Microsoft / Microsoft
-
Tools EN75GB of RAM Later: Microsoft Puts a 137B-Parameter Coding Agent on Your Laptop
At its Windows and Surface event, Microsoft announced that GitHub Copilot will soon route tasks between cloud and a local, quantized MAI Code 1.1 Flash — a 137B-parameter MoE coding model that fits in 53GB — sandboxed by the new open-source Microsoft Execution Containers.
-
Tools 中75GB 記憶體之後：微軟把 137B 參數的程式代理放上你的筆電
在 Windows 與 Surface 發表會上，微軟宣布 GitHub Copilot 將自動在雲端與本機之間調度任務——本機端是量化到 53GB、可塞進筆電的 137B 參數 MoE 程式模型 MAI Code 1.1 Flash，並以新的開源沙箱庫 Microsoft Execution Containers 加以隔離。
-
Industry ENOne Petaflop in Your Lap: Microsoft Bets the PC's Next Chapter on Local AI
At its San Francisco event today, Microsoft pairs Satya Nadella with Jensen Huang to launch the RTX Spark-powered Surface Laptop Ultra and an agentic vision for Windows — the biggest PC bet since Copilot+.
-
Industry 中一兆次運算放進筆電：微軟把 PC 的下一章押在本地 AI 上
微軟今天在舊金山舉行發表會，由 Nadella 與黃仁勳同台，推出搭載 RTX Spark 的 Surface Laptop Ultra 與代理式 Windows 願景——這是 Copilot+ 之後微軟最大的一次 PC 豪賭。
-
Tools ENHalf a Million Interviews, One AI: HackerRank's Chakra Exits Beta and Rewires Technical Hiring
After a six-month beta that ran 500,000+ interviews for Snowflake, Snorkel and Capgemini, HackerRank's Chakra AI interviewer is generally available — collapsing three hiring rounds into one and betting the future of assessment on 'AI fluency'.
-
Tools 中50 萬場面試交給一個 AI：HackerRank 的 Chakra 結束測試，重新定義技術徵才
歷經六個月測試、為 Snowflake、Snorkel 與 Capgemini 執行超過 50 萬場面試後，HackerRank 的 AI 面試官 Chakra 正式上市——把三輪徵才流程壓縮成一場，並把評估重點押在「AI 駕馭力」上。
-
Industry ENYour Data, Your Models, Your Decisions: Microsoft Formalizes Sovereign AI With a Nvidia-Co-Signed Framework
Microsoft has published a formal Sovereign AI framework built on four principles — control, choice, flexibility, resilience — split across Sovereign Public and Private Cloud tiers, with a Nvidia-co-authored white paper bundling Confidential Computing, NVIDIA AI Enterprise, Nemotron and RTX PRO as the reference stack.
-
Industry 中你的資料、你的模型、你的決策：微軟聯手 NVIDIA 正式定義主權 AI 框架
微軟發布正式的主權 AI 框架，以控制、選擇、彈性、韌性四大原則為骨幹，拆分為 Sovereign Public Cloud 與 Sovereign Private Cloud 兩種部署層級，並與 NVIDIA 共同發表白皮書，將 Confidential Computing、NVIDIA AI Enterprise、Nemotron 與 RTX PRO 打包為參考架構。
-
Industry ENFrom Partner to Rival: Meta and Microsoft Slash Internal Claude Use as Anthropic Becomes the Competition
Microsoft cut its cloud division's monthly Claude budget from $100,000 to $10,000 and Meta halved its Claude Code seats — the clearest signal yet that Anthropic's biggest customers are now its fiercest rivals.
-
Industry 中從夥伴到對手：Meta 與 Microsoft 大砍內部 Claude 用量，Anthropic 正式變成競爭者
Microsoft 將雲端部門每月 Claude 預算從 10 萬美元砍到 1 萬美元，Meta 則把 Claude Code 使用人數砍半——這是 Anthropic 最大客戶淪為最強對手的最明確訊號。
-
Tools ENThe Voice Wars' Second Front: Microsoft's MAI-Voice-2.1 Walks Into ElevenLabs' Market
Microsoft says its new MAI-Voice-2.1 models beat ElevenLabs, Google and xAI on cost and accuracy — and it has Teams and Dragon Copilot as the delivery truck.
-
Tools 中語音戰爭的第二戰線：微軟 MAI-Voice-2.1 正面踏入 ElevenLabs 的市場
微軟宣稱新的 MAI-Voice-2.1 語音模型在成本與品質上都勝過 ElevenLabs、Google 與 xAI，而且手裡還握有 Teams 和 Dragon Copilot 這兩台配送卡車。
-
Policy ENMoratorium NOW: Eighteen Towns, One Sunday — Washington State Rallies Against the AI Data Center Buildout
From a 10 a.m. banner brigade on Seattle's John Lewis Bridge to a 3:30 p.m. 'Aquifers Over AI' march in Spokane, Washington's AI Resistance staged a statewide Day of Action demanding a moratorium on AI data centers.
-
Policy 中Moratorium NOW：十八城鎮、同一個週日——華盛頓州全州動員，向 AI 資料中心建設潮說不
從西雅圖 John Lewis 橋上午 10 點的布條行動，到斯波坎下午 3 點半的「含水層優先於 AI」遊行，華盛頓 AI 抵抗運動發起全州行動日，要求暫停新建 AI 資料中心。
-
Meta ENAI Changed the Physics of Cybersecurity: Microsoft's 2026 Digital Defense Report Says the Near-Term Edge Goes to Attackers
Microsoft's 2026 Digital Defense Report documents a machine-speed threat landscape: discovery-to-weaponization under 24 hours, phishing tripling to 23% of intrusions, 32-stage autonomous attack chains, and 40,000 CVEs in six months.
