HONG KONG, Dec. 5, 2025 /PRNewswire/ -- Kuaishou Technology ("Kuaishou" or the "Company"; HKD Counter Stock Code: 01024 / RMB Counter Stock Code: 81024), a leading content community and social platform, announced that on December 3, 2025, Kling AI released the Kling Video 2.6 Model. This update introduces a milestone capability for "simultaneous audio-visual generation," fundamentally transforming the traditional workflow of AI video production model of silent visuals followed by manual dubbing. By enabling the simultaneous generation of visuals, natural voiceovers, sound effects, and ambient atmosphere in a single pass, the model reconstructs the AI video creation workflow and significantly accelerates creative efficiency.

Redefining AI Video Creation Workflow with World-Leading Chinese Voice Generation

The Kling Video 2.6 Model upgrades two major capabilities: text-to-audio-visual and image-to-audio-visual generation. Whether inputting either text or combining images with prompts, users can directly generate videos complete with speech, sound effects, and ambient sounds. The model currently supports Chinese and English voice generation, creating video content up to 10 seconds in length.

This upgrade reshapes the traditional AI video creation workflow, which typically requires generating silent footage first and using separate software for post-production audio. With Kling Video 2.6 Model, creators can instantly generate fully integrated videos with voiceovers, sound effects, and ambient sounds, significantly enhancing creators' efficiency.

Leveraging deep semantic alignment between real-world sounds and dynamic visuals, the Kling Video 2.6 Model delivers exceptional performance in audio-visual synchronization, audio quality, and semantic understanding.

With audio-visual coordination at its core, the Kling Video 2.6 Model achieves tight coordination between voice rhythm, ambient sound, and visual motion. This deep alignment ensures that visual dynamics match audio rhythms, eliminating the disjointed "mismatched audio-video" experience often found in traditional workflows.

In terms of audio quality, beyond supporting a variety of sound types such as voice, sound effects, and ambient sounds, the model generates cleaner, richly layered audio quality. The overall auditory experience closely mirrors realistic audio mixing, meeting the rigorous standards for audio detail in professional-grade production.

Regarding semantic understanding, the model demonstrates robust comprehension of textual descriptions, colloquial expressions, and complex storylines across varied scenarios. It accurately captures creator intent to deliver logically coherent audio-visual content that precisely meets user requirements. Additionally, the Kling Video 2.6 Model maintains world-leading position in Chinese voice generation performance.

One-Click "Simultaneous Audio-Visual Generation" Drives Efficiency Revolution Across Diverse Creative Scenarios: Advertising, Marketing, Social Media, and E-Commerce

The Kling Video 2.6 Model supports the generation of standalone or combined audio types - including speech, dialogue, narration, singing, rap, ambient sound effects, and mixed sound effects. This versatility facilitates broad applications in video content creation across industries such as advertising, marketing, social media, and e-commerce, significantly enhancing creative efficiency. 

For example, in advertising and marketing, the Kling Video 2.6 Model enables one-click generation of short ads featuring narration, character dialogue, and product showcases with comprehensive sound effects. This significantly lowers advertising production costs while improving efficiency.

In social media, the Kling Video 2.6 Model offers extensive applications. Its multi-character dialogue capability allows creators to produce a variety of content, including interviews, scripted performances, and comedy skits. In addition, its music performance capabilities enable diverse creative expressions, including singing, rap, and instrumental performances. With the Kling Video 2.6 Model, creators can substantially reduce costs while streamlining workflows, making social media content creation easier and more budget-friendly.

In e-commerce, leveraging solo monologue and narration capabilities, the Kling Video 2.6 Model effectively automates the creation of videos for e-commerce product showcases highlighting key selling points, helping merchants improve operating efficiency.

The launch of the Kling Video 2.6 Model further reduces the costs and complexities of video production within the content creation industries. Moving forward, Kling AI remains committed to developing practical capabilities, empowering creators with superior, easy-to-use AI video creation tools that deliver higher performance.

About Kuaishou

Kuaishou is a leading content community and social platform in China and globally, committed to becoming the most customer-obsessed company in the world. Kuaishou uses its technological backbone, powered by cutting-edge AI technology, to continuously drive innovation and product enhancements that enrich its service offerings and application scenarios, creating exceptional customer value. Through short videos and live streams on Kuaishou's platform, users can share their lives, discover goods and services they need and showcase their talent. By partnering closely with content creators and businesses, Kuaishou provides technologies, products, and services that cater to diverse user needs across a broad spectrum of entertainment, online marketing services, e-commerce, local services, gaming, and much more.

