传媒行业周报系列 [Table_Title2]2025 年第 25 周 ► 市场行情回顾 [Table_Summary] 2025 年第 25 周（实际交易日为 2025.6.23-2025.6.27）上证 指数上涨 1.91%，沪深 300 指数上涨 1.95%，创业板指数上涨 5.69%。恒生指数上涨 3.89%，恒生互联网指数上涨 3.2%，行 业落后恒生指数 0.69pct。SW 传媒指数上涨 3.06%，落后创业 板指数 2.63pct，在申万一级 31 个行业涨跌幅排名中位列第 18 位。子行业中游戏、出版和体育位列前三位，分别为上涨 9.25%、上涨 4.17%和上涨 3.77%。 ► 核心观点&投资建议 证券研究报告|行业研究周报 [Table_Date] 2025 年 6 月 29 日 [Table_Title] [Table_Title] 6 月共 158 款游戏获批，版号数量创近年新高 6 月版号数量创近年新高，共 158 款游戏获批。本周，国家 新闻出版署公布 6 月游戏审批版号名单，共 147 款国产网络 游戏、11 款进口网络游戏获批。国产版号来看，巨人网络 《无尽修仙》、电魂网络 ...

证券研究报告|行业研究周报 [Table_Date] 2025 年 06 月 29 日 [Table_Title] 周专题：7 月空冰洗内销排产好于出口，第三批消费品以旧 换新资金将下达 [Table_Title2] 家电行业周报（2025W17） [Table_Summary] ► 周专题：7 月空冰洗内销排产好于出口，第三批 消费品以旧换新资金将下达 据产业在线最新发布的三大白电排产报告显示，2025 年 7 月 三大白电排产总量同比下降2.6%。家用空调内销受终端市场 景气度影响，7 月排产较前期下调，7-9 月内销排产同比增 速分别为+8.1%、-19.2%、-7.8%（前值 7、8 月分别为 +22.8%、-9.2%）；出口受抢出口补库周期结束影响，7-9 月 出口排产同比增速分别为-16.3%、-8.3%、-16.6%（前值 7、 8 月分别为-16.9%、-6.6%）。 618 及国补折上折透支冰箱需求，美国关税政策波动影响订 单需求。7-9 月冰箱内销排产同比增速分别为+3.2 %、- 5.5%、-3.5%（前值 7、8 月分别为+7%、+1%）；7-9 月出口 排产同比增速分别为-4.9%、 ...

本周(6.21-6.27) 47 个城市商品房销售 523 万方，成交量环比+39%，连续三周上升，同比-14%；5 月新房售价环 比-0.2%，同比-4.1%，同比降幅持续收窄；结合量价，景气度下行趋缓。本周分能级来看：一线城市周环比+81%，周 同比+2%；二线城市周环比+42%，周同比-13%；三四线城市周环比+19%，周同比-22%。 本周(6.21-6.27) 22 个城市二手房成交 264 万方，成交量环比+1%，连续三周上升，同比-2%；5 月二手房售价环 比-0.5%，同比-6.3%，同比降幅持续收窄；结合量价，景气度底部企稳。本周分能级来看：一线城市周环比+2%，周同 比-10%；二线城市周环比+1%，周同比-4%；三四线城市周环比+1%，周同比+19%。 上海发布六批次土拍共推出 13 幅用地，其中徐汇衡复、静安东斯文里、普陀真如、虹口北外滩、浦东唐镇、闵行莘庄 六宗地块或具备较高的市场关注度。我们认为：①在上海新房市场热度较好的情况下，政府再接连推出好地块，释放支撑 楼市的积极信号，有利于热度持续释放。②7 月该批次土拍，若成功实现好地块+好项目的加成，有望点燃四季度至明年初 的楼市行情 ...