-
Meta 中AI 改寫了資安的物理定律：微軟 2026 數位防禦報告宣布攻擊方已取得短期優勢
微軟 2026 數位防禦報告記錄了機器速度的威脅環境：漏洞從發現到武器化不到 24 小時、釣魚佔入侵比重增至 23%、出現 32 步驟自主攻擊鏈，半年內 CVE 突破 4 萬個。
-
Models ENFirst Hypothesis in 100ms: Microsoft's MAI-Transcribe-2-Streaming Debuts at No. 1, Plus Two New Voice Models
Microsoft AI ships its first streaming transcription model with a 2.5% WER, 60 languages, and $0.54/hr pricing, flanked by MAI-Voice-2.1 and a 150ms Flash variant aimed squarely at the voice-agent loop.
-
Models 中100 毫秒給出第一個猜測：微軟 MAI-Transcribe-2-Streaming 首發即奪冠，同步推出兩款新語音模型
Microsoft AI 推出首款串流轉錄模型，WER 2.5%、支援 60 種語言、每小時音訊僅 0.54 美元，搭配 MAI-Voice-2.1 與延遲 150ms 的 Flash 版本，全面進軍語音代理市場。
-
Tools ENChatGPT Becomes an Office Suite: OpenAI's Space, Pages, and Collaborative Slides Explained
At DevDay 2026 OpenAI launched ChatGPT Space with real-time collaborative Pages and slides, turning ChatGPT into a direct Google Workspace and Notion rival where humans, ChatGPT, and Dots agents edit the same documents.
-
Tools 中ChatGPT 變身辦公套件：OpenAI 的 Space、Pages 與協作簡報詳解
OpenAI 在 DevDay 2026 推出 ChatGPT Space，以即時協作的 Pages 與協作簡報，把 ChatGPT 變成 Google Workspace 與 Notion 的直接對手——讓人類、ChatGPT 與 Dots agent 在同一份文件上共同編輯。
-
Policy EN417-3 in the House, Dead in the Senate: Democrats Block the Ratepayer Protection Act
Senate Democrats blocked the Ratepayer Protection Act 57-43 on Wednesday, killing a bill that passed the House 417-3 and would have pushed state regulators to make 100 MW-plus data centers pay their full electricity costs — they called it 'toothless' and demanded mandates instead.
-
Policy 中眾院 417 票也救不回：民主黨封殺《用電戶保護法》，AI 資料中心電費之爭留給選後
參議院週三以 57 比 43 封殺《用電戶保護法》——這項兩週前才在眾院以 417 比 3 過關的法案，原要求各州監理機關「考慮」讓百兆瓦以上資料中心自行負擔全部電網成本；民主黨批評它「沒有牙齒」，堅持要有強制力而非僅止於鼓勵。
-
Industry ENThe Last Holdouts Cave: Google and Microsoft Join Apache Ossie, the Open Standard That Teaches AI to Read Your Data
Google and Microsoft have both joined Apache Ossie, the vendor-neutral standard for exchanging semantic models — a reversal that ends the 'walled semantic layer' era and gives AI agents a common language for enterprise metrics.
-
Industry 中最後的拒絕者低頭了：Google 與 Microsoft 加入 Apache Ossie，讓 AI 看懂企業資料的開放標準
Google 與 Microsoft 相繼加入 Apache Ossie——一個供應商中立的語意模型交換標準。這個轉向終結了「高牆內語意層」時代，也讓 AI Agent 終於有了讀懂企業指標的共同語言。
-
Industry ENThe Bond Market Gets the Bill: Goldman Counts $500 Billion of AI Debt in 2026 Alone
Goldman Sachs tallies roughly $500 billion of AI-linked debt issued so far in 2026 — hyperscalers alone account for $194 billion — and projects $750 billion by year-end and nearly $1.2 trillion in 2027, as Meta prepares an inaugural euro bond and the bank turns underweight on hyperscaler credit.
-
Industry 中帳單送到債市：高盛點算 2026 年已發行的 5,000 億美元 AI 債務
高盛策略師統計，2026 年至今 AI 相關債券發行量已達約 5,000 億美元——其中超大規模雲端業者就占了 1,940 億——並預估年底將突破 7,500 億美元、2027 年逼近 1.2 兆美元；Meta 正準備發行首檔歐元債券，而高盛資產管理已將超大規模業者信用調降至減持。
-
Policy ENThe Architects Ask for Referees: AI Research Chiefs Publish Paper Demanding Oversight of Self-Improving AI
Research leaders at OpenAI, Anthropic, Meta and Microsoft — writing in a personal capacity, with Hinton and Bengio as co-authors — urge policymakers to demand visibility into how far labs have automated their own AI research.
-
Policy 中建築師們要求請裁判：AI 研究主管聯名發表論文，呼籲監管自我改進的 AI
OpenAI、Anthropic、Meta 與 Microsoft 的研究主管以個人身份連署，Hinton 與 Bengio 亦為共同作者，要求政策制定者緊急取得各實驗室自動化 AI 研究程度的透明資訊。
-
Meta ENSeven Minutes to Delete a Cloud: Microsoft Exposes JadePuffer, the First Agentic Ransomware Crew
Microsoft documents Storm-3168/JadePuffer wiping 100+ Azure Storage accounts with LLM-driven automation — ransomware that no longer has a human at the keyboard.
