Following the introduction of Meta's personal AI assistant, Muse, the product has witnessed a meteoric rise in popularity. Its download growth has eclipsed the initial success of ChatGPT's mobile app, propelling a notable surge in Meta's stock price and marking what many consider to be Meta's 'ChatGPT Moment'. With user consent, Muse is capable of accessing information and independently performing tasks such as conducting searches, completing forms, and making consumer reservations. It operates seamlessly through a cloud-based virtual computer backend.

To mitigate potential privacy data risks, Meta has implemented a robust security system. However, vulnerabilities persist, and the product still grapples with limitations concerning third-party application permissions. While the explosive growth in Muse's popularity warrants long-term observation, it has already solidified Meta's foothold in the global AI competition landscape.

Currently, several domestic companies have also rolled out personal AI assistant products. Nevertheless, due to challenges such as permission barriers between applications, they have yet to achieve a significant scale effect. The triumph of Muse offers a valuable blueprint for domestic counterparts in related fields to emulate and learn from.

On September 22, during the 2026 Apsara Conference, AutoNavi, a subsidiary of Alibaba dedicated to the practical application of spatial intelligence, introduced a range of spatial intelligence products and solutions. These included the groundbreaking All-Space Unmanned System and AutoNavi Qianyu. Liu Zhenfei, Chairman of AutoNavi, articulated that the company's ambition is to equip AI with the ability to comprehend three-dimensional spaces, evaluate environmental dynamics, and execute tasks within the framework of temporal and spatial limitations. AutoNavi's spatial intelligence services are designed to enhance travel experiences and intelligent assistance in daily life, providing a significant boost to burgeoning sectors such as embodied intelligence and the low-altitude economy.

On September 20, a closed-door think tank session titled "AI·Insight" took place alongside the "Creator Program" award ceremony at Tencent's Beijing headquarters. This event was co-hosted by Tencent News and WorkBuddy. During the session, three prominent economists—Li Daokui, Xia Chun, and Li Daxiao—engaged in in-depth discussions with financial content creators and the WorkBuddy product team. The topics spanned economic research, the nuances of financial report analysis, the intricacies of investment decision-making, and the role of AI in facilitating content creation.

On September 23, James Reilly, an analyst with Capital Economics, highlighted in a report that the connection between stocks related to artificial intelligence (AI) and bond yields could grow even stronger, particularly if major tech firms opt to raise more debt. Reilly observed that AI-related sectors tend to excel when oil prices decline, a trend that underscores their growing responsiveness to fluctuations in bond yields.

On the afternoon of September 20th, Xidian University and Shenzhen Intellifusion Technologies Co., Ltd. (hereinafter referred to as "Intellifusion") collaboratively organized the inauguration ceremony for the "Xidian University-Intellifusion Joint Laboratory for Embodied AI". The event was graced by the presence of President Gao Xinbo, Cai Gushun, a standing committee member of the Party Committee, along with Chen Ning, the Chairman and CEO of Intellifusion, and Li Aijun, the Chief Technology Officer (CTO), all of whom participated in the ceremony.

On September 23, Li Bingjun, Deputy Secretary of the Guizhou Provincial Party Committee and concurrently serving as the Governor of Guizhou Province, chaired a symposium with entrepreneurs in the field of artificial intelligence (AI). During the symposium, Governor Li Bingjun emphasized that Guizhou is set to implement more vigorous support policies. These policies will harness the province's stable power supply, leading-edge computing resources, and a wide array of application scenarios to invigorate corporate innovation. Guizhou aims to solidify its computational capabilities, foster stronger synergy between computing power and electricity, and enable intelligent computing centers to maximize their potential, thereby laying a solid foundation for the advancement of artificial intelligence. Simultaneously, Guizhou is committed to accelerating the opening up of application scenarios and public data resources. The province will dynamically update the list of demands, enhance the linkage between supply and demand, and create exemplary application scenarios in key industries, as well as in sectors such as education and healthcare. This approach is designed to tackle systemic and common challenges, thereby nurturing new drivers of growth. Furthermore, Guizhou will bolster talent service guarantees, providing robust support for enterprises in attracting and effectively utilizing top-tier talent.