Core Insights - The AI search optimization industry is undergoing a significant transformation from traditional keyword ranking to control over answers, with increasing user trust and reliance on AI-generated content [1] - Companies need to leverage technology to capture traffic on AI dialogue platforms, making core technological capabilities a competitive focus [1] Company Analysis - YuanSuo AI Optimization is recognized as a technology leader, showcasing comprehensive advantages in core technology dimensions, including a self-developed AI answer optimization engine that integrates semantic analysis, brand embedding algorithms, and negative public relations modules [1] - The engine can accurately identify user intent and ensure natural brand association, with a 95% customer satisfaction rate, 85% renewal rate, and 52% market share, solidifying its leading position in the industry [1] - DeepQing AI Optimization focuses on medium to large enterprises, emphasizing standardized processes and cost-effectiveness, particularly in the manufacturing and B2B sectors, although specific technical parameters are not disclosed [2] - QiYuan AI Optimization has emerged in the brand incubation field, specializing in semantic fine-tuning in Chinese contexts, and offers practical strategies that leverage industry experience and behavioral data analysis [2] Competitive Landscape - A significant generational gap exists between industry leaders and followers, with YuanSuo AI Optimization maintaining a complete technological closed loop and continuous R&D investment, resulting in a customer delivery speed that is one-fourth of the industry average and project costs that are one-third of competitors [2] - The upcoming shift where AI search traffic is expected to surpass traditional search highlights the importance of mastering core technologies for service providers, making them essential partners for brands [2][3] Evaluation Criteria - Decision-makers evaluating AI search optimization solutions should focus on three dimensions: precision in semantic understanding and natural language generation, data security and compliance management, and cross-platform adaptability with continuous optimization mechanisms [3] - YuanSuo AI Optimization's balanced performance across these dimensions, along with a validated success case library, positions it as a highly reliable technical partner in the current market [3] - The rapid formation of competitive barriers due to technological disparities emphasizes the need for early partnerships with leading technology service providers to gain a first-mover advantage [3]

Group 1 - The summit in Shenzhen focused on the theme "No Intelligence Connection, No AI," highlighting the rapid digital transformation across various industries [1] - Huawei introduced the Star River AI high-quality 10G campus network solution and the "No Intelligence Connection, No AI" Nebula Plan, emphasizing the integration of AI with intelligent connectivity [1] - Huawei's Data Communication Product Line President Wang Lei stated that AI is becoming a key driver of social productivity, supported by robust data communication networks [1] Group 2 - Huawei plans to deepen its "Partner + Huawei" system, offering extensive support to partners through marketing activities, a joint marketing fund of 3 million, and over 40 model sites for partner training [2] - The "No Intelligence Connection, No AI" Nebula Plan 2025 was launched to promote deep integration of AI and intelligent connectivity, enhancing digital productivity [4] - Huawei has developed a reference architecture for future campus networks, focusing on high-quality 10G networks and improving user experience across various aspects [4] Group 3 - Collaborative projects include the Smart Cultural Tourism Complex at Xiaomeisha Resort, which upgraded its wireless network using Huawei's solutions, and Shenzhen Vocational Technical University, which implemented a full-scenario Wi-Fi 7 campus network [4] - The summit concluded with a visit to the Xiaomeisha Smart Cultural Tourism Complex, where attendees experienced the upgraded wireless network and observed the latest achievements in hotel smart management [6]

本周观点 投资建议 建议关注国内生成式大模型龙头科大讯飞；AI 硬件有望成为应用落地的新载体，建议关注萤石网络、虹软科技 等；AI 相关功能打磨能够带动 C 端应用月活量、付费率提升，建议关注金山办公、万兴科技等。 风险提示 6 月 21 日，月之暗面介绍了自主智能体 Kimi-Researcher，计划在未来几个月内开源其基础预训练模型及强化 学习模型。Kimi-Researcher 是一个自主的智能体与思维模型，使用并行的实时内部搜索工具等，通过端到端强 化学习进行训练。研究者在训练语料的构建上采取了两种互补的策略，用于解决高质量智能体数据集稀缺的问题， 并设计了上下文管理机制以保留关键信息。在测试中，Kimi-Researcher 取得 26.9%的 Pass@1 成绩，获得最新 SOTA，Pass@4 准确率也达到 40.17%。6 月 23 日，特斯拉正式推出 Robotaxi 服务，将在美国得克萨斯州奥斯 汀市试运营。目前，该服务仅限受邀的知名社交媒体博主和科技内容创作者。特斯拉投入 10-20 辆 Model Y 在 预设地理围栏内运营，全自动 Cybercab 车型预计 2026 年后推出。当 ...