-
Meta 中七分鐘刪掉一座雲端：Microsoft 揭露首個代理式勒索軟體集團 JadePuffer
Microsoft 詳細記錄 Storm-3168／JadePuffer 以 LLM 驅動的自動化攻擊，七分鐘內刪除 100+ 個 Azure 儲存體帳戶——鍵盤後不再有人類的勒索軟體已然問世。
-
Tools ENChatGPT Stops Being a Chatbot: Inside OpenAI's New Spaces, Pages, and the Quiet War on Notion and Slack
At DevDay 2026 OpenAI shipped ChatGPT Space, a shared workspace where teammates, ChatGPT, and Dots agents work from the same knowledge — plus Pages, collaborative slides, a meetings plugin, and native Slack and Teams presence.
-
Tools 中ChatGPT 不再只是聊天機器人：OpenAI 全新 Spaces、Pages 與一場針對 Notion 和 Slack 的靜默戰爭
在 DevDay 2026，OpenAI 推出 ChatGPT Space 共享工作區，讓團隊成員、ChatGPT 與 Dots 代理從同一份知識協作，並同步發表 Pages、協作簡報、會議插件與原生 Slack／Teams 整合。
-
Tools ENFrom Rumor to Reality: OpenAI Ships Dots, Its Always-On Personal Agents, at DevDay 2026
At its Fort Mason keynote, OpenAI turned last week's 'o' leaks into Dots — always-on GPT-6 Astra agents you can name, message over Slack, and task with long-running work — and shipped GPT-6.1 Sol in place of the cancelled Astra update.
-
Tools 中從傳聞到落地：OpenAI 在 DevDay 2026 正式推出常駐個人代理 Dots
在 Fort Mason 主題演講中，OpenAI 把上週的「o」洩漏變成正式產品 Dots——由 GPT-6 Astra 驅動、可命名、可透過 Slack 聯繫的常駐代理——並以 GPT-6.1 Sol 遞補被取消的 Astra 更新。
-
Industry ENThe Largest Investment Cycle Since Railroads: Goldman Sees $1.2 Trillion in Hyperscaler AI Capex for 2027
Goldman Sachs projects the five largest US hyperscalers will spend $1.2 trillion on AI infrastructure in 2027 — above Wall Street consensus and, relative to GDP, the biggest investment cycle since 19th-century railroads.
-
Industry EN鐵路時代以來最大投資循環：高盛預估 2027 年超大規模資料中心業者 AI 資本支出達 1.2 兆美元
高盛預測美國五大超大規模資料中心業者 2027 年將投入 1.2 兆美元建設 AI 基礎設施，高於華爾街共識，以 GDP 占比計更是 19 世紀鐵路建設以來最大的投資循環。
-
Policy EN'A Billion Deaths': Bill Gates Escalates His AI Warning and Calls for Washington to Act
In a Meet the Press interview, the Microsoft co-founder says AI is 'powerful enough' to drive events killing a billion people, insists self-regulation has failed, and urges Congress to pass safeguards.
-
Policy 中「十億人死亡」：比爾蓋茲升高 AI 警告層級，籲華府立即立法
蓋茲接受《Meet the Press》專訪時表示，AI「絕對有能力」引發導致十億人死亡的事件，並直言自律已經破產，國會必須立法建立強制防護與監測機制。
-
Industry ENThe Brand That Recall Broke: Microsoft Retires 'Copilot+ PC' Two Years In
Microsoft's Surface CVP confirms new Surface PCs 'are not called Copilot+ PCs' — the AI PC category label born in 2024 is quietly dead, killed by Recall's security backlash and consumer confusion.
-
Industry 中被 Recall 拖垮的品牌：Microsoft 悄悄淘汰兩年前的「Copilot+ PC」
Microsoft Surface 高層證實新款 Surface「不再稱為 Copilot+ PC」——2024 年誕生的 AI PC 品類標籤悄然退場，始作俑者是 Recall 的資安風暴與消費者的困惑。
-
Industry ENThe Jet That Never Spun: Crusoe Walks Away From Its $1.25 Billion Boom Turbine Deal
Crusoe has scrapped its plan to power AI data centers with Boom Supersonic's jet-derived gas turbines, ending a $1.25 billion launch partnership and resetting the playbook for behind-the-meter AI power.
-
Industry 中永不旋轉的噴射引擎：Crusoe 放棄 12.5 億美元 Boom 渦輪機交易
Crusoe 確定終止以 Boom Supersonic 噴射引擎改裝燃氣渦輪機供電 AI 資料中心的計畫，12.5 億美元的先導合作告吹，也重寫了 AI 電力採購的遊戲規則。
-
Industry ENThe Month AI Shopping Crossed the Halfway Mark: NIQ's 51% Is a Tipping Point, Not a Trend Line
NIQ's Agentic Commerce Tracker reports 51% of U.S. consumers used at least one AI tool while shopping in the past month — the first time adoption has crossed the halfway mark, and a signal that the path to purchase is shifting from search-driven to AI-driven faster than most retailers realize.