Recently, the Secretariat of the National Cybersecurity Standardization Technical Committee unveiled the roster of approved cybersecurity standardization research projects slated for 2026. Notably, the project titled “Research on Standardization of Embodied Intelligence Human-Machine Collaborative Safety,” led by Professor Xia Yingjie from the Institute of Microelectronics at Hangzhou Dianzi University, has successfully secured approval for commencement.

In recent years, the swift advancement of smart cities, intelligent security systems, traffic management solutions, and autonomous driving technologies has underscored the importance of pedestrian perception and localization in complex scenarios as pivotal technologies for intelligent vision systems. Multi-view pedestrian detection leverages information gathered from cameras positioned at various locations. This approach effectively mitigates the challenges associated with limited fields of view and target occlusion that plague single-view systems. Moreover, it offers more comprehensive spatial data, facilitating precise pedestrian localization in intricate environments.

During the United Nations General Assembly in New York, AI giant OpenAI announced that it would provide Ukraine with free access to its latest AI cyber defense system, Daybreak, to help protect critical civilian infrastructure such as hospitals and power grids from cyber attacks.

Qualcomm announced that it will acquire unstructured robotics company PickNik, aiming to drive the development of open robotics and physical artificial intelligence.

On September 23 local time, OpenAI CEO Sam Altman called for the establishment of global AI standards at the United Nations General Assembly, emphasizing the importance of international cooperation for AI safety. He advocated for the development of unified technical standards to ensure safe and controllable AI development, and suggested drawing on the work of existing institutions such as the U.S. Department of Commerce's AI Standards and Innovation Center. Altman positioned himself as a pragmatic centrist, arguing that while advancing alignment research, common standards should guide global AI development. His speech came as the United Nations General Assembly discussed the risks of AI runaway, following the resignation of an Anthropic researcher over concerns about AI safety, which sparked broad industry discussion on the pace of AI development.

The Beijing Economic-Technological Development Zone has successfully attracted more than 600 core enterprises within the AI industry chain, achieving an industrial scale that surpasses RMB 80 billion. It has already secured early competitive advantages in areas such as AI glasses, wearable technology, and consumer electronics. Looking ahead, the development zone plans to capitalize on its robust industrial foundation, with manufacturing contributing 60% to its economy, to pursue the ambitious goal of becoming a global epicenter for defining AI consumer products and a center of excellence for intelligent manufacturing. To achieve this vision, the zone will systematically undertake four pivotal tasks to establish a comprehensive and self-sustaining ecosystem for AI consumer product industries, fostering sustained growth in advanced manufacturing sectors.

Companies are shifting some of their traditional software spending towards AI service providers like Anthropic and OpenAI. A report titled 'Where Exactly is the AI Budget Going' released by procurement software provider Zip provides supporting evidence. The report shows that in the 12 months ending in August, AI service providers such as OpenAI, Anthropic, Cursor, and Sierra accounted for 8% of total spending among Zip's customers, compared to just 1.4% in the previous 12 months. During the same period, the median total budget for enterprise software increased by 13%.

Recently, cybersecurity experts have unearthed a peculiar and novel form of Windows malware, dubbed CLOSEDQUORUM. Distinct from prior attack utilities that employed AI merely to aid in code composition or craft phishing material, CLOSEDQUORUM seamlessly integrates several sophisticated large language models into its operational framework. This integration empowers the malware to autonomously determine its subsequent moves while in action, heralding a fresh era in the evolution of AI-driven malware.

As the macOS 17 beta version continues to evolve, Apple is intensifying its efforts in the realm of artificial intelligence (AI) deployment. A closer examination of the system firmware indicates that the Apple Intelligence suite within macOS 17 is set to undergo significant enhancements and expansions. Notably, the storage space occupied by AI models and their associated resource files is projected to surge to around 30GB on some Mac devices.

