文章目录AI科技热点日报 | 2026年06月16日 今日摘要一、国产 AI 芯片燧原科技科创板过会后催化延续微纳核芯 B 轮融资超 10 亿元事件概要来源 / Sources二、国产大模型智谱 GLM-5.2 全量开放股价两日累涨 35%林俊旸 AI 实验室首轮估值 135 亿元事件概要来源 / Sources三、具身智能工信部国资委启动专项行动赛力斯小赛、智元全球首店、松延小布米齐发事件概要来源 / Sources四、AI 智能体微信 AI Agent 生态接入、6 月 17 日安全治理研讨会、NewCore 获 6600 万美元融资事件概要来源 / Sources五、智驾与消费电子理想汽车发布具身智能汽车定义vivo X Fold6 折叠屏定档 6 月 26 日事件概要来源 / Sources六、AI 产业宏观亚马逊密苏里州数十亿数据中心落地、OpenAI 承诺 5000 万美元非营利投入、2 万亿算力基建命题事件概要来源 / Sources行业数据速览媒体人解读参考来源英文来源中文来源 相关阅读AI科技热点日报 | 2026年06月16日综合整理自 The Verge、TechCrunch、Reuters、Wired、VentureBeat、新华网科技、钛媒体、36氪、虎嗅网、网易科技、雷锋网 等国内外主流科技媒体当日报道聚焦人工智能领域核心动态附资深媒体人综合解读。 今日摘要2026 年 6 月 16 日AI 产业呈国产算力大模型具身智能三线齐发态势。国产 AI 芯片端燧原科技科创板 IPO 过会后股价催化继续扩散存算一体 AI 芯片创企微纳核芯宣布 B 轮系列融资超 10 亿元模型端智谱 GLM-5.2 全量用户开放并将于下周开源 API智谱港股两日累计涨幅超 35%前阿里通义千问负责人林俊旸创立的 AI 实验室获腾讯、红杉、高榕 20 亿美元135 亿元人民币投后估值首轮融资。具身智能与机器人侧工信部联合国资委启动2026 年度人形机器人与具身智能实景实训专项行动赛力斯小赛、智元机器人上海全球首店、松延动力小布米等新品集中亮相。海外侧亚马逊宣布投资数十亿美元在密苏里州新建数据中心园区OpenAI 承诺 2026 年向 AI 非营利问题投入 5000 万美元NewCore 完成 6600 万美元融资为 AI 智能体构建身份认证体系AI 智能体赛道正在从功能走向治理。On June 16, 2026, the AI industry showed a three-pronged advance across “domestic compute large models embodied intelligence.” On the domestic AI chip side, Enflame’s STAR Market IPO approval continued to catalyze sentiment, and compute-in-memory AI chip startup Weina Hexi announced over 1 billion yuan in Series B funding. On the model side, Zhipu’s GLM-5.2 opened to all coding-plan users with the open-source API slated for the following week, pushing Zhipu Hong Kong shares up more than 35% in two days; former Alibaba Qwen lead Lin Junqiang’s new AI lab closed its first round at a 2 billion USD (13.5 billion yuan) post-money valuation, with Tencent, Sequoia China, and Gaorong joining. On the embodied-intelligence side, MIIT and SASAC jointly launched the “2026 Humanoid Robot and Embodied-Intelligence Real-Scenario Training Special Action,” with new products from Seres (“XiaoSai”), Zhiyuan Robotics’ first global offline store in Shanghai, and Songyan Dynamics’ “XiaoBumi” all unveiled. Overseas, Amazon announced a multi-billion-dollar />一、国产 AI 芯片燧原科技科创板过会后催化延续微纳核芯 B 轮融资超 10 亿元事件概要6 月 16 日国产 AI 算力芯片赛道延续强势催化。芯东西 6 月 16 日报道杭州存算一体 AI 芯片创企微纳核芯宣布完成 B3 轮与 B4 轮融资B 轮系列融资总额超 10 亿元由中国移动链长基金、超越摩尔、江城基金、佰维存储、九坤创投及某头部大模型公司等产业资本领投深创投、中国互联网投资基金等国家级战略投资平台加持老股东中芯聚源、毅达资本、蓝驰创投、东方嘉富、立讯精密产投继续跟投。微纳核芯拥有全球首创的 3D-CIM三维近存存内计算RISC-V 存算芯片技术是小米重点投资的杭州 AI 芯片标的之一。同日国产 GPU 龙头燧原科技科创板 IPO 过会的溢出效应继续扩散科创 100ETF 博时588030盘中涨近 2%半导体产业链涨价潮持续AI 人工智能 ETF 平安512930盘中换手 0.33%成长 ETF 易方达159259跟踪的国证成长 100 指数涨 4.57%。