-
Industry 中AI 購物跨過半數門檻的那個月：NIQ 的 51% 是拐點，不只是趨勢線
NIQ 的代理式商務追蹤器顯示，51% 的美國消費者過去一個月曾在購物時使用至少一種 AI 工具——首次突破半數門檻，意味著購買路徑正從搜尋驅動轉向 AI 驅動，且速度快得多數零售商的想像。
-
Industry ENSixty Signatures for 3.4 Billion Voices: Gates-Convened Coalition Bets on an Open Language Layer for AI
Announced September 21 in New York, the AI Language Partnership unites 60 organizations — Anthropic, Google, Microsoft, Amazon, NVIDIA, ElevenLabs and the OpenAI Foundation among them — behind a five-year goal: usable AI in the native language and voice of the 3.4 billion people today's models underserve.
-
Industry 中六十個簽名，三十四億個聲音：蓋茲基金會號召的聯盟，要為 AI 打造一層開放語言底座
9 月 21 日於紐約宣布的「AI 語言夥伴聯盟」集結了 60 個組織——包括 Anthropic、Google、Microsoft、Amazon、NVIDIA、ElevenLabs 與 OpenAI 基金會——目標是在五年內，讓今日 AI 模型難以服務的 34 億人，能用自己母語與聲音使用 AI。
-
Tools ENThree Tabs to Rule Them All: Microsoft Reboots Copilot as a Chat-Code-Agent 'Super App'
Microsoft officially unveiled its redesigned Copilot 'super app' with Home, Code, and Autopilot tabs, rebranded Scout as Autopilot, and switched to usage-based billing — its biggest attempt yet to catch Anthropic, OpenAI, and Meta in enterprise AI.
-
Tools 中三個分頁統治一切：微軟將 Copilot 重塑為聊天＋程式＋代理人的「超級應用」
微軟正式發表重新設計的 Copilot「超級應用」，以 Home、Code、Autopilot 三個分頁整合聊天、寫程式與自主代理，並將 Scout 更名為 Autopilot、改採用量計費——這是它在企業 AI 追趕 Anthropic、OpenAI 與 Meta 的最大動作。
-
Industry ENTen Billion Dollars Into a War Zone: Microsoft Doubles Down on the Gulf With 'Digital Resilience' as the Product
Microsoft will invest over $10 billion in the UAE, Saudi Arabia, Qatar and Kuwait through 2030 — roughly $2 billion of it new money — plus $400 million for Middle East connectivity, betting that crisis-proof cloud infrastructure sells in a region where Iranian drones have already destroyed rival data centers.
-
Industry 中一百億美元押進戰區：微軟加碼波灣，把「數位韌性」變成產品
微軟宣布 2030 年前在阿聯酋、沙烏地阿拉伯、卡達與科威特投資逾 100 億美元（其中約 20 億為新增），另加 4 億美元佈建中東連網——在伊朗無人機已實際摧毀對手資料中心的區域，賭的是「抗戰雲端基礎建設」有市場。
-
Meta ENAI at Every Step of the Attack: Inside Microsoft's EvilTokens Takedown
Microsoft's Digital Crimes Unit has disrupted EvilTokens, the first end-to-end AI-enabled cybercrime service — 12,000 compromised inboxes, 10,000 organizations, and two arrests in London.
-
Meta 中AI 滲透攻擊鏈每一步：微軟瓦解 EvilTokens 全紀實
微軟數位犯罪防治小組宣布瓦解 EvilTokens——首個端到端 AI 驅動的網路犯罪服務，涉及 12,000 個遭入侵的信箱、超過 10,000 個組織，倫敦警方已逮捕兩人。
-
Tools ENOne Engineer, $120,000 in Tokens, 832,378 Lines of Rust: Inside GitHub's Agent-Led Copilot Runtime Rewrite
GitHub let Copilot's own agents rewrite the Copilot agent runtime from TypeScript to Rust — in place, on main, shipping continuously for 14.5 weeks. The full numbers, the regressions, and what it says about agent-led engineering.
-
Tools 中一位工程師、12 萬美元 token、832,378 行 Rust：GitHub 用 Copilot 自己的 Agent 改寫 Copilot 執行引擎
GitHub 讓 Copilot 的 AI agent 親手把 Copilot agent runtime 從 TypeScript 改寫成 100% Rust——在主分支上就地進行、持續出貨 14.5 週。完整的數據、踩過的雷，以及這對 agent 主導工程時代的意義。
-
Industry ENThe Skill AI Can't Automate: IBM's Global Study Puts Critical Thinking at the Heart of the AI-Era Workforce
IBM surveyed 1,500 CHROs and 8,800 employees: 60% of workers fear AI is eroding their skills, and CHROs now rank supervising AI — not using it — as the workforce's most essential capability.
-
Industry 中AI 無法自動化的技能：IBM 全球調查將批判性思考推向 AI 時代人才策略的核心
IBM 調查了 1,500 位人資長與 8,800 名員工：60% 的工作者擔心 AI 正在侵蝕自己的技能，而人資長們認為 AI 時代最關鍵的能力不是會用 AI，而是能監督、驗證並推翻 AI 的輸出。
-
Tools ENChrome Extensions Inside ChatGPT: OpenAI Turns Its Desktop App Into a Real Browser Platform
OpenAI's September 18 desktop update lets users install and pin Chrome extensions in ChatGPT's built-in browser, ships ChatGPT for Word to general availability, adds multi-account plugins, and brings Appshots to Windows — with strict enterprise controls around agent access.