Forward-Looking Statements

Certain statements included in this press release, other than statements of historical fact, are forward-looking statements. Forward-looking statements generally can be identified by the use of forward-looking terminology such as "may", "might", "can", "could", "will", "would", "anticipate", "believe", "continue", "estimate", "expect", "forecast", "intend", "plan", "seek", or "timetable". These forward-looking statements, which are subject to risks, uncertainties, and assumptions, may include our business outlook, estimates of financial performance, forecast business plans, growth strategies and projections of anticipated trends in our industry. These forward-looking statements are based on information currently available to the Group and are stated herein on the basis of the outlook at the time of this press release. They are based on certain expectations, assumptions and premises, many of which are subjective or beyond our control. These forward-looking statements may prove to be incorrect and may not be realized in the future. Underlying these forward-looking statements are a large number of risks and uncertainties. In light of the risks and uncertainties, the inclusion of forward-looking statements in this press release should not be regarded as representations by the Board or the Company that the plans and objectives will be achieved, and investors should not place undue reliance on such statements. Except as required by law, we are not obligated, and we undertake no obligation, to release publicly any revisions to these forward-looking statements that might reflect events or circumstances occurring after the date of this press release or those that might reflect the occurrence of unanticipated events.

For investor and media inquiries, please contact:
Kuaishou Technology
Investor Relations
Email: [email protected] 

香港2025年12月5日 /美通社/ -- 領先的內容社區及社交平台快手科技（「快手」或「公司」；港幣櫃台股份代號：01024 / 人民幣櫃台股份代號：81024）宣佈，12月3日，可靈推出視頻生成2.6模型，該模型提供了里程碑式的「音畫同出」能力，徹底改變了傳統AI視頻生成模型「先無聲畫面、後人工配音」的工作流程。它能夠在單次生成中，輸出包含自然語言、動作音效以及環境氛圍音的完整視頻，重構了AI視頻創作工作流，極大提升創作效率。 

重構AI視頻創作工作流 中文語音生成效果全球領先 

可靈2.6模型升級了文生音畫、圖生音畫兩大功能，輸入文本或是輸入圖片結合提示詞文本，均可直接生成帶有語音、音效及環境音的視頻。語音當前支持生成中文以及英文，生成視頻長度最長支持10秒。

該升級重構了傳統AI視頻創作「先生成無聲的視頻，再結合其他軟體完成後期聲音製作」的工作流，創作者使用可靈2.6模型，能夠直接生成包含人聲、環境與效果音效的視頻，極大提升了創作者的工作效率。 

通過對物理世界聲音與動態畫面的深度語義對齊，可靈2.6模型在音畫協同、音频質量和語義理解上表現亮眼。

音畫協同上，可靈2.6模型生成的視頻，在語音節奏、環境音與畫面動作上緊密呼應，實現了對畫面動態與聲音節奏的深度對齊，避免了傳統工作流可能產生的「畫面一套、聲音一套」的割裂體驗。

音頻品質上，在支持人聲、音效、環境聲等多類型聲音生成的基礎上，生成的音頻音質更乾淨、層次更豐富，整體聽感更接近真實的混音效果，滿足專業級創作對聲音細節的高要求。

語義理解上，該模型對多種場景下的文本描述、口語表達和複雜劇情有較強的語義理解能力，能夠更準確地把握創作者意圖，從而輸出邏輯更嚴密、更貼合用户需求的音畫內容。同時，可靈2.6模型在中文語音生成效果上保持全球領先。

一鍵「音畫同出」 廣告營銷、自媒體、電商等創作場景迎來效率革命

可靈2.6模型支持的聲音類型包括說話、對話、旁白、唱歌、Rap、環境音效、混合音效等多種聲音的單獨或混合生成，能夠廣泛地應用於廣告營銷、自媒體、電商等行業的視頻內容創作，極大提升創作效率。 

例如在廣告營銷行業，可靈2.6模型能夠一鍵生成含有旁白介紹、演員對白、商品展示等綜合聲音效果的廣告短片，極大降低廣告片拍攝的成本，提升效率。 

在自媒體領域，可靈2.6模型的應用則更加廣泛。通過多人對白能力，創作者可以創作包含訪談節目、劇情演繹、搞笑短劇等多類型的內容創作，還可以借助模型的音樂表演能力，進行唱歌、說唱表演、樂器演奏等類型的內容創作。運用可靈2.6模型，可大幅度降低自媒體創作者的創作成本與難度。

在電商領域，通過單人獨白、旁白解說等能力，可靈2.6模型能夠完成商品展示、產品講解等電商領域的視頻內容創作，提升商家的經營效率。 

可靈2.6模型的推出，進一步降低了內容創意行業視頻創作的成本與難度，未來可靈AI也會持續開發更多實用功能，為創作者提供更優質、便捷的AI視頻創作工具。 

關於快手

快手作為中國乃至全球領先的內容社區及社交平台，致力於成為全球最癡迷于為客戶創造價值的公司。作為一家以人工智能為核心驅動和技術依託的科技公司，快手專注於通過持續的技術創新和產品升級，不斷豐富服務和應用場景，為客戶創造價值。在快手，用戶通過短視頻和直播來記錄和分享他們的生活，發現所需，發揮所長。通過與內容創作者和企業緊密合作，快手提供的技術、產品和服務可滿足用戶的多元化的需求，包括娛樂、線上營銷服務、電商、本地生活、遊戲等。

