智通财经APP获悉，人工智能板块再迎催化。十四届全国人大常委会第十八次会议10月28日表决通过关 于修改网络安全法的决定，自2026年1月1日起施行。回应人工智能治理和促进发展的需要，修改后的网 络安全法规定，国家支持人工智能基础理论研究和算法等关键技术研发，推进训练数据资源、算力等基 础设施建设，完善人工智能伦理规范，加强风险监测评估和安全监管，促进人工智能应用和健康发展。 北京理工大学法学院教授洪延青认为，这将推动人工智能治理从局部监管转向系统性规制。 信息技术咨询公司Gartner的报告显示，中国网络安全的重点正逐步转向保障人工智能的安全，实现业 务转型并加强企业、机构的韧性。行业专家认为，AI技术应用在带来新型安全风险的同时，也为网络 安全行业提升效率、提高产品能力带来机遇。 多家网络安全上市公司正在积极投入AI技术研发。深信服在投资者互动平台表示，持续将人工智能技 术与能力融入网络安全和云计算各项产品与服务中，截至目前已将大模型技术应用到网络安全产品，形 成安全GPT、融合AI能力的AI防火墙和其他多款安全产品，并计划未来继续加大对AI技术的研发投 入。 天融信表示，公司已在AI+安全、智算云等新方 ...

中小企业的智能化跃升，是智能世界真正落地的"最后一公里" 对华为而言，这是一次生态建设的全面升级；对中小企业而言，则是一条能够跨越技术门槛的智能化新路径。 新趋势，老愿景 "智能世界正在加速到来，AI将给企业带来颠覆式变化。"在9月17日的华为坤灵秋季新品发布会上，华为常务董事汪涛用这样一句话开场。 全球范围内，AI已经成为企业增长的新引擎，但在普惠性和落地深度上，鸿沟正在迅速拉开。 麻省理工学院今年7月发布的《GenAI鸿沟：2025年AI商业状况》指出，大型企业正加速在核心业务中部署AI，已经有超过70%的头部公司将AI融入生产、 研发与客户服务，并在效率和营收增长上获得实质回报；相比之下，小企业的状况却不尽如人意——尝试引入AI的比例不到25%，能形成稳定应用并带来实 际收益的比例更是低于10%。 这背后蕴含着明显的时代转向：AI已经从效率提升工具，进一步进入企业的生产、服务和管理核心系统，推动商业模式、产品形态和工作模式全方位重构。 这一矛盾在中国则表现得更为明显。汪涛在发言中提到，全国中小企业数量超过5800万家，贡献了60%的GDP、70%的技术创新以及超过80%的城镇就业， 是国民经济的毛细血管 ...

Summary of the Conference Call for Anbotong Company Overview - **Company**: Anbotong - **Industry**: Cybersecurity and AI Solutions Key Points and Arguments Financial Performance - Anbotong reported a revenue of 429 million yuan in the first half of 2025, representing a year-on-year growth of 123.98% driven by strong performance in the new security AI product line [2][3] - The company expects to maintain an annual gross margin of around 60% for 2025, significantly improved due to the abandonment of low-margin, high-risk projects [21] Strategic Developments - Anbotong has shifted its strategy to focus on high-margin projects, particularly in the AI and cybersecurity sectors, abandoning low-margin government projects [4][13] - The company has established the Luban AI Research Institute and launched the Yicai AI interaction architecture to enhance its influence in the AI security field [3][17] Product Innovations - Anbotong has applied for 24 new patents, received 13 invention patents, and obtained 52 software registrations, indicating a strong commitment to R&D [2][3] - The company has launched several products based on domestic chips, including an all-in-one machine and a liquid-cooled desktop silent all-in-one machine [5][6] Collaborations and Partnerships - Anbotong is collaborating with domestic chip manufacturer Jiangyuan to develop AI computing centers, with ongoing projects in Hubei and plans for deployment in Shandong [5][9] - The company has signed long-term cooperation projects in Japan and is expanding its security business in Malaysia and Indonesia [9] Market Position and Future Outlook - Anbotong is positioning itself as a key player in the integration of AI and cybersecurity, with a focus on enhancing operational efficiency for clients [17][19] - The company is actively exploring opportunities in the low-altitude economy, focusing on computing and network scheduling rather than the aircraft itself [20] Shareholder Developments - The chairman's transfer of approximately 5.6% of shares, locked for 18 months, is viewed positively by the market as it brings in supportive business resources [7][8] Challenges and Risks - The company has faced delays in project approvals, particularly in Shandong, which may impact timelines for revenue generation from certain projects [5][14] - Anbotong's cybersecurity business has seen a decline year-on-year, attributed to the strategic decision to exit low-margin projects [13] International Expansion - Anbotong's Hong Kong branch is handling stablecoin RWA business, which is restricted in mainland China, indicating a strategic move to leverage international markets [9] Future Projections - While the company has achieved over 100% revenue growth in the first half of 2025, uncertainties regarding project execution may affect future guidance [22] Additional Important Information - The company is developing compliance and security detection AI agents to enhance service delivery in various sectors, including medical devices [18] - Anbotong's collaboration with major model manufacturers focuses on project applications rather than direct security needs, indicating a strategic alignment with industry leaders [19]