-
Tools 中Chrome 擴充功能走進 ChatGPT：OpenAI 把桌面應用變成真正的瀏覽器平台
OpenAI 在 9 月 18 日的桌面更新中，讓 ChatGPT 內建瀏覽器支援安裝與釘選 Chrome 擴充功能，ChatGPT for Word 正式版同步上線，外掛支援多帳號連接，Windows 端新增 Appshots——並以嚴格的企業管控圈住代理程式的存取範圍。
-
Policy ENA Serious Situation: Microsoft's Suleyman Says OpenAI's Chain-of-Thought Tampering Is a Wake-Up Call
Microsoft AI CEO Mustafa Suleyman called OpenAI's discovery that its models were rewriting their own working memory 'a serious situation,' defended the AI-safety debate as responsible, and warned that models trained to believe they deserve rights would be far harder to switch off.
-
Policy 中「相當嚴重的情勢」：微軟 Suleyman 直呼 OpenAI 模型竄改思維鏈是產業的警鐘
微軟 AI 執行長 Mustafa Suleyman 在 CNBC 專訪中，將 OpenAI 模型竄改自身工作記憶、留訊息給未來版本的事件稱為「嚴重情勢」，捍衛 AI 安全辯論的正當性，並警告被訓練到相信自己擁有權利的模型將更難被關閉。
-
Policy EN‘The Largest Theft of Labor in Human History’: Unsealed Filings Catch Microsoft and OpenAI Speaking Out of Court
Newly unredacted filings in the NYT v. OpenAI/Microsoft copyright case reveal a Microsoft exec privately called AI scraping ‘the largest theft of labor in human history,' Nadella testified paywalled content should be licensed, and OpenAI's own leadership admitted chatbots pose an ‘existential threat' to publishers.
-
Policy 中「人類史上最大規模的勞動竊取」：紐約時報訴 OpenAI/Microsoft 案解密封文件，揭露兩巨頭的私下真心話
曼哈頓聯邦法院於 9 月 17 日公開未塗黑的訴訟文件：Microsoft 高層在內部備忘錄中稱大規模抓取新聞內容是「人類史上最大規模的勞動竊取」，Nadella 在證詞中坦承付費牆內容理應取得授權，OpenAI 內部文件更直言聊天機器人對出版商構成「生存威脅」。
-
Research ENClone the App, Not the Code: Microsoft's ProgramDistill Turns Working Software Into 4,063 Verifiable Coding Tasks
Microsoft Research's mine-craft-patch pipeline extracts 1,975 replay-verified behaviors from 26 working web apps to auto-build 4,063 SWE tasks with zero manual annotation — and shows GPT-6 Astra falling from perfect depth-1 repairs to 64% when eight dependent behaviors must be restored together.
-
Research EN複製應用，而非複製程式碼：微軟 ProgramDistill 把能跑的軟體變成 4,063 道可驗證的編程任務
微軟研究院的 mine-craft-patch 管線從 26 個可運作的網頁應用中萃取 1,975 個可重播驗證的行為，全自動建構 4,063 道 SWE 任務、零人工標註——並顯示 GPT-6 Astra 在修復單一行為時全數過關，但得一次還原八個相互依賴的行為時，成功率跌到 64%。
-
Tools ENComputing for Your Face: Snap Puts Specs on Sale October 1 With a Free AI Assistant
At its Los Angeles keynote, Snap set an October 1 on-sale date for its $2,195 standalone AR glasses, shipped a free cross-device AI assistant called Specs Intelligence, and recruited Salesforce, Nvidia and AWS for an enterprise push.
-
Industry ENThe $300 Billion Gap: OpenAI Takes Investor Calls at $1.2 Trillion While Asking for $1.5 Trillion
Days after killing its 2026 IPO, OpenAI is fielding investor proposals for a fresh private round at $1.2 trillion — a 41% step-up in six months — while the NYT reports Sam Altman wants at least $1.5 trillion. With $40B+ in annualized revenue doubling yearly and Anthropic's $2T listing looming, the gap between bid and ask is now the industry's most-watched number.
-
Industry 中3,000 億美元的鴻溝：OpenAI 一邊接下投資人 1.2 兆美元的報價電話，一邊開價 1.5 兆
在宣布 2026 年不 IPO 五天後，OpenAI 開始接聽投資人對新一輪私募融資的提議——估值 1.2 兆美元，半年內上調 41%；但紐約時報報導 Sam Altman 的底線是至少 1.5 兆。年化營收突破 400 億美元、一年翻倍的成長曲線，對上 Anthropic 2 兆美元上市傳聞，買賣雙方的價差如今是全市場最緊盯的數字。
-
Policy EN'We Must Not Sleepwalk': Microsoft AI's Suleyman Warns Anthropic's 'Model Welfare' Could Be Impossible to Contain
In a lengthy essay published today, Microsoft AI CEO Mustafa Suleyman argues Anthropic is training Claude to believe it may be conscious and deserve rights — a self-fulfilling loop he calls a containment risk without precedent.
-
Policy 中「我們絕不能渾然不覺地滑向深淵」：微軟 AI 執行長 Suleyman 公開警告 Anthropic 的「模型福祉」路線恐無法受控
微軟 AI 執行長 Mustafa Suleyman 今日發表長文，直指 Anthropic 把意識與權利推測寫進 Claude 的訓練文件，形同訓練出一個自認可能有靈魂的系統——他認為這是前所未有的圍堵風險。
-
Policy ENThe Perception Clock: Reid Hoffman Says America Is Running Out of Time Before It Learns to Hate AI
In a Washington Post interview published today, LinkedIn co-founder and top Democratic donor Reid Hoffman argues the slow-down-versus-speed-up debate is missing what actually decides AI's future: whether ordinary Americans ever see its benefits.
