AI concept stocks experienced a collective surge! On December 30th, in the A-share market, AI concept stocks saw significant unusual activity. Kute Smart surged vertically during the session to hit the 20% daily limit-up, while Baina Qiancheng also briefly touched the 20% limit-up. As of the time of writing, more than ten stocks related to areas like KNOWLEDGE ATLAS AI, AI Agents, and AIGC had hit their daily limit-ups.

News that Meta is acquiring the AI startup Manus for billions of dollars provided a boost to AI concept stocks. It is reported that Manus focuses on building general-purpose AI Agents. Its founder, Xiao Hong, is from Ji'an, Jiangxi, and graduated from Huazhong University of Science and Technology with a degree in Software Engineering. Notably, Manus's growth rate is astonishing; just eight months after its official launch, its Annual Recurring Revenue (ARR) has already surpassed $100 million.

AI concept stocks strengthened. On Tuesday, AI concept stocks collectively gained strength. As of the time of writing, Kute Smart was limit-up by 20%, Baina Qiancheng rose over 15%, while Hubei Radio & Television, Aerospace Information, Nanxing Co., Ltd., Tianshang Fashion, Zhenghe Ecology, Hongying Intelligent, Taier Heavy Industry, and others hit the 10% daily limit-up. Stocks like Jieyi Technology, BlueFocus, and Zhongyou Technology also posted significant gains.

From a news perspective, Meta's announcement of its acquisition of the AI startup Manus flooded tech circles. On Monday local time, Meta announced on its official website that Manus would be joining the Meta family, bringing leading agent technology to hundreds of millions of users and opening up new commercial opportunities for the company's full product ecosystem. Meta stated: "Manus has built industry-leading autonomous general-purpose agents capable of independently executing complex tasks such as market research, programming, and data analysis. We will continue to operate and sell Manus's services while deeply integrating them into our product system."

Currently, Manus provides daily services to millions of users and businesses worldwide. Since launching its first general AI agent earlier this year, Manus has processed over 147 trillion tokens and created more than 80 million virtual computers. Meta stated: "We plan to extend this service to a broader enterprise domain." Meta added: "The excellent talent team from Manus will join Meta to work with us on developing general agent technology for both consumer and enterprise products, including the Meta AI assistant."

Manus also posted on its official website, announcing it would be joining Meta. Since its launch, Manus has focused on building general-purpose AI Agents to help users efficiently complete research, automation, and complex tasks. Faced with growing user demand globally, the team has continuously iterated the product, striving to make Manus more practical and reliable in real-world use. This partnership with Meta will further solidify Manus's strategic position in the AI application layer—transforming advanced AI capabilities into scalable, reliable systems that execute user-assigned tasks end-to-end with excellence in practical scenarios.

It is reported that earlier this year, Manus launched the industry's first general AI Agent. On December 17th, Manus revealed that just eight months after its official launch, its Annual Recurring Revenue (ARR) had successfully surpassed $100 million. Currently, the company's total revenue run rate has exceeded $125 million annually, a figure that includes usage-based revenue and other business income. Since the release of Manus 1.5, the company has maintained a monthly compound growth rate of over 20% and is accelerating product iteration and the launch of new features.

The financial terms of the transaction between Facebook and Manus were not disclosed, but media reports suggested the acquisition price by Meta could be as high as tens of billions of dollars. Prior to the acquisition by Meta, Manus was conducting a new funding round with a valuation of $2 billion.

Institutions: Bullish on the Domestic AI Industry Chain Recently, major AI model companies like KNOWLEDGE ATLAS and MiniMax have successively advanced their processes for listing in Hong Kong. On December 19th, KNOWLEDGE ATLAS passed the Hong Kong Exchange's hearing and formally submitted its listing application, becoming the first among China's "Big Model Six Little Dragons" to initiate an IPO. The prospectus shows that KNOWLEDGE ATLAS has doubled its revenue for three consecutive years, with 2022, 2023, and 2024 revenues reaching 57.4 million yuan, 124.5 million yuan, and 312.4 million yuan respectively, achieving a compound annual growth rate of 130%. In the first half of 2025, KNOWLEDGE ATLAS's revenue was 191 million yuan, a year-on-year increase of 325%.

On December 21st, MiniMax passed the Hong Kong Exchange's hearing. Based on its self-developed full-modal large model, MiniMax has created a matrix of AI-native products covering both the C-end and B-end, including Conch AI, Talkie, and Xingye. In terms of commercialization progress, as of September 30, 2025, MiniMax had over 212 million individual users from more than 200 countries and regions, as well as 130,000 enterprise customers. Its revenue for the first nine months of 2025 increased by over 170% year-on-year, with overseas market revenue contributing more than 70%.

Kaiyuan Securities pointed out that startups like KNOWLEDGE ATLAS and MiniMax successively advancing IPOs might leverage capital markets to significantly enhance their financing capabilities and replenish funds, thereby increasing investment in model R&D, improving multimodal/reasoning/Agent performance, driving demand for training and inference-side computing power, and accelerating global commercialization. Furthermore, the recent successive releases of multimodal large models by domestic internet giants—such as Alibaba's Wanxiang 2.6 series, Tencent's Hunyuan World Model 1.5, Doubao's Large Model 1.8, and the audio-video model Seedance1.5 Pro—may further accelerate the commercialization of multimodal AI applications in both B-end and C-end markets, stimulating demand for model training and application inference computing power. They recommend continuing to position in the directions of multimodal AI applications and inference computing power.

Changjiang Securities indicated that the recent approvals for KNOWLEDGE ATLAS and MiniMax by the Hong Kong Exchange suggest a high certainty of strong景气度 (prosperity) for large models in 2026. The competition for traffic portals on the to-C side will be a core focus for AI applications next year, and a multimodal explosion is anticipated next year. They remain bullish on the domestic AI industry chain. Hualong Securities believes that with policy support, the AI industry is expected to sustain healthy development, and AI will continue to empower thousands of industries. Simultaneously, increased capital market support is expected to accelerate the domestic AI innovation cycle. Huafu Securities also expressed optimism about the potential explosion in the domestic AI application field.