Tech Weekly | Scalpers' markup margins shrink for iPhone Duo; celebrities also struggle to hold AI "face-stealing" livestream sellers accountable

Scalpers' markup room on the iPhone Duo shrinks

On October 23, Apple's first foldable-screen phone, the iPhone Duo, went on sale. The markup on this phone, which had previously been listed on e-commerce platforms, could reach 4,000 yuan. However, scalpers buying stock at offline stores in Shenzhen told Yicai that the actual markup on the day of release had shrunk noticeably. Around 9 a.m., according to scalpers' quotes, the 512GB and 1TB storage versions of the iPhone Duo both carried a markup of about 1,500 yuan, while the 2TB version carried a markup of only a few hundred yuan.

[Comment] Behind the shrinking markup, according to scalpers' feedback, the reason is that there is already a lot of stock on the market, or that after restocking too much they dare not buy at too high a price. While scalpers are not short of stock, Apple's official channels are out of stock of the iPhone Duo. On October 23, selecting iPhone Duo models with different storage capacities on Apple's official website showed an estimated shipping date of January 22. Shortly after iPhone Duo pre-orders opened, Tmall's Apple official flagship store already showed an estimated shipping time pushed to January 23 of next year.

Embodied intelligence financing cools: investors are no longer "head over heels"

Recently, industry insiders reported that some parties related to listed companies are no longer looking at the robotics sector; investment institution sources also mentioned that many RMB funds are now choosing to wait and see, unwilling to bet easily. According to an incomplete tally by Yicai, at least 185 financing events occurred in the embodied intelligence sector in the third quarter of this year, with financing amounts reaching 33.335 billion yuan. IT Juzi data shows that total financing in China's embodied intelligence track reached 93.5 billion yuan in the first half of this year. By that calculation, the single-quarter financing level in the first half was 46.75 billion yuan. Industry insiders noted that there is a certain lag in companies' release of financing information, which often comes later than the time the financing was finalized. This means that data more accurately reflecting current market reactions will require further waiting.

[Comment] In the first half of this year, financing in the embodied intelligence track was hot, and the technological and commercialization progress of some companies lagged behind their valuation levels. Entering the second half, the industry has reached an important watershed. Multiple industry insiders believe this cooling does not mean the market is over; investment funds are shifting from "casting a wide net" to "precision investing."

On October 18, actress Wu Yue posted on social media that someone had used AI to impersonate her in livestream selling and demanded that the relevant accounts return the money defrauded. Screenshots posted by Wu Yue showed that an account named "Sister Hua's Skincare Diary" used promotional language such as "celebrity-endorsed, recommended by Wu Yue," claiming that Wu Yue had teamed up with manufacturers to give skincare products to fans, and the voice in the video was also disguised to sound like Wu Yue herself.

Former "Si Le Me" app announces pricing, up to 199 yuan/year

Recently, the app "Zai Me Zai Me" (formerly "Si Le Me"), which sparked huge controversy because of its original name, officially announced its pricing standards, drawing widespread attention. On October 22, the app's founder, Lü Gongchen, said that "Zai Me Zai Me" focuses on serving elderly people living alone. The software system is divided into versions 1.0 and 2.0: "The 1.0 system is 99 yuan/year, and the 2.0 system is 199 yuan/year. They are centrally procured by the government, and elderly people living alone can install and use them for free."

Lü Gongchen introduced that the app is currently being promoted in the Zhejiang area, and next they are also trying to further promote it in Guangdong, Shandong, Gansu and other places.

[Comment] At the beginning of this year, an app named "Si Le Me" went viral thanks to daily check-ins, overdue reminders and its highly provocative name. In May, the app was renamed "Zai Me Zai Me" and landed in Hangzhou, shifting from a consumer-facing tool for young people living alone to a safety service for elderly people living alone. Over the past six months, the company has been seeking to transform from a traffic product into government and enterprise services.

Housekeeping companies rush to stake out the embodied intelligence data "business"

On October 23, at the 2026 China Housekeeping Industry High-Quality Development Summit, Tangdun Shijia, together with Qiongche Intelligent and Dahuan Robotics, jointly initiated the establishment of the Tangdun Shijia Home Service Embodied Intelligence Data Collection Laboratory. The lab focuses on real scenarios, real work and real data to accumulate data and practical evidence for the research and development of housekeeping robots. In this cooperation, housekeeping company Tangdun Shijia mainly provides scenarios for home data collection, the hardware equipment for collection is provided by Dahuan Robotics, and data processing and training are completed by Qiongche Intelligent.

[Comment] Over the past two years, many robotics companies have chosen to cooperate with housekeeping companies or platforms that provide housekeeping services. The "evolution" of embodied intelligence large models cannot be separated from training data. Acquiring real home-scenario data is not easy, and privacy is a major issue. According to Chen Zhe, founder and CEO of Tangdun Shijia, the first batch of households that agreed to have data collected in their homes numbered 100, and they will carry out desensitized, batch processing of data on scenarios such as bedrooms and bathrooms as well as photos of personal items.