前瞻性聲明

除過往事實的陳述外，本新聞稿載有若干前瞻性陳述。前瞻性陳述一般可透過所使用前瞻性詞彙識別，例如「或會」、「可能」、「可」、「可以」、「將」、「將會」、「預期」、「認為」、「繼續」、「估計」、「預計」、「預測」、「打算」、「計劃」、「尋求」或「時間表」。該等前瞻性陳述受風險、不確定因素及假設的影響，可能包括業務展望、財務表現預測、業務計劃預測、發展策略及對我們行業預期趨勢的預測。該等前瞻性陳述是根據本集團現有的資料，亦按本新聞稿刊發之時的展望為基準，在本新聞稿內載列。該等前瞻性陳述是根據若干預測、假設及前提作出，當中許多涉及主觀因素或不受我們控制。該等前瞻性陳述或會證明為不正確及可能不會在將來實現。該等前瞻性陳述涉及大量風險及不明朗因素。鑒於上述風險及不明朗因素，本新聞稿內所載列的前瞻性陳述不應視為董事會或本公司聲明該等計劃及目標將會實現，故投資者不應過於依賴該等陳述。除法律要求的情形外，我們並無責任公開發佈可能反映本新聞稿日期後發生的事件或情況或可能反映意料之外事件的該等前瞻性陳述的任何修訂。

投資者及媒體問詢
快手科技
投資者關係
郵箱：[email protected]

CHANGSHA, China, Dec. 5, 2025 /PRNewswire/ -- Zoomlion Heavy Industry Science & Technology Co., Ltd. ("Zoomlion"; 1157.HK) was named one of the first 15 industry-leading smart factories at the 2025 World Intelligent Manufacturing Conference on November 27, 2025. The Company's breakthrough in shared mixed-flow production for multiple types of heavy equipment has enabled a collaborative smart factory cluster for construction machinery. By combining flexibility, scale, and automation, the model is shaping a new standard for the industry's digital and intelligent transformation.

Zoomlion's major breakthrough in intelligent manufacturing addresses a significant industry challenge, namely, efficiently producing low-volume, multi-model heavy equipment on a shared mixed-flow production line. At its Smart Industrial City, the Company built a shared smart excavator factory, achieving three world-first milestones. Zoomlion became the first to implement full-process intelligent manufacturing across six core stages, run a mixed-flow line for over 100 models, and integrate collaborative production across multiple equipment categories.

This shared manufacturing model supports a modular, highly scalable cluster for construction equipment. Designed for interoperability, it spans major equipment categories and connects hundreds of manufacturing steps. Three shared manufacturing centers enable multi-plant integration, pushing steel plate utilization above 90 percent and cutting construction costs by 15 percent. Automated logistics replaces storage with real-time transport, reducing inventory by 70 percent and ensuring precise distribution. Combined with a cross-regional R&D platform and a flexible shared supply-chain model, Zoomlion has built a comprehensive ecosystem that advances efficiency and digital transformation across the industry.

Zoomlion's shared intelligent manufacturing model delivers leading efficiency driven by data and AI. At its excavator smart factory, AI covers over 80 percent of processes, enabling self-organizing production with zero changeover time. The full cycle, from steel plate cutting to final assembly, takes only 6.5 days, and one excavator rolls off the line every six minutes. With the industry's shortest takt time and production cycle, the factory supports sales-driven, customized output while sharply reducing inventory and eliminating the disconnect between production and demand.

The model's success is already scaling globally. Zoomlion has applied its intelligent production system to more than 20 factories worldwide, while also enabling innovation in sectors like agricultural machinery, energy equipment, and emergency response vehicles. Building on advances in AI and robotics, the company plans to deepen technology integration, accelerate innovation in next-generation processes, and share its smart manufacturing model with global partners across industries.

According to foreign media, Meta (META.US) may completely cut 10% of its expenses, with the reduction in the Metaverse department possibly being as high as 30%, Citi Research issued a research report saying.

Although Meta's 10% cut for its expenses annually has become the norm since 2022, this significant reduction in Metaverse investment will undoubtedly free up resources for the Super Intelligence Labs to support the launch of its new AI products in 2026.

Citi Research believed that the above news indicate that Meta can achieve sustained product-driven growth by reallocating resources to the most promising areas. Meta remains its top pick. Therefore, the broker reiterated rating at Buy and a target price of US$850.
~


AASTOCKS Financial News
Website: www.aastocks.com

CMSI released a research report forecasting accelerated integration of AI models in 2026. The US led in powerful agents and multimodal systems, such as GPT-5 and Gemini 3, but China is rapidly closing the gap through highly efficient open-source models, bringing intense competition to the global market.

Regarding China's internet sector, the broker noted that AI advertising and game drove TENCENT (00700.HK)'s growth, while rational competition in local services benefited BABA-W (09988.HK)(BABA.US). AI upgrades and cost efficiency advantages accelerated overseas expansion in cloud, e-commerce and content.

CMSI remained optimistic about the internet industry, and favored companies with characteristics such as solid profit growth, competitive moat, sound risk-return profile and sizable shareholder returns. Its top picks include TENCENT, BABA-W and BILIBILI-W (09626.HK).

Regarding US equities, CMSI reiterated its preference for large tech companies, believing that they are best positioned to address AI challenges, leverag strong core businesses to subsidize AI investments meanwhile implementing monetization strategies.

Its top pick for US stock market is Alphabet (GOOG.US) > Meta (META.US) > Amazon.com (AMZN.US) > Microsoft (MSFT.US).
~

AASTOCKS Financial News
Website: www.aastocks.com

美國馬里蘭州羅克維爾市和中國蘇州2025年12月5日 /美通社/ -- 致力於在腫瘤等領域開發創新藥物的領先的生物醫藥企業 -- 亞盛醫藥（納斯達克代碼：AAPG；香港聯交所代碼：6855）宣佈，公司原創1類新藥奧雷巴替尼（商品名：耐立克^®）聯合化療治療新診斷費城染色體陽性（Ph+）急性淋巴細胞白血病（ALL）患者的全球註冊III期臨床研究（POLARIS-1，NCT06051409）獲美國食品藥品監督管理局（FDA）和歐洲藥品管理局（EMA）同意開展。作為耐立克^®在歐美監管機構獲批的第二個全球註冊III期研究，POLARIS-1研究在多國家多中心同步入組，將加速耐立克^®全球特別是歐美市場的上市進程。

POLARIS-1研究為國際多中心、隨機對照、開放性的全球III期臨床試驗，旨在評估耐立克^®聯合化療在新診斷Ph+ ALL患者中的有效性和安全性。POLARIS-1研究已於2023年獲中國CDE臨床試驗許可並迅速啟動。

值得一提的是，POLARIS-1研究的最新進展也將在即將召開的2025年美國血液學會（American Society of Hematology，ASH）年會上亮相，這是該研究數據的首次公佈。目前已披露的摘要數據顯示：在耐立克^®聯合低強度化療治療的初治Ph+ ALL患者中，三週期分子微小殘留病（MRD）陰性率、分子MRD陰性完全緩解（CR）率均可達65%左右，較國外同類產品在同樣條件下的療效明顯提升；即使是對一些高危亞型如攜帶IKZF1plus基因變異的患者也達到了很好的療效。同時，該聯合方案的安全性表現優異，副作用發生率低且可控制。

Ph+ ALL在成人ALL患者中約占20%-30%，具有主要在老年中發病、復發率高、無病生存期短、預後差等特點。在小分子靶向藥物TKI問世之前，單純化療治療Ph+ ALL的5年總生存（OS）率低於20%。TKI的應用顯著改變了該領域的治療前景，但一代以及二代TKI治療對於Ph+ ALL患者仍存在一定局限性。

亞盛醫藥首席醫學官翟一帆博士表示：「這一針對Ph+ ALL初治患者的全球註冊性Ⅲ期研究獲得美國FDA和歐洲EMA批准，是耐立克^®在Ph+ ALL領域的國際重大突破。秉承『解決中國乃至全球患者尚未滿足的臨床需求』這一使命，我們將積極推進耐立克^®在該適應症的臨床開發，以期早日獲批，從而造福全球患者。」

耐立克^®是亞盛醫藥原創1類新藥，為口服第三代TKI，是中國首個獲批上市的第三代BCR-ABL抑制劑。其在中國的商業化推廣由亞盛醫藥和信達生物共同負責。目前，耐立克^®已在中國獲批多個CML耐藥領域治療的適應症，且所有獲批適應症均已被納入國家醫保藥品目錄。同時，耐立克^®在Ph+ ALL領域深度佈局，針對Ph+ ALL領域的治療已獲《CSCO惡性血液病診療指南》連續推薦，並獲CDE突破性療法認定。

2024年6月14日，亞盛醫藥與跨國製藥企業武田就耐立克^®簽署了一項獨家選擇權事宜。一旦行使選擇權，武田將獲得開發及商業化耐立克^®的全球權利許可，惟中國大陸、中國香港特別行政區、中國澳門特別行政區、中國台灣等地區除外。

關於亞盛醫藥

亞盛醫藥是一家綜合性的全球生物醫藥企業，致力於研發創新藥，以解決腫瘤等領域全球患者尚未滿足的臨床需求。2019年10月28日，公司在香港聯交所主板掛牌上市，股票代碼：6855.HK；2025年1月24日，公司在美國納斯達克證券交易所掛牌上市，股票代碼：AAPG。

亞盛醫藥已建立豐富的創新藥產品管線，包括抑制Bcl-2和 MDM2-p53 等細胞凋亡通路關鍵蛋白的抑制劑；新一代針對癌症治療中出現的激酶突變體的抑制劑等。

公司核心品種耐立克^®是中國首個獲批上市的第三代BCR-ABL抑制劑，且獲批適應症均被成功納入國家醫保藥品目錄，目前，亞盛醫藥正在開展耐立克^®一項獲美國FDA許可的全球註冊III期臨床研究（POLARIS-2），用於治療既往接受過治療的慢性髓細胞白血病慢性期（CML-CP）成年患者。此外，耐立克^®聯合治療新診斷費城染色體陽性急性淋巴細胞白血病（Ph+ ALL）患者和治療琥珀酸脫氫酶（SDH）缺陷型胃腸道間質瘤（GIST）患者的全球註冊III期研究正在開展中。