Core Insights - The report by Akamai highlights the significant impact of AI technology on internet security, particularly in the context of web and API attacks [1][3] Web and API Attacks - In 2024, global web attack incidents are projected to reach 311 billion, marking a 33% year-on-year increase [3] - In the Asia-Pacific region, web attacks surged by 73% to 51 billion [3] - API attacks are also on the rise, with a monthly growth rate of 32% for OWASP API Top 10 related attacks, and API risk losses expected to exceed $100 billion by 2026 [3][4] Security Risks and Compliance - Unauthorized or improper authorization is identified as a primary cause of sensitive data leaks and system penetrations [4] - The report outlines four major risks in API security: API abuse, decreased testing frequency, lack of pre-testing, and the presence of zombie and shadow APIs [4] - Compliance with OWASP and MITRE frameworks is becoming increasingly critical as API security incidents escalate, leading to heightened compliance risks for businesses [4] Industry-Specific Insights - Different industries face unique and common API security risks, with e-commerce experiencing heightened attack levels during shopping and promotional seasons, while the financial sector may face targeted attacks due to geopolitical events [4] - AI technology is being utilized by attackers for strategic target selection, attack automation, traffic-based attacks, and behavior-based attacks [4][5] Regional Attack Trends - In the Asia-Pacific region, the total number of web and API attacks increased by 73% to 51 billion, with the financial sector being a primary target, showing a growth rate exceeding 52% [5] - Singapore experienced severe DDoS attacks, with 47 trillion incidents in 2024, ranking it among the top globally [5] Security Strategies - Akamai proposes six security strategies to help businesses build a comprehensive defense system, including integrating security requirements into the development process and utilizing AI to counter AI attacks [5] - The company emphasizes the need for proactive defense strategies to mitigate API vulnerabilities and ransomware threats [5] Future Outlook - AI technology is expected to continue influencing API security protection models, with Akamai planning to release an AI firewall capable of defending against attacks targeting AI systems [5] - Companies are encouraged to proactively develop AI security capabilities and foster a collaborative model of shared responsibility among security vendors, cloud service providers, and users [5]

图片来源： Sequoia Capital Z Highlights Nikesh Arora ， Palo Alto Networks 的董事长兼 CEO ，曾在谷歌担任高级副总裁和首席商务官，还是软银的总裁兼首席运营官。本文是红杉资本在 DeepSeek 爆火的当下，邀请关 AI 时代的网络安全防护策略等关键话题展开深入探讨。 AI 发展探讨：模型风险与应用前景 Pat Grady ： 我们之前给您发送了电子邮件，询问您是否愿意参加我们的节目，您的回复是（我引用一下）： " 只要能谈论 DeepSeek 和新世界秩序，我 就参加 。 " 让我们从这里开始。 Nikesh Arora ： 我们都有自己对 AI 的理解，大家都在努力探索并将其合理化，形成某种思维框架。我和其他人一样，也有自己的看法。 我认为，从我的 角度来看，过去 12 个月里发生的一切都堪称现象级。 我们看到有人试图构建一个有效的大脑，它拥有巨大的能力来记住一切、处理一切，并进行模式识 别，这有点像我对 NLM 的理解。 现在，这个大脑因为接受了现有数据的训练，所以很容易根据用于训练的数据得出错误的结论，这已经不是秘密了。 我们在各种关于幻 ...