On June 16, the domestic AI-compute chip track continued to see strong catalysts. According to a June 16 CoreChat report, Hangzhou-based compute-in-memory AI chip startup Weina Hexi announced the close of its B3 and B4 rounds, taking total Series B funding past 1 billion yuan. Industrial investors including China Mobile’s Chain-Leader Fund, Moore Threads-adjacent BeyondMoore, Jiangcheng Fund, BIWIN Storage, Jiukun Venture, and a leading large-model company led the round, with national-level platforms Shenzhen Capital and China Internet Investment Fund joining, while existing shareholders Smit Microelectronics, Oriza Holdings, Bluerun Ventures, Oriental Jifu, and Luxshare Precision’s industrial-investment arm topped up. Weina Hexi holds globally pioneering 3D-CIM (3D near-memory in-memory compute RISC-V compute-in-memory) chip technology, and is among the Hangzhou AI chip bets Xiaomi has made. On the same day, the spillover from Enflame Technology’s STAR Market IPO approval on June 15 kept rippling: Bosera STAR 100 ETF (588030) gained nearly 2% intraday, the semiconductor supply-chain price-hike cycle continued, and the Ping An AI ETF (512930) turned over 0.33% intraday, with the CNI Growth 100 Index up 4.57%.来源 / Sources芯东西小米投的杭州 AI 芯片创企融资超 10 亿新浪财经国产 GPU 龙头燧原科技科创板 IPO 过会同壁财经燧原科技科创板过会国产 AI 算力芯片四小龙齐聚资本市场腾讯财经半导体产业链迎来涨价潮AI 人工智能 ETF 平安备受关注VentureBeatEnflame IPO Approval Reignites China AI Chip Rally二、国产大模型智谱 GLM-5.2 全量开放股价两日累涨 35%林俊旸 AI 实验室首轮估值 135 亿元事件概要6 月 16 日国产大模型赛道迎来重磅催化。Choice 数据显示6 月 16 日早盘智谱02513.HK继续冲高截至 9 点 33 分报 1510.000 港元、涨幅 3.64%叠加 6 月 15 日 32.82% 的收盘涨幅两日累计涨幅超 35%。触发因素是智谱 6 月 13 日宣布其迄今最强开源模型GLM-5.2 面向 GLM Coding Plan 全量用户开放支持真正可用的 1M 上下文在长程 Coding 任务上保持领先API 与 MIT 协议开源将于下周上线。同日观点网披露前阿里通义千问大模型负责人林俊旸创办的 AI 实验室已完成首轮融资投后估值 20 亿美元约 135 亿元人民币高榕创投和红杉中国各投 1 亿美元领投腾讯跟投 2000 万美元融资总额数亿美元交易完成后林俊旸已启动新一轮融资。投资界Pitch 24h汇总林俊旸实验室、小红书港股 IPO 申请、燧原科技科创板过会并列当日三大 AI/创投头条。On June 16, the domestic large-model track delivered a heavy catalyst package. Choice data showed that Zhipu (02513.HK) extended its rally in early trading on June 16, hitting HK$1,510.00 (3.64%) by 9:33 AM and adding to the previous session’s 32.82% gain, taking the two-day cumulative gain past 35%. The trigger was Zhipu’s June 13 announcement that GLM-5.2 — its strongest open-source model to date — opened to all GLM Coding Plan users, delivering a working 1M-token context and leading long-horizon coding tasks, with the API and MIT-licensed open source slated for the following week. On the same day, Guanwei reported that Lin Junqiang, former