On October 20, at the finals of the Feishu Doubao Work AI Pioneer Future Talent Competition, 12 groups of contestants focused on real business problems of enterprises, using Feishu and Doubao Work to complete research, collaboration and delivery of results. Feishu CEO Xie Xin said that Feishu connects, on one end, the advanced enterprises that students aspire to, understanding enterprises' demand for AI talent; on the other end, it also sees that college students still face uncertainty in career choices in the AI era. The competition hopes to bridge this talent mismatch and let students and enterprises meet in real business.

[Comment] The competition centers on real enterprise problems and delivery of results, shifting the use of AI tools toward solving real problems and landing business outcomes. Compared with traditional resume screening, it better observes students' collaboration, boundary awareness and ability to take responsibility for results. What enterprises need is not people who can use AI to write PPTs, but people who can use AI to solve real problems and deliver results.

On October 24, according to a relevant market report, 53.1% of advertisers will use AIGC technology in creative content generation. The other side of efficiency gains is escalating risk: black and gray industries use AI to mass-generate highly realistic materials and forge authoritative endorsements, with risks distributed across the entire process of "seeing an ad, clicking a link, completing a transaction," making manual review alone increasingly difficult to cope with. To this end, the Blue Book on Internet Advertising Governance in the AIGC Era systematically presents the practical approach of "using AI to govern AI."

[Comment] Technological iteration runs ahead of institutions, but governance is not about restraining innovation; it is about building a solid foundation of trust for AIGC advertising. AI cannot create a responsibility vacuum. Advertisers bear primary responsibility for authenticity and cannot use "content generated by AI" as an exemption, while platforms bear gatekeeping obligations commensurate with their technical capabilities.

Leading smartphone manufacturers step up full-terminal esports ecosystem

On October 21, the Honor WIN2 series was officially released, with the WIN2 and WIN RT2 going on simultaneous first sale, starting at 3,149 yuan after national subsidies. Against the industry backdrop of continuous consolidation of vertical esports brands, leading manufacturers have begun shifting toward ecosystem-based layouts. In less than a year, the Honor WIN esports matrix has completed the layout of three core terminals—smartphones, tablets and gaming laptops—while also supporting peripherals such as WIN earphones and game controllers, building a complete esports ecosystem.

[Comment] The overall smartphone market has entered stock competition, pure vertical esports brands continue to contract, and the sector's voice is shifting toward leading manufacturers. The core purpose of major companies incubating esports-oriented sub-brands is to capture highly sticky young users and avoid price competition with their main brands. But the competitive logic of the smartphone sector has already changed: the dividend from simply stacking hardware is fading, and whether a brand can enable multi-device collaboration and form a stable user circle will determine whether its esports product line can stand firm over the long term.

ERP vendors promote large-scale implementation of enterprise Agents

On October 23, Kingdee iterated its Lingji enterprise intelligent agent operating system at the Global Visionaries Conference, launching Lingji Manage, enterprise context and ERP connectors. Xu Shaochun, chairman of the board and CEO of Kingdee Group, said that enterprise AI transformation is first a cognitive transformation, then a management change, and finally a technological revolution. He believes that technology can be purchased, computing power can be rented, and models can be called, but only cognition must be traversed by oneself. When enterprise leaders truly complete the cognitive shift, AI transformation will no longer be a technical project but the beginning of organizational evolution.

[Comment] After intelligent agents gradually acquire business execution capabilities, overseas markets once worried that the value of traditional SaaS would be diluted, triggering short-term fluctuations in the software sector. Currently, domestic software vendors generally rely on business foundations to carry Agent capabilities rather than replacing existing systems with large models. But large-scale implementation faces prominent challenges: divergent industry business rules and high costs of transforming heterogeneous systems, while in strongly regulated scenarios such as finance, risks such as AI hallucinations, permission boundaries and audit traceability continue to hinder intelligent agents from penetrating into core business operations.

The He Xiangjian Foundation Young Scientist Program will add intelligent science as a new funding direction

On October 24, the 2026 He Xiangjian Young Scientist announcement ceremony was held at the Midea Global Innovation Center in Shunde, Foshan. As with the first session last year, this year's second session funded 20 young scientists in the two major fields of medicine and life sciences, and energy, environment and climate change, each receiving 2 million yuan in research funding disbursed over three years. He Xiangjian, founder of Midea Group, established a science fund with 3 billion yuan of his own money in 2023. The fund provides grants to selected young scientist projects to support original and breakthrough basic, translational and applied research. Starting from the third session, the program will add intelligent science as a new funding direction.

[Comment] Compared with other domestic science funds, the He Xiangjian Science Fund sets no assessment indicators and no proportion limits on the use of funds for the young scientist projects it supports, enabling them to dare to tackle more basic, longer-term and potentially failure-prone high-difficulty topics. Compared with the first session, the selected group in this session is overall younger, with more outstanding female scientists emerging; while encouraging frontier exploration, the review orientation further strengthens application translation and pays more attention to the potential of research to promote the landing of results. Starting from the third session, the program will add intelligent science as a new funding direction.