公司另一重磅品種利生妥^®是中國首個獲批上市的國產原創Bcl-2抑制劑，已獲批用於既往經過至少包含布魯頓酪氨酸激酶（BTK）抑制劑在內的一種系統治療的成人慢性淋巴細胞白血病/小淋巴細胞淋巴瘤（CLL/SLL）患者。目前，亞盛醫藥正在開展利生妥^®四項全球註冊III期臨床研究，分別為獲美國FDA許可的治療經治CLL/SLL患者的GLORA研究；治療初治CLL/SLL患者的GLORA-2研究；治療新診斷老年或體弱急性髓系白血病（AML）的GLORA-3研究；以及獲美國FDA、歐洲EMA與中國CDE同步批准開展的治療新診斷中高危骨髓增生異常綜合征（MDS）患者的GLORA-4研究。

截至目前，公司4個在研新藥共獲16項FDA和1項歐盟孤兒藥資格認定，2項FDA快速通道資格以及2項FDA兒童罕見病資格認證。憑借強大的研發能力，亞盛醫藥已在全球範圍內進行知識產權佈局，並與武田、默沙東、阿斯利康、輝瑞、信達等領先的生物製藥公司，以及梅奧醫學中心（Mayo Clinic）、丹娜法伯癌症研究院（Dana-Farber Cancer Institute）、美國國家癌症研究所（NCI）和密西根大學等學術機構達成全球合作關係。

亞盛醫藥已在原創新藥研發與臨床開發領域建立經驗豐富的國際化人才團隊，以及成熟的商業化生產與市場營銷團隊。亞盛醫藥將不斷提高研發能力，加速推進公司產品管線的臨床開發進度，真正踐行「解決中國乃至全球患者尚未滿足的臨床需求」的使命，以造福更多患者。

前瞻性聲明

本新聞稿包含根據美國《1995年私人證券訴訟改革法案》，以及經修訂的《1933年證券法》第27A條和《1934年證券交易法》第21E條所界定的前瞻性陳述。除歷史事實陳述外，本新聞稿中的所有內容均可能構成前瞻性陳述，包括亞盛醫藥對未來事件、經營成果或財務狀況所發表的意見、預期、信念、計劃、目標、假設或預測。

這些前瞻性陳述受到諸多風險和不確定性的影響，具體內容已在亞盛醫藥向美國證券交易委員會（SEC）提交的文件中詳細說明，包括2025年1月21日提交的經修訂的F-1表格註冊說明書和2025年4月16日提交的20-F表格中的「 風險因素」和「 關於前瞻性陳述及行業數據的特別說明」章節、2019年10月16日提交的首次發行上市招股書中的「前瞻性聲明」、「風險因素」章節，以及我們不時向SEC或HKEX提交的其他文件。這些因素可能導致實際業績、運營水平、經營成果或成就與前瞻性陳述中明示或暗示的信息存在重大差異。本前瞻性聲明中的陳述不構成公司管理層的利潤預測。

因此，該等前瞻性陳述不應被視為對未來事件的預測。本新聞稿中的前瞻性陳述僅基於亞盛醫藥當前對未來發展及其潛在影響的預期和判斷，且僅代表截至陳述發表之日的觀點。無論出現新信息、未來事件或其他情況，亞盛醫藥均無義務更新或修訂任何前瞻性陳述。

SAN FRANCISCO and SUZHOU, China, Dec. 5, 2025 /PRNewswire/ -- Innovent Biologics (HKEX: 01801) today announced that the global strategic collaboration with Takeda (TSE: 4502, NYSE: TAK) has closed and become effective following the satisfaction of all closing conditions. The collaboration, initially announced on October 22, 2025, aims to accelerate the global development and commercialization of Innovent's next-generation immuno-oncology (IO) and antibody-drug conjugate (ADC) therapies, including the global partnership on IBI363 (PD-1/IL-2^α-bias) and IBI343 (CLDN18.2 ADC), and an option for an early-stage program IBI3001 (EGFR/B7H3 ADC).

Dr. Hui Zhou, Chief R&D Officer for Oncology Pipeline at Innovent, stated, "IBI363 and IBI343 represent our next-generation therapies designed to address critical unmet needs in global cancer treatment. With clear, aligned development plans, Innovent's deep understanding of these assets, combined with Takeda's extensive experience and strong development and commercialization capabilities, we are poised to maximize the clinical potential of these assets across multiple indications. We look forward to the collaboration with our partner going forward."

Under the agreement:

• Innovent and Takeda will co-develop IBI363 globally, and co-commercialize IBI363 in the U.S., with Takeda leading the co-development and co-commercialization efforts under joint governance and aligned development plan. In addition, Innovent has granted Takeda exclusive commercialization rights for IBI363 outside Greater China and the U.S. Takeda has global manufacturing rights to supply IBI363 outside of Greater China, with such rights being co-exclusive with Innovent for commercial supply in the U.S.
• Innovent has also granted Takeda exclusive global rights to develop, manufacture and commercialize IBI343 outside of Greater China.
• Additionally, Takeda receives an exclusive option to license global rights for IBI3001, a first-in-class EGFR/B7H3 bispecific ADC in Phase 1 stage, outside Greater China.

