of brokerage stocks are low compared to historical averages, with the A-share brokerage index trading at a PB of 1.37x, indicating potential for valuation recovery [4][5]. Catalysts for Sector Turnaround - Factors such as reduced capital market disturbances, positive policy signals, and a shift in investor preferences towards lower volatility investments are expected to catalyze a recovery in the brokerage sector [5][6]. Stock Selection Criteria - The report recommends focusing on three main lines for stock selection: high-quality undervalued leading brokerages, mid-sized brokerages benefiting from regional economic advantages, and opportunities arising from regional mergers and acquisitions [6][11].

1. Report Industry Investment Rating No relevant content provided. 2. Core Viewpoints of the Report - In February 2026, OpenClaw topped the GitHub hot list, accelerating the layout of AI Agent. This report systematically analyzes its theoretical framework and implementation path in the fixed - income investment and research scenario [3]. - AI Agent leads fixed - income investment and research from "dialogue response" to "autonomous action". It is a goal - driven autonomous intelligent system that can solve the pain points of traditional large models [3]. - The mainstream development platforms of AI Agent are divided into three categories, and different platforms are suitable for different users and application scenarios [3]. - AI Agent is suitable for the pain points of the fixed - income industry and can achieve multi - scenario full - process automation in the future [3]. - Privatized deployment is the inevitable choice for financial institutions to implement AI Agent, with the optimal solution of "domestic open - source models + privatized deployment" [3]. - Currently, the implementation of AI technology in the fixed - income field still faces challenges such as data security and compliance, model hallucination, high business adaptation threshold, lack of interpretability, and computing power cost pressure [3]. 3. Summary According to the Directory 3.1 AI Agent's Definition, Knowledge System, and Architecture Classification - **Definition and Development Background**: AI Agent is an intelligent system based on LLM that can independently understand user goals, disassemble tasks, plan execution paths, and call tools to complete complex tasks. It emerged to solve the pain points of traditional large models, and its development was promoted by relevant research and the breakthrough of reasoning large models [9][10]. - **Difference from Traditional Q&A AI and Copilot**: AI Agent has high autonomy, can achieve the user - specified final goal, and can perform complex interactions. It is different from traditional Q&A AI and Copilot in terms of autonomy, goal, interaction mode, task - handling ability, and typical application scenarios [12]. - **Knowledge System and Architecture**: AI Agent is a closed - loop system. Its operation logic can be disassembled into five steps: environmental perception, planning and decision - making, action execution, feedback verification, and reflection and optimization [14][15]. - **Three Mainstream Architecture Paradigms**: There are three mainstream architecture paradigms: reactive, deliberative, and hybrid. Each has its own advantages, limitations, and typical application scenarios [18][22][24]. 3.2 AI Agent's Support Technologies - **Large Language Model (LLM) Base Technology**: LLM is the "brain" of AI Agent. Different tasks require different base models, and factors such as Token, Embedding, Temperature, Top P, and context window affect its performance [28][30]. - **Prompt Engineering**: It is the "language art" of communicating with LLM/Agent. There are principles for writing prompts, and there are also common prompt - writing skills in Agent development [31][32][35]. - **Retrieval - Augmented Generation (RAG) Technology**: RAG breaks the knowledge boundary of large language models. It has a two - stage process and is mainly applied in three fields, but it also faces some technical bottlenecks [38][40][47]. - **Tool Call and Plugin Development Technology**: Tool call enables AI Agent to perform specific operations. The tool - call process has four steps, and there are six types of commonly used tools in the fixed - income business scenario [48][50][52]. - **Memory System Construction Technology**: The memory system of AI Agent is divided into short - term memory and long - term memory, which is crucial for its coherent behavior, personalized service, and continuous learning [53]. - **Multi - Agent Collaboration Technology**: Multi - Agent collaboration allows multiple AI Agents to work together to complete more complex tasks, with advantages such as professional division of labor, reduced complexity, controllable and compliant processes, and strong traceability [55]. - **Workflow Orchestration Technology**: Workflow orchestration fixes the execution steps of Agent into a standardized process, with advantages such as strong stability, traceability, batch - processing ability, and low - code development [58][59]. 3.3 AI Agent's Mainstream Development Platforms and Implementation Paths - **Mainstream Development Platform Comparison Analysis**: The current AI Agent development platforms are divided into three categories: code development frameworks, low - code/no - code platforms, and open - source privatized platforms. Each has its own positioning, architecture characteristics, applicable scenarios, advantages, and limitations, and there are corresponding platform - selection suggestions [60][61]. - **Typical Implementation Process of AI Agent**: Taking the "fixed - income sentiment monitoring Agent" on the Coze platform as an example, the process includes defining the goal and boundary, writing the Agent's persona and prompts, configuring tools and plugins, configuring the knowledge base, orchestrating the workflow, configuring multi - Agents (optional), testing and tuning, and publishing and integrating [63][64][65]. - **Privatized Deployment Scheme of AI Agent**: Privatized deployment is the mainstream choice for financial institutions. There are light - weight and enterprise - level distributed privatized deployment schemes, each with its own applicable scenarios, architectures, components, and model - selection suggestions [66][69][70]. 3.4 AI Agent's Application, Challenges, and Outlook in the Fixed - Income Field - **Fixed - Income Business Pain Points and AI Agent's Adaptability**: The fixed - income business has pain points such as low investment and research efficiency, homogeneous customer service, lagging risk management, and high compliance costs. AI Agent can address these issues through automation, personalization, real - time monitoring, and standardization [72][73][74]. - **Fixed - Income Field Knowledge Base**: It includes regulatory compliance, macro and market basics, sub - category exclusive investment and research, risk management, investment and research strategies, and case and technical support [75]. - **AI Agent's Application Outlook in Fixed - Income Investment and Research Scenarios**: It can be used in automated data collection and cleaning, macro and interest rate research, credit research and individual bond analysis, and automated research report generation, which can significantly improve work efficiency and reduce risks [76][77][78]. - **AI Agent's Application Outlook in Intelligent Investment Advisory and Asset Allocation Scenarios**: It can be used in customer risk profiling and assessment and fixed - income asset allocation recommendation, with different architectures for different tasks [83][84]. - **Challenges and Outlook**: AI Agent faces challenges such as data security and compliance risks, model hallucination and accuracy issues, business adaptation and implementation thresholds, interpretability and audit requirements, and computing power and cost pressure. In the future, it will evolve into an important part of the fixed - income team, and human - in - the - loop mechanism will become the norm [85][86].