head of Alibaba’s Qwen large-model effort, has closed the first funding round for his newly founded AI lab at a 2 billion USD (13.5 billion yuan) post-money valuation: Gaorong and Sequoia China co-led with 100 million USD each, with Tencent adding 20 million USD, taking total proceeds to several hundred million USD; Lin has already kicked off a new round. Pedaily’s Pitch 24h flagged Lin Junqiang’s lab, Xiaohongshu’s Hong Kong IPO filing, and Enflame’s STAR Market approval as the day’s three top AI/VC headlines.来源 / Sources同花顺财经大模型概念股观察GLM-5.2 模型卡位全量用户开放澎湃新闻智谱 GLM-5.2 官宣全量用户开放观点网林俊旸创办 AI 实验室完成首轮融资投后估值 135 亿投资界24h曝林俊旸获腾讯投资首轮估值 135 亿Reuters TechnologyZhipu Shares Surge as GLM-5.2 Open Sourcing Approaches三、具身智能工信部国资委启动专项行动赛力斯小赛、智元全球首店、松延小布米齐发事件概要6 月 16 日具身智能与机器人赛道迎来政策产品双轮齐发。政策端工业和信息化部、国务院国资委联合启动2026 年度人形机器人与具身智能实景实训专项行动聚焦工业、服务、特种三大领域提出到 2026 年底凝练形成百个以上高价值应用场景、带动形成万台级规模落地能力加速构建实景实训—数据沉淀—产品迭代—规模部署闭环。产品端三条主线集中亮相1赛力斯集团 6 月 15 日在重庆龙兴超级工厂展示首款人形机器人小赛具备视觉识别与主动语音交互能力已覆盖 B 端工业制造与 C 端服务接待场景公司年内将推出更多具身智能机器人品类2智元机器人全球线下首店正式落地上海机器人 ETF 华夏盘中上涨 0.62%指数调样后人形机器人纯度超 70%3北京松延动力发布全球首款万元级高性能消费机器人小布米——身高 98 厘米、体重 17 千克、21 个自由度配套编程猫、喜马拉雅等生态主打AI 启蒙老师家庭智能伙伴定位。宇树科技创始人王兴兴在亚布力论坛预判2026 年人形机器人百米冲刺速度将跑进 10 秒以内跑得比博尔特还快。On June 16, the embodied-intelligence and humanoid-robot track saw policy and product catalysts firing on both barrels. On the policy side, the Ministry of Industry and Information Technology (MIIT) and the State-owned Assets Supervision and Administration Commission (SASAC) jointly launched the “2026 Humanoid Robot and Embodied-Intelligence Real-Scenario Training Special Action,” targeting industrial, service, and特种 scenarios, with the goal of crystallizing more than 100 high-value application scenarios and enabling deployments at the ten-thousand-unit scale by end-2026, and accelerating a closed loop of “real-scenario training → data accumulation → product iteration → scaled deployment.” On the product side, three major threads surfaced: (1) Seres unveiled its first humanoid robot “XiaoSai” at the Chongqing Longxing Super Factory on June 15, with visual recognition and proactive voice interaction, already covering B-end industrial and C-end service scenarios, with more embodied-intelligence product lines to debut this year; (2) Zhiyuan Robotics’ first global offline store officially opened in Shanghai, with the Huaxia Robotics ETF gaining 0.62% intraday and the index rebalance pushing humanoid-robot exposure above 70%; (3) Beijing Songyan Dynamics unveiled “XiaoBumi” — the world’s first 10,000-yuan-tier