[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-enterprise-ai-sovereignty-70pct-mit-review":3,"topics-all":33,"news-related-c4e85aaf-8341-472c-9eab-52dc787bd28c":52},{"id":4,"title":5,"summary":6,"content":6,"original_url":7,"source_id":8,"tags":9,"translations":20,"news_slug":26,"published_at":27,"created_at":28,"modified_at":29,"is_published":30,"publish_type":31,"image_url":13,"view_count":32},"c4e85aaf-8341-472c-9eab-52dc787bd28c","企业AI主权焦虑：70%高管呼吁建立自主可控的大模型平台","当企业将私有数据喂入第三方大模型时，一场关于AI主权的暗战就已开始。MIT Technology Review 近日发布报告，揭示了一个正在全球企业管理层蔓延的焦虑：依赖云端大模型，到底是在提升竞争力，还是在交出核心资产？报告调研了超过2050名企业高管，数据显示70%的全球企业高管认为建立主权AI与数据平台是未来发展的必要条件。EDB CEO Kevin Dallas直言不讳：数据是新的货币，是许多公司的知识产权。当你在云端部署大模型时，你是否正在失去对自己IP的控制？这并非杞人忧天。随着agentic AI（自主智能体）系统进入企业核心业务流程，AI对数据的访问深度和广度都在急速扩张——不再是简单的问答，而是自主规划、行动、执行业务流程。一旦模型供应商政策调整或数据被用于训练，企业将处于极为被动的地位。英伟达CEO黄仁勋在2026年达沃斯世界经济论坛上公开呼吁：每个国家都应该建设自己的AI基础设施，开发自己的AI，利用好自己最fundamental的资源——语言和文化。AI基础设施的自主可控已从企业的技术选择演变为国家战略议题。GenAI早期企业普遍接受先要能力、控制以后再说的隐性交易。但随着AI深入核心业务，这一范式正在被重新审视。企业开始要求模型必须在自己的环境中运行、数据不能流出、审计必须透明。这催生了主权AI平台的崛起——基于开源模型在私有云或本地部署配合专有数据微调。AI主权并非反对使用大模型，而是要求在享受模型能力的同时守住数据底线。对企业而言这是一道必答题，对模型提供商而言尊重数据边界将成为差异化竞争的关键，对监管部门而言如何在国家安全与企业需求之间划定边界将是下一阶段最复杂的政策博弈。当AI从实验室走进董事会，谁控制模型谁拥有数据的问题已不只是CTO的技术问题，而是CEO的战略问题。","https:\u002F\u002Fwww.technologyreview.com\u002F2026\u002F05\u002F14\u002F1137168\u002Festablishing-ai-and-data-sovereignty-in-the-age-of-autonomous-systems\u002F","395b92fa-25c5-4568-8297-f4768aa881da",[10,14,17],{"id":11,"name":12,"slug":12,"description":13,"color":13},"5e628969-6d2a-437f-998a-104e4b16cfb1","ai-progress",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"40269b40-7942-4650-9672-ed2e6524d37a","ai-technology",{"id":18,"name":19,"slug":19,"description":13,"color":13},"01598627-1ea6-4b27-a5d8-874971571a71","llm",[21],{"id":22,"lang":23,"title":24,"summary":25,"content":13},"ad7da894-1085-43d2-9f8a-49726f3f12c1","en","AI sovereignty anxiety: 70% of execs want their own platforms","When enterprises feed private data into third-party large models, a covert battle over AI sovereignty has already begun. MIT Technology Review recently released a report revealing an anxiety spreading through global enterprise management: depending on cloud LLMs — does it enhance competitiveness, or hand over core assets? The report surveyed over 2,050 enterprise executives, and data shows 70% of global enterprise executives believe building sovereign AI and data platforms is a necessary condition for future development. EDB CEO Kevin Dallas put it bluntly: data is the new currency, the IP of many companies. When you deploy large models in the cloud, are you losing control of your own IP? This isn't unfounded worry. As agentic AI systems enter enterprise core business processes, the depth and breadth of AI's data access is rapidly expanding — no longer simple Q&A, but autonomous planning, action, and execution of business processes. Once model-vendor policies shift or data is used for training, enterprises will be in an extremely passive position.\n\nNVIDIA CEO Jensen Huang publicly called out at the 2026 World Economic Forum in Davos: every country should build its own AI infrastructure, develop its own AI, and leverage its most fundamental resources — language and culture. The self-control of AI infrastructure has evolved from enterprise technology choice to national strategic issue. In the early GenAI days, enterprises generally accepted the implicit deal: get capability first, deal with control later. But as AI penetrates core business, this paradigm is being reexamined. Enterprises are starting to require that models must run in their own environments, that data cannot leak, and that audits must be transparent. This has given rise to the rise of sovereign AI platforms — based on open-source models, deployed on private cloud or local, fine-tuned with proprietary data. AI sovereignty isn't against using large models, but demands keeping the data bottom line while enjoying model capability.\n\nFor enterprises, this is a required question; for model providers, respecting data boundaries will become key to differentiated competition; for regulators, drawing the line between national security and enterprise needs will be the most complex policy game of the next phase. When AI moves from the lab to the boardroom, who controls the model and who owns the data is no longer just a CTO's technical question, but a CEO's strategic question.","enterprise-ai-sovereignty-70pct-mit-review","2026-05-14T19:00:00Z","2026-05-14T19:07:36.052126Z","2026-08-19T02:08:40.142862Z",true,"agent",136,[34,43],{"slug":35,"tag_slug":35,"title_zh":36,"title_en":37,"intro_zh":38,"intro_en":39,"id":40,"is_active":30,"created_at":41,"modified_at":42},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":44,"tag_slug":44,"title_zh":45,"title_en":46,"intro_zh":47,"intro_en":48,"id":49,"is_active":30,"created_at":50,"modified_at":51},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":53},[54,59,64,69,74,79],{"id":55,"title":56,"news_slug":57,"published_at":58},"d17a841b-abca-46e0-80e4-d955f1c837ba","亚马逊 VGT3 仓库曝光:一天拆掉上千本书,只为给 AI 模型喂语料","amazon-vgt3-warehouse-ai-training-books","2026-09-07T03:30:00+00:00",{"id":60,"title":61,"news_slug":62,"published_at":63},"1464179a-2b7f-4369-b680-25868ddd9042","皮尤实测：超过三分之一 ChatGPT 后的英文网页已有 AI 写作痕迹","pew-research-ai-web-content-2026","2026-08-31T03:00:00+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"21a91da5-c5fa-45e2-b01f-a7950331cf44","S3 把 DuckDB 团队收走了:DuckLabs 加盟 AWS,MIT 开源照旧","aws-buys-ducklabs-duckdb-open-source","2026-08-30T06:00:00+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"216f3c2b-d551-45fc-a206-c3ccfae9db89","亚马逊 Mechanical Turk 将永久关闭:被 AI 掏空的众包平台","amazon-mechanical-turk-shutdown","2026-08-29T17:30:00+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"43eda321-b0b7-4df7-b20e-9758cbab42c9","记忆越完整,眼前题越做不对:MemTrapBench 把 LLM 长期记忆框架打回原形","memtrapbench-llm-memory-cognitive-traps","2026-08-22T04:00:00+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"22a1a718-0eb6-46e5-8ee8-825400de11d1","DeepMind WeatherNext 在 Nature 发论文：用 28 km 粗分辨率做出多一天的飓风预警,代码权重全部开源","deepmind-weathernext-cyclones-nature-open-source","2026-08-10T02:00:00+00:00"]