Takeda will pay Innovent an upfront payment of US$1.2 billion, including a US$100 million equity investment in Innovent through new share issuance at premium, i.e., HK$112.56 per share. Furthermore, Innovent is eligible for development and sales milestone payments for IBI363, IBI343, and IBI3001 (if option exercised) totaling up to approximately $10.2 billion, for a total deal value of up to $11.4 billion. Innovent is also eligible to receive potential royalty payments for each molecule outside Greater China, except with respect to IBI363 in the U.S., where the parties will share profits or losses (40/60 Innovent/Takeda).

Detailed information about the collaboration can be found at the official website of Innovent Biologics.

About Innovent Biologics

Innovent is a leading biopharmaceutical company founded in 2011 with the mission to empower patients worldwide with affordable, high-quality biopharmaceuticals. The company discovers, develops, manufactures and commercializes innovative medicines that target some of the most intractable diseases. Its pioneering therapies treat cancer, cardiovascular and metabolic, autoimmune and eye diseases. Innovent has launched 17 products in the market. It has 1 new drug applications under regulatory review, 4 assets in Phase 3 or pivotal clinical trials and 15 more molecules in early clinical stage. Innovent partners with over 30 global healthcare companies, including Eli Lilly, Roche, Takeda, Sanofi, Incyte, LG Chem and MD Anderson Cancer Center.

Guided by the motto, "Start with Integrity, Succeed through Action" Innovent maintains the highest standard of industry practices and works collaboratively to advance the biopharmaceutical industry so that first-rate pharmaceutical drugs can become widely accessible. For more information, visit www.innoventbio.com, or follow Innovent on Facebook and LinkedIn.

Forward-looking statement

This news release may contain certain forward-looking statements that are, by their nature, subject to significant risks and uncertainties. The words "anticipate", "believe", "estimate", "expect", "intend" and similar expressions, as they relate to Innovent, are intended to identify certain of such forward-looking statements. Innovent does not intend to update these forward-looking statements regularly.

These forward-looking statements are based on the existing beliefs, assumptions, expectations, estimates, projections and understandings of the management of Innovent with respect to future events at the time these statements are made. These statements are not a guarantee of future developments and are subject to risks, uncertainties and other factors, some of which are beyond Innovent's control and are difficult to predict. Consequently, actual results may differ materially from information contained in the forward-looking statements as a result of future changes or developments in our business, Innovent's competitive environment and political, economic, legal and social conditions.

Innovent, the Directors and the employees of Innovent assume (a) no obligation to correct or update the forward-looking statements contained in this site; and (b) no liability in the event that any of the forward-looking statements does not materialize or turn out to be incorrect.

The number of initial unemployment insurance weekly claims in the US last week fell to the lowest level in over 3 years, strengthening investors' confidence in a potential rate cut by the Fed in December.

US stocks showed mixed movements on Thursday. The DJIA opened higher but closed down 31 points or nearly 0.1% to 47,850. The Nasdaq fluctuated and closed up 51 points or 0.2% to 23,505, while the S&P 500 picked up 7 points or 0.1% to close at 6,857, with both indices rising for 3 consecutive trading days.

Meta (META.US) closed up 3.4% at close, amid rumors that Mark Zuckerberg planned to significantly cut the budget for its metaverse project, with layoffs potentially starting as early as next month.

Large techs showed varied performances, with Nvidia (NVDA.US) climbing 2.1%. Tesla (TSLA.US) added 1.7%, while Microsoft (MSFT.US) ticked up 0.7%. Alphabet (GOOG.US) faded 0.7%, while Apple (AAPL.US) lost 1.2%. Micron Technology (MU.US) cascaded 3.2%.
~

AASTOCKS Financial News
Website: www.aastocks.com

香港2025年12月4日 /美通社/ -- 在全球人工智能發展從模型競爭轉向應用深化的關鍵時刻，機器人產業作為「AI走向真實世界」的關鍵載體，正迎來前所未有的政策聚焦與歷史性發展機遇。美國、歐洲、日本、韓國等主要經濟體紛紛將政策重心向機器人產業傾斜，視其為未來製造業、國防體系與社會服務的基礎設施。在這一全球性產業浪潮中，首程控股（697.HK）憑藉其獨特的「投資+應用」雙引擎模式與平台化生態構建能力，正成為中國機器人產業鏈中一支不可忽視的新興力量。

全球政策東風：機器人產業進入加速落地期

當前，全球科技政策呈現出由泛人工智能向機器人領域聚焦的清晰轉向。這一趨勢標誌着機器人產業正從「技術突破期」邁向「政策扶持與商業落地期」的關鍵拐點。

據悉，隔夜美國白宮正在醖釀頒佈機器人產業專項行政命令，並計劃在年底前成立跨部門機器人工作組。此舉意味着機器人很可能繼人工智能之後，成為美國下一個國家級科技戰略工程。歷史表明，美國僅在戰略性基礎設施領域才會動用行政命令級別的推動力，這無疑將機器人產業提升至國家戰略高度。

多國加速戰略佈局：日本將機器人視為解決人口老齡化與勞動力短缺的核心方案；韓國明確提出在2030年前躋身全球機器人強國前三的目標；歐盟啓動了多項具身智能前沿研究項目；而中東與東南亞地區則將機器人納入智慧城市與產業升級的核心規劃。這種全球性的政策共振，共同加速了機器人從「實驗室技術」向「國家基礎設施」的演進節奏。

在這一「政策、技術、商業」三重動力同步成熟的節點，產業競爭的核心已轉變為：誰能夠率先在投資佈局、應用落地與產業協同組織上建立穿透力，誰就更可能主導下一代智能科技產業的發展。

作為熱門港股市場機器人概念股公司之一的首程控股在機器人領域模式最獨特之處在於形成了「投資-應用」的強閉環。公司不僅是資金的提供者，更是產業資源的組織者。它通過推動被投企業的產品在真實場景中進行二次開發、運營驗證，解決了初創企業「技術落地難」的核心痛點。這種將前沿技術從實驗室直接導向電力巡檢等高確定性B端場景的能力，構成了首程難以複製的產業穿透力，也為自身帶來了穩定的現金流與深遠的產業影響力。目前，公司的投資版圖覆蓋了機器人本體、關鍵零部件、動力系統、算法及具身智能等核心領域。其投資的企業包括宇樹科技、雲深處、銀河通用、星海圖、松延動力等，幾乎貫穿了中國機器人產業從底層技術到上層應用的關鍵鏈路。這種佈局確保了公司在未來具身智能產業爆發時，能全面把握各環節的高增長紅利。

雙端落地引擎：夯實B端基礎，開創C端入口

首程控股的商業化路徑清晰分為To B（企業端）與To C（消費者端）兩大陣地，二者相互支撐，共同構建增長飛輪。

首先，在To B端實現深度場景融合，驗證商業化價值。在B端，首程聚焦於機器人商業化確定性最高的領域。通過將生態內企業的產品導入電力巡檢等剛性需求場景，不僅完成了技術可行性與可靠性的實證，更在此過程中積累了寶貴的場景數據與工程經驗，反哺產品迭代，為未來的規模化複製奠定了堅實基礎。

其次，在To C端構建全國性消費觸達網絡，針對中國消費級機器人市場存在的「認知缺口」與「體驗缺口」，首程控股開創性地打造了線上線下融合的消費觸達體系，包括建立線下體驗店。公司打造的「機器人科技體驗店」（目標一年內拓展至20家）已在北京、成都等地成功落地。這些門店集展示、體驗、銷售於一體，被譽為「機器人行業的4S店」。國慶期間單店日銷售額突破十萬元、日客流過萬的業績，證明了將機器人從「看不見、摸不着」的科技概念轉化為「可試、可玩、可買」的消費品的巨大潛力。這是推動中國機器人消費化最關鍵的一步。此外，公司正在密集籌備的機器人專屬直播間，將承擔產品展示、科普教育、銷售轉化與用户數據沉澱等多重功能。當「線上直播引流」與「線下體驗轉化」相結合，首程有望成為中國機器人面向大眾消費市場的最強入口。

首程控股的戰略並非簡單的「投資加銷售」，其本質是構建了一個能夠推動行業技術與需求雙向增長的平台型生態系統。這一系統形成了一個強大的自我強化增長飛輪：投資佈局 → B端落地驗證 → C端市場教育 → 數據反哺企業 → 估值與收益提升 → 再投資深化佈局。在這個飛輪中，前期投資奠定了產業根基，B端應用驗證了商業價值併為技術優化提供真實反饋，C端網絡則加速了市場教育並開闢了廣闊的營收渠道，最終的數據與資本收益又反哺於新一輪、更深度的產業佈局。

隨着全球進入「機器人優先」的新時代，機器人產業有望如新能源汽車、智能手機一樣，開啓一個長達數年的規模化增長週期。首程控股憑藉其全鏈條的投資佈局、深度的B端場景穿透、創新的C端消費入口以及獨特的平台型組織能力，已初步構建起國內最具產業穿透力的機器人生態之一。這不僅使其能夠充分受益於行業貝塔性的增長，更賦予了其通過生態協同創造阿爾法價值的潛力。在時代窗口已然打開的背景下，首程控股正成為一個值得密切關注、並可能在未來數年實現價值重估的產業龍頭標的。

SINGAPORE, Dec. 4, 2025 /PRNewswire/ -- ZTE Corporation (0763.HK / 000063.SZ), a global leading provider of integrated information and communication technology solutions, announced that its Chief Development Officer, Cui Li, spoke at AI Innovation Asia 2025, hosted by Economist Impact, the thought-leadership arm of The Economist Group.

On the panel "How May AI Help You? Agentic AI and the Customer Experience", Cui Li shared ZTE's strategic vision for agentic AI and illustrated how it is already reshaping customer experience and operational models across the company. She also underscored its role in strengthening resilience, enhancing oversight and accountability, and urged organizations to prepare now for the era of agentic AI.

Q1: Setting the scene – How is agentic AI impacting customer experience in your sector?

Agentic AI is redefining the user experience—not only about UI design, but from response to understanding and co-creation. For ZTE, under our "AI for All" strategy, we are integrating agentic AI into four key areas: networks, computing, homes, and personal devices.

For example, we enable level 4 and beyond autonomous networks with three engines—Nebula Telecom Large Model, big data, and digital twin. In real practice, ZTE and China Mobile co-created multi-agents that can detect network problems and enable self-healing, cutting troubleshooting time by 47%.