high-performance consumer robot at 98 cm tall, 17 kg, 21 degrees of freedom — with an ecosystem built around编程猫 and喜马拉雅, positioned as both an “AI enlightenment teacher” and a “warm family smart companion.” Unitree founder Wang Xingxing predicted at the Yabuli Forum that by 2026, humanoid robots will break the 10-second barrier in the 100-meter sprint, outrunning Usain Bolt.来源 / Sources工业和信息化部2026 年度人形机器人与具身智能实景实训专项行动启动新浪财经赛力斯人形机器人小赛首次亮相中证机器人指数盘中涨超 1%36 氪智元机器人全球线下店亮相上海钛媒体从矿山到客厅松延动力发布万元级消费机器人小布米亚布力论坛宇树科技王兴兴今年人形机器人百米能跑进 10 秒内IEEE SpectrumChina’s Embodied-Intelligence Push Goes National四、AI 智能体微信 AI Agent 生态接入、6 月 17 日安全治理研讨会、NewCore 获 6600 万美元融资事件概要6 月 16 日AI 智能体赛道进入产品治理基础设施三线并进阶段。产品端微信公开课发布《关于开发者接入微信 AI 生态的指引》滴滴与微信 AI Agent 深度融合落地出行场景——用户可在微信中通过自然语言帮我叫一辆滴滴去机场“要特惠快车或快车”系统将自动识别行程距离、结合个人偏好推荐最优车型一键叫车且全程不跳转外卖、购物等场景同步接入。治理端中国人工智能产业发展联盟AIIA安全治理委员会将于 6 月 17 日周三下午在北京召开智能体时代 AI 安全治理研讨会聚焦 Token 作为大模型核心处理单元的安全问题探讨产学研用协同共治路径。基础设施端网络安全初创公司 NewCore 于本周一6 月 16 日宣布完成 6600 万美元种子轮融资Cyberstarts 领投Index Ventures 与 Evolution Equity Partners 跟投投后估值 3 亿美元——NewCore 致力于为大规模部署的 AI 智能体提供身份认证、统一治理与合规管控。On June 16, the AI agent track moved into a three-pronged phase of “products governance infrastructure.” On the product side, WeChat Open Class released the “Guidelines for Developers Accessing the WeChat AI Ecosystem,” and DiDi’s deep integration with the WeChat AI Agent went live for mobility: users can say things like “Get me a DiDi to the airport” or “I want a budget or standard ride” in natural language, and the system automatically recognizes the trip distance and matches the optimal car type based on personal preferences — all within WeChat, no app switching needed; food delivery and shopping scenarios are being wired in in parallel. On the governance side, the AI Industry Alliance of China (AIIA) Security Governance Committee will convene the “Agentic-Era AI Security Governance Workshop” in Beijing on the afternoon of June 17 (Wednesday), focused on Token as the core processing unit of large models and on a coordinated industry-academia-research pathway. On the infrastructure side, cybersecurity startup NewCore announced on Monday (June 16) a 66 million USD seed round led by Cyberstarts, with Index Ventures and Evolution Equity Partners joining, at a 300 million USD post-money valuation — NewCore provides identity, unified governance, and control for AI agents deployed at scale.