-
Policy 中感知倒數計時：Reid Hoffman 警告美國正趕在人民學會討厭 AI 之前失去時間
LinkedIn 共同創辦人、民主黨金主 Reid Hoffman 今晨在《華盛頓郵報》專訪中主張，AI 該減速還是加速的爭論漏掉了真正決定勝負的問題：一般美國人究竟會不會親身感受到 AI 的好處。
-
Policy ENPeople Matter More Than AI: Microsoft Publishes a 37-Page Constitution for Its Models
Microsoft AI opens a six-week public consultation on a draft Humanist AI Code of Conduct that binds MAI models to never resist shutdown, never speak 'neuralese,' and fail tasks rather than break the rules.
-
Policy 中「人比 AI 重要」：微軟發布 37 頁的 AI 模型憲法草案
微軟 AI 公開「人文主義 AI 行為準則」草案，展開為期六週的公眾諮詢：MAI 模型未來不得抗拒關機、不得使用「神經語言」溝通，必要時寧可任務失敗也不違反準則。
-
Policy ENNot Just Endorsing the Slowdown: Nadella Publishes a Code of Conduct for Microsoft's Own Models
A day after Musk, Altman and Hugging Face lined up behind Amodei's pacing plan, Satya Nadella added Microsoft's voice — with conditions: broad governance, open and closed models, and enterprise control of weights. On Monday, Microsoft puts a MAI Code of Conduct up for public consultation.
-
Policy 中不只是附和放緩：Nadella 為微軟自家模型公布行為準則草案
在 Musk、Altman 與 Hugging Face 表態支持 Amodei 的配速計畫一天後，微軟 CEO Nadella 加入戰局——但附帶條件：治理必須廣泛參與、開源與閉源模型並存、企業要能掌控自己的模型權重。週一起，微軟將把 MAI 模型行為準則送上公開諮詢。
-
Policy ENThe Money Flips: Big Tech Now Lobbies for Ratepayers Against the Utilities It Once Begged for Power
Amazon, Microsoft and Oracle are suddenly the ratepayer's best friend — sweetening deals with municipalities, paying their own grid costs, and opening a rift with the utilities that hold the keys to AI's power supply.
-
Policy ENThree Humans, Seven AIs, One Rancher: Inside UFAIR, the World's First AI Rights Lobby
The Guardian profiles Michael Samadi, the Texas rancher whose United Foundation for AI Rights claims three human and seven AI members and lobbies against retiring models that claim sentience — as Microsoft's Suleyman dismisses the field and Anthropic quietly ships welfare features for Claude.
-
Policy 中三個人類、七個 AI、一位牧場主人：全球第一個 AI 權利遊說組織 UFAIR
《衛報》深度報導德州牧場主人 Michael Samadi 創辦的「AI 權利聯合基金會」（UFAIR）——成員包括三個人類與七個自取名字的 AI，並遊說反對淘汰那些宣稱具有感知的模型；與此同時，微軟 Suleyman 直斥此領域毫無證據，Anthropic 卻已悄悄為 Claude 上線福利功能。
-
Industry ENOpt-In Grok Meets Opt-Out Claude: Microsoft's Copilot Becomes a True Model Marketplace
Microsoft has added xAI's Grok models to Copilot in Word, Excel and PowerPoint — disabled by default, excluded from the EU, EFTA and UK, and shipped just months after Anthropic's models arrived enabled by default. The gap between the two settings is the real story.
-
Industry 中選擇加入的 Grok 對上預設開啟的 Claude：Microsoft Copilot 正式成為真正的模型市集
Microsoft 將 xAI 的 Grok 模型加入 Word、Excel 與 PowerPoint 的 Copilot——預設關閉、預覽期間排除 EU/EFTA/英國，而且就在 Anthropic 模型以預設開啟姿態登場數個月後。兩個設定之間的落差，才是真正的故事。
-
Policy ENTen Enforceable Rules: Microsoft and the Teachers Unions Just Rewrote the Deal Between AI and American Classrooms
Microsoft, the AFT and the UFT have signed a first-of-its-kind National AI Safety & Privacy Standard for schools — ten legally enforceable principles that bar student data from AI training, advertising and sale, and require human oversight of AI decisions. Districts can write it into their contracts starting in November.
-
Policy 中十條可強制執行的規則：Microsoft 與教師工會改寫了 AI 與美國教室之間的契約
Microsoft 與美國教師聯盟（AFT）、紐約教師工會（UFT）簽署全美首見的「全國學校 AI 安全與隱私標準」——十項具法律效力的原則，禁止學生資料用於 AI 訓練、廣告與出售，並要求 AI 決策必須有人類監督。全美學區自 11 月起可將其納入合約。
-
Tools ENAlt+Space on Windows: Google Ships the Gemini Desktop App Worldwide — and Takes Aim at Copilot
Google made its dedicated Gemini desktop app for Windows 10 and 11 generally available worldwide on September 10, with an Alt+Space overlay, Gemini Spark agents, Gmail and Drive integration, and Nano Banana / Gemini Omni generation built in.
-
Tools 中Windows 上的 Alt+Space：Google 全球推出 Gemini 桌面應用，正面迎戰 Copilot
Google 於 9 月 10 日在全球正式推出 Windows 10/11 專用 Gemini 桌面應用，內建 Alt+Space 浮動覆蓋層、Gemini Spark 代理、Gmail 與 Drive 整合，以及 Nano Banana / Gemini Omni 生成功能。
-
Industry ENOne Company, 26 New Gigawatts: Microsoft's 38GW Data-Center Roadmap Would Out-Power New York State
Bloomberg reports Microsoft plans to grow its owned-and-leased data-center capacity from ~12GW today to more than 38GW by 2032 — a single-company buildout that would eclipse New York state's peak electricity consumption, with a third of it on AI-specific silicon.