Q2: Strategic transformation – How will increasing digital autonomy through agentic AI reshape how organizations build resilience and adaptability?

Actually, we're in a highly uncertain era now. This requires us to start with the end in mind, namely, always seek stability from uncertainty and build a strong moat like the snowball effect. Also we must stay agile to spot changes and pivot on a dime, evolving from a machine-like organization into an organic, adaptive one.

And then we need to know how AI can truly help us. Large AI models are already performing at or beyond PhD level. While agents take a step further by integrating memory and tools, acting as the real-world application of models. And agentic AI can coordinate different agents to automate more complex, time-consuming tasks. Of course, this is an ideal status. The truth is, whether agents or agentic AI, they are still at a very early stage, and face many technical challenges. But given the hypergrowth of AI, I believe that solutions will be right in place very soon.

And then I believe AI can exert effects only when we stick with it to the long haul. Intelligence is built on digital and network foundations. Without digital transformation, a company will not truly achieve intelligence, not to mention becoming more resilient or adaptable—it's just like trying to run before you can walk. Also, going intelligent requires knowledge engineering, process restructuring, and an AI mindset, which is a marathon, not a sprint.

At ZTE, our digital journey started in 2016, and then intelligent transformation in 2022. Here is our experience: infrastructure first, but keep a balance between hardware and software; carry out systematic planning from the top down, to ensure everyone is on the same page; make continuous investment, as disruptive breakthroughs come from every small step; and finally, start from high-value and concrete scenarios, then iterate fast to address any uncertainties along the way.

Q3: Oversight & accountability – As AI systems make more decisions autonomously, how can companies maintain oversight, ensure accountability, and preserve digital sovereignty? 

In a nutshell, to keep humans in the loop. Tasks like design, review, decision-making, and supervision—they still need to be done by people, who should remain ultimately accountable. Automation is just a means, not an end. What humans should really worry about is not being replaced—it's stepping back or being absent in this process.

Every coin has two sides, so do these models. Their generalization, emergent abilities, and continuous evolution—yes, these are the features of a truly game-changing tech. But they inherently come with hallucinations, black-box problems, and so on. In addition, human has social intelligence and morality, something that AI can never truly master. So essentially, AI is built on statistical models—it lacks common sense in the real world, let alone navigating complex trade-offs like humans do.

More importantly, rolling out AI in a business needs the deep integration with know-how. We have to consider factors like accuracy, security, compliance, and division of responsibility, and take into account both workflows and KPIs. From ZTE's hands-on experience, I'd like to give some tips: First, companies should develop their own knowledge engineering projects and domain-specific large models—plus RAG and digital twins to be professional and reliable. Second, identify the concrete problems that agents need to solve. A one-size-fits-all agent often ends up doing nothing well. Third, know when to use agents or workflows. Agents are better at handling complex tasks with variable execution paths, while workflows are more accurate and efficient in highly predictable scenarios. Last, enable end-edge-cloud collaboration to ensure both cost efficiency and security. While among all these key parts, humans are still the ones steering AI in the right direction and creating real value.

Q4: Looking ahead – How do you see agentic AI evolving from task-based automation to integrated business partners, and what immediate actions should organisations take to future-proof themselves for agentic AI?

From a tech perspective, we can think of agents or agentic AI as proactive digital workers. Beyond simple or repetitive tasks, they can connect entire workflows, realize cognitive automation, and even self-evolve. Agents now work well in scenarios that are well-structured, info-heavy, fault-tolerant, and have clear feedback loops. But they often get stuck in the lab when the real-world environments become more complex or risky. So like I said, agents and agentic AI are still early. In the next year or two, they will mainly focus on vertical industries. After that, they will take on complex tasks with greater autonomy, getting more generalized, adaptive, and able to learn and evolve. Agents now are developing very fast. Gemini 3, which was just launched last month, sets a new bar for AI models with SOTA reasoning, multimodal understanding, and agentic capabilities.

For organizations, I think embracing AI is the only way to go. Deploying AI isn't just about connecting to APIs—it's reshaping processes, structures, teams. Companies need to first make mid- and long-term plans, and be adaptable enough to keep up with tech and market changes. Next, start from high-value, business-specific scenarios, and then iterate fast. That's how we can truly master AI. Also, it is reshaping our talent strategy. In the future, three types of talent will matter most: AI experts, who drive this technology forward; AI power users, who foster innovation and enhance efficiency; and people who go beyond AI with high-order thinking and a healthy mindset. Finally, to give AI full play, companies should restructure themselves and plan for a future of "human-AI symbiosis".

AI Innovation Asia 2025 is a high-level dialogue platform connecting enterprise leaders, technology pioneers, and policymakers. Centered on 15 in-depth thematic sessions and insights from over 40 industry experts, the platform focuses on the commercialization pathways of frontier technologies, such as generative AI and Agentic AI, to help businesses turn technical insight into tangible growth and navigate sustainable digital transformation across Asia-Pacific's complex market landscape.

MEDIA INQUIRIES:
ZTE Corporation
Communications
Email: [email protected]