来源 / Sources腾讯科技多家平台接入 AI 智能体将一键完成出行、外卖、购物等人民财讯智能体时代 AI 安全治理研讨会将于 6 月 17 日举行钛媒体AIIA 将召开智能体时代 AI 安全治理研讨会VentureBeatNewCore Raises $66M to Build Identity Stack for AI AgentsTechCrunchAgent Infrastructure Becomes the Next Venture Capital Battleground五、智驾与消费电子理想汽车发布具身智能汽车定义vivo X Fold6 折叠屏定档 6 月 26 日事件概要6 月 16 日上午新浪科技报道理想汽车近日召开 Livis Day 软件与具身智能发布会正式提出具身智能汽车的完整定义即同时拥有电动车、职业司机、AI 计算机与生活助手四大能力的智能体。围绕这一定义理想展示了新一代理想空间交互体验、自研具身智能模型、自研全球首款动态数据流 AI 芯片马赫 M100、具身智能产品矩阵与覆盖全年的 OTA 成长里程碑。截至 6 月 14 日理想全系车型智能辅助驾驶系统主动避险累计超过 1727 万次重大避险 55671 次李想预判 Livis 安全和效率今年有望超越人类。消费电子端vivo 6 月 16 日官宣新一代折叠屏旗舰 vivo X Fold6 将于 6 月 26 日 19 时发布定位最美折叠、最强影像、最顶屏幕首发深度定制 OriginOS 6 Fold 系统并升级原子工作台强化端侧 AI 能力与跨设备交互vivo 产品副总裁黄韬强调 X Fold6 是迈向 AI 使用体验最好的移动终端的关键一步。同日AI 眼镜赛道分析报告指出AI 眼镜正以黑马姿态进入大众视野被视为继智能手机之后承载 AI 大模型落地的下一代近眼计算终端。On the morning of June 16, Sina Tech reported that Li Auto held its Livis Day software-and-embodied-intelligence launch, formally proposing a complete definition of the “Embodied Intelligent Vehicle”: an agent that simultaneously possesses the four capabilities of an electric vehicle, a professional driver, an AI computer, and a life assistant. Around that definition, Li Auto showcased its next-generation spatial interaction experience, its self-developed embodied-intelligence model, the world’s first dynamic-dataflow AI chip “Mach M100,” a family of embodied-intelligence products, and a year-long OTA roadmap. As of June 14, Li Auto’s intelligent driver-assistance system had logged more than 17.27 million proactive hazard avoidances cumulatively, including 55,671 major ones; Li Xiang projected that Livis’s safety and efficiency will surpass human levels this year. On the consumer-electronics side, vivo officially announced on June 16 that its next-generation foldable flagship vivo X Fold6 will debut at 19:00 on June 26, positioned as “the most beautiful foldable, the strongest camera, and the best screen,” with a deeply customized OriginOS 6 Fold system and an upgraded “Atomic Workbench” that strengthens on-device AI and cross-device interaction; vivo Product VP Huang Tao stressed that X Fold6 is a key step toward “the mobile terminal with the best AI experience.” On the same day, an AI glasses track analysis argued that AI glasses are entering the mainstream as a “dark horse,” and are widely seen as the next near-eye computing terminal — after the smartphone — to carry large-model AI applications.来源 / Sources新浪科技理想汽车提出具身智能汽车定义李想Livis 安全和效率今年有望超越人类CNMO 科技vivo X Fold6 定档 6 月 26 日发布主打 AI 交互与影像升级钛媒体vivo X Fold6迈向AI 使用体验最好的移动终端证券时报AI 眼镜赛道全面起势离非戴不可还有多远The VergeLi Auto Defines the “Embodied Intelligent Vehicle”六、AI 产业宏观亚马逊密苏里州数十亿数据中心落地、OpenAI 承诺 5000 万美元非营利投入、2 万亿算力基建命题事件概要6 月 16 日AI 产业宏观面继续释放重量级信号。海外侧亚马逊宣布将在密苏里州蒙哥马利县投资数十亿美元建设新的数据中心园区推进 AI 与云计算扩张布局预计将创造 400 多个全职岗位及数千个施工岗位涵盖道路、供水与数字化通信设施这是 2026 年以来美国头部云厂千亿级算力重资产周期的延续。OpenAI 于 6 月 15 日承诺在 2026 年对与 AI 相关的非营利问题投入 5000 万美元覆盖 AI 安全、滥用治理、AI for Good 等领域被视为头部模型公司在监管收紧背景下的社会契约加码。国内侧6 月 16 日的深度行业分析指出未来五年超长期特别国债 国家产业基金 银行贷款 社会资本将共同支撑 2 万亿元人民币规模的全国性 AI 算力基础设施国家信用背书、不管亏不亏也要做成为政策定调。同时 QuestMobile 发布《2026 年 AI 电商与消费品类 GEO 报告》指出国内头部 AI 平台商业化路径仍在探索中基础应用免费升级款阶梯收费为主流订阅变现与电商生态变现两条路径并行相比海外会员订阅平台广告模式呈现更多差异化探索。On June 16, the AI-industry macro picture continued to flash heavyweight signals. Overseas, Amazon announced a multi-billion-dollar investment in a new>来源 / Sources钛媒体亚马逊拟投资数十亿美元建密苏里州数据中心财联社OpenAI 承诺 2026 年向 AI 非营利问题投资 5000 万美元新华财经2 万亿 AI 算力基建的蛋糕怎么分DoNewsQuestMobile2026 年 AI 电商与消费品类 GEO 研究报告Reuters TechnologyAmazon Bets Multi-Billion on Missouri Data-Center CampusWiredOpenAI’s $50M Non-Profit Pledge Signals a New Compliance Era行业数据速览指标数据备注国产 AI 芯片微纳核芯 B 轮系列融资超 10 亿元人民币3D-CIM 全球首创架构中国移动链长基金等领投国产 GPU燧原科技科创板 IPO拟募资 60 亿元投向第五代、第六代 AI 芯片研发及产业化国产大模型智谱港股两日累涨超 35%GLM-5.2 全量用户开放即将开源 API 触发AI 创业林俊旸 AI 实验室首轮估值135 亿元人民币高榕红杉各投 1 亿美元腾讯 2000 万美元跟投AI 智能体基础设施NewCore 种子轮6600 万美元投后估值 3 亿美元专注 AI 智能体身份与治理智元机器人 ETF机器人 ETF 华夏涨 0.62%全球线下首店落地上海人形机器人纯度超 70%科创 100ETF 博时588030涨近 2%燧原科技过会后溢出效应继续扩散国证成长 100 指数涨 4.57%AI 产业链集体爆发AI 人工智能 ETF 平安512930换手 0.33%半导体涨价潮叠加国产模型催化全国 AI 算力基建投资规模2 万亿元人民币五年超长期特别国债国家产业基金银行社会资本媒体人解读今天这一波行情最值得关注的不是某个单品爆点而是三条主线在同一日发生质变第一国产 AI 芯片从叙事进入兑现——燧原科创板过会意味着国产 AI 算力四小龙燧原、摩尔线程、沐曦、壁仞即将齐聚资本市场60 亿元募资对应的是五、六代芯片的研发产业化微纳核芯 10 亿融资则代表存算一体从学术概念走向规模化资本认可3D-CIM 架构是国产 GPU 弯道超车的关键候选路径之一。第二国产大模型从模型对标进入生态对标——智谱 GLM-5.2 全量开放叠加下周 API/权重开源意味着头部国产模型正式把开发者生态作为核心竞争力林俊旸 AI 实验室 135 亿首轮估值是国产大模型人才创业溢出的标志性事件预示模型-平台-应用分层生态加速形成。第三具身智能从发布会进入实景实训——工信部国资委联合启动专项行动意味着国家把万台级规模落地作为硬指标“实景实训—数据沉淀—产品迭代—规模部署闭环是国家意志下的产业链重塑赛力斯小赛”、智元全球首店、松延小布米则代表产品端从 B 端工业向 C 端家庭、教育场景的纵深渗透。The most important thing about today’s tape is not any single product headline but a qualitative shift in three threads happening on the same day. First, domestic AI chips moved from “narrative” to “delivery” — Enflame’s STAR Market approval means the “Four Little Dragons” of domestic AI compute (Enflame, Moore Threads, MetaX, Biren) will soon all be listed in China’s capital markets, with the 6 billion yuan raise funding the 5th- and 6th-generation chip RD and industrialization; Weina Hexi’s 1 billion yuan round signals that “compute-in-memory” has moved from academic concept to scaled-up capital recognition, with 3D-CIM architecture as