🇨🇳 China AI News Daily — 2026年05月08日
🇨🇳 China AI News Daily
2026年05月08日 · 10 articles
🤖 New AI Model & Product Releases
海光DCU完成与混元Hy3 preview大模型的深度适配 [Hygon DCU Completes Deep Integration with Tencent Hunyuan Hy3 Preview Large Model]
36氪获悉,近日,海光信息完成深算3号DCU与腾讯混元Hy3 preview大模型的深度适配。据了解,腾讯混元Hy3 preview是腾讯最新发布的大模型版本,具备295B总参数规模,并支持256K超长上下文,在复杂推理、Agent能力和代码生成等方面实现提升。 [Hygon Information has completed deep integration between its Shensuang 3 DCU chip and Tencent's latest Hunyuan Hy3 preview model, which features 295B parameters and 256K context length with improvements in complex reasoning, agent capabilities, and code generation.]
DeepSeek拟募资最高500亿元 [DeepSeek Plans to Raise Up to 50 Billion Yuan in Funding]
据报道,DeepSeek拟募资最高500亿元,这将成为中国人工智能公司有史以来最大的一轮融资。(财联社) [DeepSeek is reportedly planning to raise up to 50 billion yuan, which would become the largest single funding round ever for a Chinese AI company.]
36氪首发 | 清华系AI Infra厂商完成数亿元融资,以GPU为核心重构计算机系统架构 [Exclusive: Tsinghua-Affiliated AI Infrastructure Firm Raises Hundreds of Millions of Yuan to Rebuild Computer Architecture Around GPUs]
硬氪获悉,北京容芯致远科技有限公司(以下简称'容芯致远')近日完成天使轮数亿元融资。本轮由北京绿色能源和低碳产业基金与赛富投资基金领投,顺禧基金、富华资本、万利达集团、长江创新投、水木清华校友基金、梅花创投等跟投。云岫资本此前参与了公司种子轮投资,本轮继续跟投,并担任长期独家财务顾问。 [Beijing-based Tsinghua-affiliated AI infrastructure startup Rongxin Zhiyuan has closed an angel-round funding of hundreds of millions of yuan led by Beijing Green Energy and Low-Carbon Industry Fund and SAIF Partners, to rebuild computer system architecture with GPUs at its core.]
Plaud获头部大厂投资,目前估值达20亿美元|硬氪独家 [Exclusive: AI Recording Device Maker Plaud Backed by Major Tech Giant, Now Valued at $2 Billion]
硬氪独家获悉,AI卡片录音笔公司Plaud已于2025年年中拿下腾讯的融资,估值达10亿美元;目前,Plaud公司估值已涨至约20亿美元。同时,硬氪从相关人士处了解到,Plaud正在推进与腾讯会议之间的硬件合作。 [AI voice recorder maker Plaud, which secured Tencent investment at a $1B valuation in mid-2025, has now reached a valuation of approximately $2 billion and is reportedly developing hardware integration with Tencent Meeting.]
小红书四年AI 路:FOMO、犹豫,到突然加速 [Xiaohongshu's Four-Year AI Journey: From FOMO and Hesitation to Sudden Acceleration]
所有互联网大中厂都渴望在AI时代博得位置,在这场比赛中,小红书曾是克制的那个。在一个搜索属性与社区属性并存,以真实经验分享为核心的产品中,活人感与AI、温情和算法,始终像天平的两端。 [This feature traces how Xiaohongshu has shifted from a cautious, restrained approach to AI integration to a rapid acceleration, as the platform grapples with balancing its authentic community feel against algorithmic AI-driven features.]
📜 AI & Technology Governance Policies
专家透露:今年将出台政策,应对AI可能带来的就业风险 [Experts Reveal: Policies to Address AI-Related Employment Risks to Be Released This Year]
中国今年将出台一系列公共政策,以应对人工智能使用可能带来的就业风险.例如,要求企业在应用AI时,必须对其就业影响进行评估... [China plans to release a series of public policies this year to address AI-related employment risks, including requirements for companies to conduct employment impact assessments before deploying AI.]
国家发改委、国家能源局、工信部、国家数据局4部门重磅发布! [Four Ministries Including NDRC, NEA, MIIT, and National Data Administration Release Major Joint Action Plan]
据国家能源局网站5月8日消息,国家发改委、国家能源局、工信部、国家数据局印发的《关于促进人工智能与能源双向赋能的行动方案... [China's NDRC, National Energy Administration, MIIT, and National Data Administration jointly issued an action plan on May 8 to promote mutual empowerment between AI and the energy sector.]
国家数据局局长最新发言:释放可信数据空间发展信号! [National Data Administration Director's Latest Speech: Signals for Development of Trusted Data Spaces]
2026 第十届数字金融与金融安全大会启幕文/国家发展和改革委员会党组成员,国家数据局党组书记、局长刘烈宏国家发展改革委党组... [National Data Administration Director Liu Liehong signaled the development of trusted data spaces at the 2026 10th Digital Finance and Financial Security Conference.]
📖 Long Reads: Social Impact of AI
AI梦魇将至?——就业冲击、平台垄断与政策抉择 [AI Nightmare Approaching? — Employment Shocks, Platform Monopolies, and Policy Choices]
核心发现芝加哥布斯教授、前印度央行行长Raghuram Rajan对AI就业冲击提出双重审视:一方面,技术采纳的摩擦与惯性会延缓失业浪... [Former RBI Governor Raghuram Rajan offers a dual analysis of AI's employment impact, examining both the friction slowing job displacement and the structural risks of platform monopolization.]
苏州3个项目入选人工智能教育国家级优秀案例 [Three Suzhou Projects Selected as National Model Cases for AI in Education]
近日由教育部教育管理信息中心中央网信办数据与技术保障中心联合开展的全国教育领域'人工智能安全向善优秀案例'征集活动结果... [Three Suzhou education projects were selected in a joint Ministry of Education and Cyberspace Administration initiative to identify national model cases of safe and beneficial AI use in education.]
Sources: WeChat public accounts · 36氪 · 量子位 · FT中文网 · 纽约时报中文网
Don't miss what's next. Subscribe to China AI News Digest: