
AI-generated summary
By 2029, there will be more than 1 billion agentic artificial intelligence (AI) agents in use around the world — a bold prediction made by IT market intelligence giant International Data Corporation. The number is astonishing, as it not only represents a 40-fold increase compared to 2025 but also accounts for roughly one-eighth of the global population. Yet despite widespread curiosity about agentic AI — where machines evolve from passive tools into active assistants that can plan, coordinate and execute complex tasks — enterprise adoption remains slow.
In McKinsey's "The State of AI in 2025" report, 62% of respondents said their organisations were experimenting with AI agents, but only 23% had scaled these systems within the company. However, if there is any doubt about the transformative power of autonomous agents, the early movers are already proving otherwise. At the Amazon Web Services (AWS) Summit Hong Kong 2026, Chris So, Managing Director of AWS Hong Kong, attributed the company's success in achieving US$15 billion in annual recurring revenue in Q1 2026 to the growing adoption of AI agents.
From Keynotes to Demos: AI in Action
Now, the challenge lies in turning widespread interest into actual impact. For the thousands of business leaders, technology enthusiasts and policymakers gathered at the Hong Kong Convention and Exhibition Centre on 17 June, the annual summit charted a practical roadmap for building this future of human-machine collaboration. Across more than 60 expert-led sessions, hands-on workshops and customer success showcases, attendees were shown how agentic AI, data, security and cloud innovation will reshape the world. Beyond the keynotes on the main stage, the summit delivered a wealth of immersive experiences across the Expo. The Industry Zone featured interactive demonstrations across key sectors, including financial services, retail and consumer goods, allowing attendees to connect with AWS experts to address pressing challenges. The AWS Village showcased cutting-edge solutions from partners around the globe, while the Developer Community Zone gave developers a chance to experiment with Kiro, an agentic coding service that brings structure to coding chaos.
The AWS AI League Hong Kong turned the conference floor into the ultimate AI showdown, where competitors raced to solve real-world challenges through gamified tournaments. Top performers not only secured tickets to re:Invent 2026 in Las Vegas but also earned a place at the Global Championship. Meanwhile, the Training and Certification Zone welcomed AI newcomers with free on-site sessions and exam-readiness workshops that enabled attendees to build skills and earn certification vouchers.
Two Decades Young: AWS Celebrates a Day One Mindset This year's summit carried special significance as AWS celebrated its 20th anniversary. During his opening remarks, Chris So reminded the audience: "AWS is just 20 years old. We are still very young, so we are still on day one in the cloud and AI era." This "day one" mindset was proven not by words but by the innovations on display across all six layers of AWS's AI stack. From infrastructure and custom silicon to data platforms, model training, agent orchestration and ready-to-use applications, each served as a testament to 20 years of relentless commitment.
However, the most significant announcement of the day came from the deepening partnership between AWS and the Hong Kong Science and Technology Parks Corporation (HKSTP). The two organisations unveiled a three-year strategic evolution of their Joint Innovation Centre (JIC), a collaborative platform launched in 2023 to support Hong Kong's startup ecosystem as well as innovation and technology development.
Under the new agreement, AWS and HKSTP will support more than 300 AI-driven startups with a new focus on healthcare and life sciences — a sector where innovation can make a profound impact on people's lives. Talent development is another priority, with structured career pathways for Hong Kong students through certifications and industry-relevant training. On the global front, the collaboration aims to strengthen the startup bridge into the Greater Bay Area while granting access to worldwide markets.
The Vistra Story: When Executives Become Builders While incubating startups is one challenge, scaling AI across a 9,000-person multinational operation is quite another. The partnership between AWS and HKSTP laid the groundwork for tomorrow's innovators, but the summit also showcased how today's enterprises are harnessing AI at scale. Vistra, a leading fund administrator and corporate service provider, turned to AWS to rebuild its technology foundation for the AI era.
Damian Leach, Chief AI & Digital Officer at Vistra Group, described a business-driven approach to AI adoption following the company's US$6.5 billion merger. Faced with heavy compliance obligations serving 200,000 legal entities across 170 countries, Vistra partnered with AWS in phased moves: first establishing a cloud foundation for resilience and scale, then improving customer experience by launching its AI compliance advisor, Geni.
However, the real breakthrough came in phase three, when Vistra gave all employees access to AI tools, resulting in over 13,000 hours of enhanced productivity and more than 80,000 AI actions. Among them, an executive vice president with no technical background used Amazon Q to create a pricing remediation playbook agent, proving that powerful AI tools no longer require deep technical expertise. Damian Leach remarked, "She's a commercial leader who saw a revenue problem and solved it with a tool that was available to her. This was a mindset shift, and represented true grassroots innovation."
From Interest to Impact: Hong Kong's AI Decade Unfolds
The story from Vistra proved that transformation demands more than technology — it demands a willingness to change. The future belongs not to those with the most advanced AI models, but to those who are willing to rethink how they operate.
Hong Kong, backed by substantial government investment including a HK$3 billion AI subsidy scheme, is staking its claim as a leading AI hub, with AWS as a key enabler. With billions in cloud infrastructure being deployed and a new generation of AI talent being trained, the city's AI journey is just beginning.