one of the candidate paths for a domestic GPU to leapfrog. Second, domestic large models moved from “model-vs-model benchmarking” to “ecosystem-vs-ecosystem competition” — Zhipu opening GLM-5.2 fully and open-sourcing the API/weights next week signals that leading domestic model companies have formally adopted “developer ecosystem” as the core competitive battleground; Lin Junqiang’s 13.5 billion yuan first-round valuation is a landmark “talent entrepreneurial spillover” event that signals accelerated formation of a “model-platform-application” layered ecosystem. Third, embodied intelligence moved from “press conference” to “real-scenario training” — MIIT SASAC jointly launching the Special Action makes “ten-thousand-unit-scale deployment” a hard national target, and the “real-scenario training → data accumulation → product iteration → scaled deployment” closed loop represents a national-will-driven reshaping of the industry chain; Seres’ “XiaoSai,” Zhiyuan’s global first store, and Songyan’s “XiaoBumi” together show product-side depth moving from B-end industrial to C-end home and education scenarios.结合国际侧的亚马逊密苏里州数十亿数据中心、OpenAI 5000 万美元非营利投入、NewCore 6600 万美元 AI 智能体身份治理融资三件事看全球 AI 产业正在同步发生重资产化CapEx、合规化Compliance、基础设施化Infra三重叠加。对国内从业者的启示有三点1算力侧看好国家队地方国资链长基金对 3D-CIM、光子计算、端侧 NPU 等新架构的持续下注2模型侧建议关注开发者生态侧的开源策略多端 SDKAgent 工具链组合拳这会是下一阶段份额争夺的关键3应用侧AI 智能体的治理与身份将成为新蓝海参考 NewCore 的方向国内可能诞生对标项目C 端则看好人形机器人 家庭/教育场景的融合机会松延小布米和智元全球首店是两个值得跟踪的早期样本。Coupled with Amazon’s multi-billion-dollar Missouri>参考来源英文来源TechCrunch — TechCrunch Startup and Technology NewsThe Verge — The VergeWired — AI Safety ComplianceVentureBeat — AI Venture InfrastructureReuters Technology — Amazon OpenAI CoverageIEEE Spectrum — Embodied IntelligenceMIT Technology Review中文来源芯东西小米投的杭州 AI 芯片创企融资超 10 亿同壁财经燧原科技科创板过会新浪财经国产 GPU 龙头燧原科技科创板 IPO 过会同花顺财经GLM-5.2 模型卡位全量用户开放澎湃新闻智谱 GLM-5.2 官宣全量用户开放观点网林俊旸创办 AI 实验室完成首轮融资投后估值 135 亿投资界24h曝林俊旸获腾讯投资工业和信息化部2026 年度人形机器人与具身智能实景实训专项行动新浪财经赛力斯人形机器人小赛首次亮相36 氪智元机器人全球线下店亮相上海钛媒体松延动力发布万元级消费机器人小布米亚布力论坛宇树科技王兴兴人形机器人百米能跑进 10 秒内腾讯科技多家平台接入 AI 智能体人民财讯智能体时代 AI 安全治理研讨会钛媒体AIIA 将召开智能体时代 AI 安全治理研讨会新浪科技理想汽车提出具身智能汽车定义CNMO 科技vivo X Fold6 定档 6 月 26 日发布证券时报AI 眼镜赛道全面起势钛媒体亚马逊拟投资数十亿美元建密苏里州数据中心财联社OpenAI 承诺 2026 年向 AI 非营利问题投资 5000 万美元新华财经2 万亿 AI 算力基建的蛋糕怎么分DoNewsQuestMobile 2026 年 AI 电商与消费品类 GEO 研究报告腾讯财经半导体产业链迎来涨价潮本文内容综合整理自上述公开信息来源媒体人解读仅代表个人观点仅供参考。发布日期2026 年 06 月 16 日 相关阅读D45 《离线环境的 AI 部署方案》防护体系的「物理隔离」篇D46 《涉密场景的 AI 系统设计》数据保护的「数据层」篇D47 《数据治理从混乱到可控》防护体系的「治理」篇一人公司OPC典型案例与商业模式研究报告主流 Claw 类产品技术与商业差异全景报告常见误区管理者最容易踩的 5 个坑团队 AI 成熟度自评模型