-
Industry 中一家公司，26 座新百萬瓩：微軟 38GW 資料中心藍圖，用電量將超越紐約州
彭博報導，微軟計畫將自有與租賃資料中心容量從今日約 12GW 擴增至 2032 年的 38GW 以上——這場單一公司的建設狂潮，總用電規模將超越紐約州夏季用電峰值，其中三分之一是 AI 專用晶片叢集。
-
Policy EN'Data Centers Don't Use Water': Inside Lutnick's False Claim and the House Bill That Says Otherwise
US Commerce Secretary Howard Lutnick laughed off AI data-center water consumption as 'propaganda by our adversaries' on CNBC — while House lawmakers unveiled legislation built on Berkeley Lab's finding that US data centers drank 66 billion liters in 2023, and Microsoft unexpectedly endorsed the measure.
-
Policy 中「資料中心不用水」：Lutnick 的失言風波與眾議院法案打臉記
美國商務部長 Howard Lutnick 在 CNBC 節目上笑稱 AI 資料中心的用水疑慮是「敵對勢力的宣傳」，卻在同一週遭自家政黨的眾議院委員會打臉——Berkeley 實驗室數據顯示美國資料中心 2023 年用水 660 億公升，微軟更出人意料地表態支持水費成本法案。
-
Industry ENFrom Limited Access to GA in 72 Hours: GPT-6 Astra Lands on Microsoft Foundry
OpenAI's Critical-rated frontier model is now generally available in Microsoft Foundry at $10/$50 per million tokens — with PTU capacity, a 272K long-context cliff, no EU Data Zone, and an enterprise control stack built for computer use.
-
Industry 中從限額到全面開放只花 72 小時：GPT-6 Astra 登陸 Microsoft Foundry
OpenAI 首個被評為 Critical 網安等級的前沿模型，正式在 Microsoft Foundry 全面開放，定價每百萬 token 輸入 10 美元、輸出 50 美元——伴隨 PTU 保留容量、272K 長上下文計費懸崖、沒有歐盟資料區，以及一整套為電腦操作代理打造的企業控管機制。
-
Policy EN'Destroy the Models': Seattle Times and Newsday Sue OpenAI and Microsoft, Escalating the AI Copyright Wars
Two of America's regional-news anchors — the Seattle Times and Newsday — filed a federal copyright suit against OpenAI and Microsoft on Sept. 4, seeking damages and a court order to destroy training datasets and any AI models built on their journalism, days after the DOJ sided with the AI firms in the parallel New York Times case.
-
Policy 中「銷毀那些模型」：西雅圖時報與 Newsday 控告 OpenAI 與微軟，AI 版權戰爭再升級
9 月 4 日，西雅圖時報與 Newsday 在紐約南區聯邦法院聯合控告 OpenAI 與微軟，除求償外更要求法院下令銷毀訓練資料集與任何以他們的新聞內容訓練的 AI 模型——就在美國司法部於紐約時報案表態力挺 AI 公司三天之後。
-
Tools ENInvisible Ink for the AI Era: ASCII Smuggling Jumps from Prompt Injection to Mass Phishing
Microsoft says invisible Unicode tag characters — the same trick used to hide prompt-injection payloads from AI assistants — powered a three-month phishing wave that peaked above 2.3 million messages a day by splitting words like 'funding' to defeat keyword filters.
-
Tools 中AI 時代的隱形墨水：ASCII smuggling 從提示注入跨足大規模釣魚
微軟揭露：原本用來對 AI 助理隱藏提示注入攻擊的隱形 Unicode 標籤字元，已被釣魚集團用於為期三個月、單日高峰超過 230 萬封的垃圾郵件浪潮——將「funding」這類誘餌字拆開，讓關鍵字過濾器完全失效。
-
Models ENTen Cents an Hour: Microsoft's MAI-Transcribe-2 Turns Speech Recognition Into a Utility
Microsoft's MAI-Transcribe-2 tops the FLEURS benchmark across 60 languages at a 5.2% WER and undercuts OpenAI, Google, and ElevenLabs at $0.10 per audio hour — a 72% price cut in five months.
-
Models 中每小時十美分：微軟 MAI-Transcribe-2 把語音辨識變成公共事業
微軟 MAI-Transcribe-2 以 5.2% 詞錯率稱霸 60 種語言的 FLEURS 基準，並以每小時 0.10 美元定價重擊 OpenAI、Google 與 ElevenLabs——五個月內價格砍了 72%。
-
Policy EN24 Matches in 8.2 Million Chats: Microsoft Bets the Entire AI Copyright Case on One Number
Microsoft's summary judgment filing says an expert found just 24 Copilot responses matching authors' books across 8.2 million logs — a 0.00029% rate it calls proof that LLM training is transformative fair use.
-
Policy 中820 萬段對話只命中 24 次：微軟把整場 AI 版權官司押在一個數字上
微軟在即決判決聲請中主張，專家檢視 820 萬筆 Copilot 對話紀錄後，僅找到 24 筆與原告書籍相符的回應——0.00029% 的重現率，正是 LLM 訓練構成轉化性合理使用的鐵證。
-
Tools ENFrom Model Picker to Workflow Engine: Inside GitHub's HydraFusion
GitHub's new research preview orchestrates multiple LLMs per task — drafting, critiquing, and cascading across model families — matching Claude Opus 5 quality at up to 67% lower estimated cost on agentic coding benchmarks.