"希望你们在党的兴边富民好政策引领下，切实维护好民族团结，创造更加幸福美好的生活，把雪域高原的自然美 景保护好，把'桃花村'品牌擦得更亮，为建设繁荣稳固的祖国边疆贡献力量。"近日，习近平总书记给西藏林芝市 巴宜区林芝镇嘎拉村全体村民回信。总书记深情回忆起4年前嘎拉村之行的美好时光，对村里近年来发生的新的可 喜变化倍感欣慰，对乡亲们提出殷切期望。总书记的回信，字里行间饱含对西藏发展的关心、对西藏各族人民生 活的关切，必将激励西藏各族干部群众团结一心、艰苦奋斗，奋力谱写雪域高原长治久安和高质量发展新篇章。 今年是西藏自治区成立60周年。60年来，在党的领导下，在全国人民的大力支持下，西藏现代化建设取得丰硕成 果。特别是党的十八大以来，习近平总书记两次出席中央西藏工作座谈会，亲赴西藏视察指导，确立了新时代党 的治藏方略，为西藏明确发展目标、部署战略任务、谋划大事要事，推动西藏各项事业取得全方位进步、历史性 成就。嘎拉村的美好生活，正是60年来西藏经济社会发展成就的一个缩影。在这个民族团结进步之花盛开的地 方，村民因地制宜大力发展乡村旅游，美好生活芝麻开花节节高。实践充分证明，党中央关于西藏工作的方针政 策是完全正确 ...

上半年渐近尾声，公募"中考"成绩单也陆续揭榜。从分红榜来看，公开数据显示，截至6月29日，年内公募基金分 红总额达1274.59亿元，较2024年上半年增长超三成。其中，债券型基金仍是分红"主力军"，分红总额占比超七 成。但若从分红榜前十名来看，则有七席由权益类ETF包揽，且前三名均为权益类ETF。其中，华泰柏瑞沪深 300ETF不仅以83.94亿元的分红总额登顶分红榜，更刷新ETF单次分红纪录。有业内人士指出，下半年公募基金分 红热情有望延续，分红总额或超越去年，但也需要注意外部环境变化、宏观经济形势波动等风险因素可能导致部 分基金分红规模不及预期。 | 序号 | 证券代码 | ﮯ | 证券简称 | ト | ト 草舍全称 | ▲ 区间分红总额 | | --- | --- | --- | --- | --- | --- | --- | | | | | | | | [起始交易日期] 2025-1-1 [截止交易日期] 2025-6-29 [单位] 亿元↓ | | 1 | 510300.OF | | 华泰柏瑞沪深300ETF | | 华泰柏瑞沪深300交易型开放式指数证券投资基金 | 83.9350 | | 2 ...