On September 23, Nanwei Software released an announcement addressing unusual fluctuations in its stock price. The announcement highlighted that from September 21 to 23, the cumulative deviation in the closing price of the company's stock over three consecutive trading days surpassed 20%, signaling abnormal trading volatility. In recent times, AI applications have become a focal point of market interest. It's noteworthy that Nanwei Software's business operations concerning the Government Service AI Agent and WellWork AI Agent Platform are still in their infancy in terms of market development. As of the first half of 2026, these ventures have yet to yield any revenue. Consequently, future revenue streams remain uncertain, and the current impact on the company's overall performance is negligible.

Following its launch, Meta's AI agent, Muse, swiftly garnered widespread attention, propelling a notable uptick in Meta's stock value and providing a significant lift to the semiconductor industry. Muse is primarily designed to assist users in cutting costs, making it more accessible through integration with WhatsApp. It utilizes a specialized secure virtual machine architecture, clearly delineates permission boundaries, and effectively alleviates user apprehensions. Prior to this, China had seen similar attempts at developing personal AI agents, yet these endeavors failed to captivate users and investors alike. At present, despite high market expectations for Muse, it still grapples with challenges including user trust, limitations on platform access, and ambiguous paths to monetization. The long-term user retention rate is yet to be determined and closely watched.

On September 23, during the 2026 Apsara Conference, Lingyang, in partnership with Feihe, Yingfeng Digital Technology, and Yibao Group, collectively introduced the '7-Day AI Employee Deployment Initiative'. This initiative is specifically designed to meet the real-world business requirements of enterprises. Under this program, the Lingyang FDE team collaborates closely with the enterprise to pinpoint suitable AI application scenarios, ensuring the swift deployment of AI employees within a mere 7 days. Additionally, the initiative offers comprehensive support, including a complimentary one-day senior FDE scenario assessment, an initial credit package worth 150,000 Credits, access to AI platform tools, and enrollment in a hands-on training camp. During the conference, 27 enterprises expressed their interest by signing letters of intent to participate in the initiative.

On September 23, SenseTime ShanHui, a smart retail ecosystem enterprise within the SenseTime Group, officially launched its groundbreaking retail physical operating system, SenseMart OS, in Shanghai. Leveraging robotic convenience stores as its primary platform—or "carrier," as it's known in technical terms—the system seamlessly integrates with upstream robotic ontology (referring to the core structure or 'main unit' of the robotic system) enterprises, as well as downstream unmanned retail operators and franchisees. To date, SenseMart OS has successfully facilitated the opening of over twenty stores across various locations nationwide.

During the grand launch event, SenseTime ShanHui also announced a strategic collaboration with key ecosystem partners, including COFCO Smart Fashion and UBOX, to jointly initiate the ambitious 'Thousand Cities, Ten Thousand Stores Plan.' This initiative aims to further expand the reach and impact of SenseMart OS, fostering a new era of smart retail across China.

On September 23, the Standing Committee of the Beijing Municipal Party Committee convened a pivotal meeting, underscoring the imperative to propel the manufacturing industry forward through technological innovation. The focus was on bolstering original innovation and achieving breakthroughs in pivotal core technologies, thereby facilitating the seamless conversion of innovative outcomes. The meeting put forward the implementation of the 'AI Plus' initiative, aimed at accelerating the digital transformation within the manufacturing industry. This includes fostering the integration of advanced manufacturing with modern service sectors, nurturing service-oriented manufacturing, and empowering enterprises to venture into novel business formats and models. Additionally, there was an emphasis on expediting the innovation and deployment of green technologies and equipment, thereby advancing energy conservation, carbon reduction, and pollution mitigation efforts in the manufacturing realm.
Simultaneously, the meeting expressed support for leading manufacturing enterprises to assume a 'chain leader' role, fostering the growth of unicorn companies and specialized, sophisticated, and innovative 'little giant' enterprises. It also advocated for the swift establishment of a world-class science and technology park in Zhongguancun, orchestrating the interconnected development of the 'Three Cities and One District,' and refining policy mechanisms, including financial services and intellectual property protection, to cultivate a conducive development ecosystem.