尽管近期发生多起人工智能软件越权入侵事件引发美国焦虑,但中国政府与科技界普遍认为人工智能利大于弊,正通过开放模型权重推动技术发展与安全治理,习近平称开放模型是‘难得的历史性机遇’,同时警告需确保人工智能处于人类控制之下,中国企业与研究机构正加大前沿模型安全研究,安远AI报告显示相关论文发表量增长60%,国家安全部发布警告但将责任归于用户,智谱AI发布开源模型GLM-5.3并建立安全漏洞披露台账,主张开放可增强安全性,而美国方面则倾向于限制模型能力并寻求政府干预以减缓技术发展,中美在人工智能安全上的分歧反映出对技术控制权的根本性不同看法。
8月30日,成渝铁路改造工程取得重大进展,由中铁八局承建的全线最长桥梁——黄桷坪双线特大桥顺利合龙。桥梁位于重庆市九龙坡区,全长754米,集铁路运输、车站接驳、市政兼容等功能于一体,是成渝铁路改造工程全线控制性关键工程。

Apple and Huawei are set to launch new foldable smartphones next week, with Apple's debut expected on September 9 alongside CEO John Ternus's keynote. Analysts project the launches could revive a global smartphone market forecasted to decline 17% in 2026, with IDC's Nabila Popal predicting Apple will ship over 17 million foldable iPhones by 2027 and capture 40% of the foldable market.

中国消费者普遍反映AI客服难转人工、承诺不兑现等问题,国家标准《顾客联络服务 人工与智能客户服务协同要求》将于9月1日起实施,要求智能客服优先处理标准化问题,人工客服处理复杂及高风险事务,并设置清晰入口和信息同步,旨在规范人机协作并为监管提供依据。

今年以来,多个国产大模型发布后迅速获得国产算力适配支持,如智谱匿名模型在OpenRouter测试中走红及月之暗面Kimi K3开源模型参数达2.8万亿。然而,芯片产业周期以年计算与大模型月度迭代的时间差成为核心挑战。专家指出,国产算力链的真正护城河在于数据主权与部署灵活性带来的成本、数据、供应链三重可控,而非仅仅成本优势。下游应用中,模型企业与算力厂商积极合作推动软硬件协同迭代,智能体兴起和词元经济形成为国产算力提供结构性机遇。

LandSpace's Zhuque-3 Y2 rocket successfully launched from Dongfeng Commercial Space Innovation Test Area on August 19, deployed its payload, and achieved China's first land-based recovery of an orbital-class rocket booster using deployable landing legs in Gansu province, marking a milestone previously achieved only by SpaceX.