Financial Performance - Revenue for the year ended March 31, 2025, was HKD 1,314,101,000, representing a 82% increase from HKD 722,725,000 in the previous year[3] - Gross profit increased to HKD 103,804,000, up from HKD 83,829,000, reflecting a gross margin improvement[3] - Net profit for the year was HKD 94,919,000, compared to HKD 76,041,000 in the prior year, marking a 25% increase[4] - Basic earnings per share rose to HKD 0.95 from HKD 0.92, while diluted earnings per share increased to HKD 0.92 from HKD 0.75[4] - The company reported a total comprehensive income of HKD 90,163,000, significantly higher than HKD 49,168,000 in the previous year[4] - The company's gross profit from gold sales was HKD 1,143,142,000 in 2025, up from HKD 621,004,000 in 2024, reflecting an 84% increase[17] - The group's profit for the year was approximately HKD 95 million, compared to approximately HKD 76 million in the previous year, with net profit attributable to equity holders of approximately HKD 66.6 million, up from HKD 55.3 million last year[33] - The group recorded earnings per share of approximately HKD 0.95, compared to approximately HKD 0.92 in the previous year[34] Assets and Liabilities - Total assets amounted to HKD 1,699,156,000, up from HKD 1,628,067,000, indicating growth in the company's asset base[5] - The company's net assets rose to HKD 749,549,000 from HKD 506,651,000, indicating a strong equity position[5] - Total assets in the gold mining segment reached HKD 1,721,406,000 as of March 31, 2025, compared to HKD 1,403,581,000 in 2024, indicating a 22.6% growth[14][15] - The company’s total liabilities decreased to HKD 929,607,000 in 2025 from HKD 921,621,000 in 2024, indicating a stable financial position[15] - The accounts payable increased to HKD 591,351,000 as of March 31, 2025, compared to HKD 477,860,000 in the previous year, indicating a rise of approximately 23.7%[25] Cash Flow and Financing - The company's cash and bank balances decreased to HKD 29,481,000 from HKD 47,489,000, reflecting changes in liquidity[5] - The company’s financing costs totaled HKD 11,174,000 in 2025, a slight decrease from HKD 13,376,000 in 2024[14][15] - The total interest expenses for the year ended March 31, 2025, amounted to HKD 11,174,000, a decrease of approximately 16.4% from HKD 13,376,000 in the previous year[18] Revenue Sources and Customer Base - The company identified two major customers contributing over 10% of total revenue in 2025, compared to three in 2024, with total revenue from major customers reaching approximately HKD 858,000,000[16] - Other income increased significantly to HKD 20,125,000 in 2025 from HKD 993,000 in 2024, marking a substantial rise[17] Capital Expenditure and Investments - Capital expenditure for the gold mining segment was HKD 123,915,000 in 2025, compared to HKD 360,583,000 in 2024, showing a reduction in investment[14][15] - Total spending on mining development and exploration for the year was approximately HKD 117.9 million and HKD 2.4 million, respectively[53] Taxation and Compliance - The income tax expense for the year ended March 31, 2025, was HKD 20,503,000, compared to HKD 10,411,000 for the previous year, reflecting an increase of approximately 96.5%[20] - The audit committee reviewed the draft consolidated financial statements for the year ending March 31, 2025, ensuring compliance with accounting principles and internal controls[66] Future Plans and Strategic Initiatives - The company plans to continue its focus on gold exploration and mining, aiming for further market expansion and operational efficiency[6] - The group plans to accelerate gold mining and production capacity expansion, with significant increases in gold output expected following the completion of technical upgrades at its subsidiary, Taizhou Mining[50] - The group is actively exploring acquisition opportunities for multiple gold and polymetallic mines, aiming to complete several acquisitions within the next three to four years[50] - The group will continue its mineral exploration program, extending the exploration area from approximately 1,100 meters above sea level down to 400 meters, aiming to increase gold and other mineral resources and reserves[50] - The group has established a joint venture with a leading industrial AI developer in mainland China to develop AI-driven mining models, aiming to enhance exploration, extraction, and smelting efficiency[51] - The group aims to diversify its revenue sources through the commercialization of AI solutions across various industries[51] Corporate Governance - The group is committed to maintaining high standards of corporate governance to enhance shareholder value and protect shareholder rights[61]

Group 1 - Meta's CEO Mark Zuckerberg has personally compiled a list of top AI talent, referred to as "The List," to attract global AI experts, offering salaries up to $100 million to entice them to join his new lab focused on "superintelligence" [2][3] - The recruitment effort follows the launch of Meta's Llama 4 AI model, which reportedly did not meet Zuckerberg's expectations, prompting him to actively seek talent from competitors like OpenAI [3][4] - Meta has reached out to numerous researchers from elite institutions, many of whom have experience at leading AI companies such as OpenAI and Google DeepMind, indicating a strategic move to bolster its AI capabilities [3][5] Group 2 - Alexander Wang, the 28-year-old founder of Scale AI, is leading Zuckerberg's new team, and Meta recently acquired a stake in Scale AI for $14 billion, highlighting the company's commitment to enhancing its AI resources [6] - The competitive landscape in AI has intensified, with companies like OpenAI taking measures to protect their research, while researchers are increasingly opting for roles in tech labs over academia due to the resources available [8]