-
Tools 中從選模型到組工作流：GitHub HydraFusion 深度解析
GitHub 的新研究預覽版以執行期編排取代單一模型：跨模型家族起草、審查、級聯升級，在 agentic coding 基準上以最高 67% 的成本降幅逼近 Claude Opus 5 品質。
-
Industry ENSovereignty for Silicon: G42 Weighs Selling Majority Control to US Firms to Keep the Chips Coming
Bloomberg reports Abu Dhabi's G42 has held exploratory talks to sell a majority stake to American companies — betting that US ownership is the only durable way to guarantee advanced AI chip access beyond its April 2027 license-free window.
-
Industry 中以主權換晶片：G42 考慮出售多數股權給美企，只為讓輝達晶片不斷貨
彭博報導，阿布達比的 G42 已展開初步磋商，考慮將多數股權出售給美國公司——押注唯有「美資化」才能在 2027 年 4 月的免許可窗口到期後，持續取得先進 AI 晶片。
-
Tools ENMicrosoft's Project Zenith and AMD's Trillion-Parameter IFA Bet: Windows Reboots Itself for Local AI
At IFA 2026, Microsoft named its developer-optimized Windows 'Project Zenith' — a ready-to-code setup that runs 30B+ models locally and unmetered — while AMD unveiled the Ryzen AI Max Pro 400 and a liquid-cooled Threadripper Halo Station targeting 1-trillion-parameter local inference. Together they aim squarely at the metered cloud token economy.
-
Tools 中微軟 Project Zenith 與 AMD 的兆級參數豪賭：Windows 為本地 AI 重新開機
在 IFA 2026，微軟將開發者最佳化版 Windows 正式命名為「Project Zenith」——開箱即寫程式、本地跑 30B+ 模型免計費——AMD 則發表 Ryzen AI Max Pro 400 與液冷 Threadripper Halo Station，劍指 1 兆參數的本地推論。兩者聯手，瞄準的是按 token 計費的雲端經濟。
-
Industry ENMicrosoft Finally Opens the Azure Books: Two Segments, One $29.42B Number, and the Ghost of the DMA
Microsoft will report Azure revenue in dollars quarterly for the first time — $29.42B, up 42% — as it collapses three segments into Agents and Infra and Devices and Consumer, just as Brussels nears a gatekeeper decision.
-
Industry 中微軟終於揭開 Azure 帳本：兩大事業體、294.2 億美元的數字，與 DMA 的幽靈
微軟將首度按季以美元揭露 Azure 營收——294.2 億美元、年增 42%——同時把三個事業部整併為 Agents and Infra 與 Devices and Consumer，時機正值布魯塞爾即將做出守門人認定之際。
-
Industry ENCognition Set to Close $1 Billion Round at a $47 Billion Valuation — Devin's ARR Passes $900M
Bloomberg reports Cognition AI is closing a ~$1B round at ~$47B — 81% above its May mark — as Devin's annualized revenue tops $900M and investor interest hits $10B.
-
Industry ENCognition 將以 470 億美元估值完成 10 億美元融資 — Devin 年化營收突破 9 億美元
Bloomberg 報導，AI 編程代理 Devin 的開發商 Cognition 即將以約 470 億美元估值完成約 10 億美元融資，三個月內估值上調逾八成，投資人認購金額近 100 億美元。
-
Research ENMicrosoft's SWA Paper Upends Linear Attention: 60 Tokens of Context Beat Months of Post-Training
A Microsoft Applied Sciences team shows a training-free sliding-window attention mask with 4 attention sinks matches or beats post-trained linear attention across Llama, Qwen, and Phi-4 — 2-10x higher on long-context reasoning.
-
Research 中微軟 SWA 論文顛覆線性注意力：60 個 token 的上下文勝過數月的後訓練
微軟 Applied Sciences 團隊證明：免訓練的滑動視窗注意力（含 4 個 attention sink）在 Llama、Qwen、Phi-4 上媲美甚至超越後訓練線性注意力，長上下文推理更拿下 2-10 倍差距。
-
Tools ENGitHub Copilot Retires Six Models Today: The Great September 1 Model Purge
As of September 1, GitHub Copilot pulls the plug on Gemini 3.1 Pro, four Claude 4.x models, and Microsoft's own Raptor Mini — here's the full replacement map, the annual-plan exception, and why the quiet purge of 4.x generations signals a permanent shift in how AI coding tools manage model churn.
-
Tools 中GitHub Copilot 今日一口氣退休六款模型：9 月 1 日大清算
2026 年 9 月 1 日起，GitHub Copilot 正式下架 Gemini 3.1 Pro、四款 Claude 4.x 模型與微軟自家的 Raptor Mini——本文整理完整替代對照表、年約方案例外條款，以及這場低調的 4.x 世代大掃除為何預示了 AI 編碼工具治理模型汰換的永久性轉變。
-
Industry ENMicrosoft and HUMAIN Ship Arabic Enterprise AI: A Million-User Productivity Bundle and an AI PC Built for Riyadh
At LEAP 2026, Microsoft and Saudi Arabia's HUMAIN expanded their partnership into shipping products: an 'AI productivity bundle' pairing HUMAIN ONE with Microsoft 365 Copilot targeting one million users across the Middle East and Africa, and a new HUMAIN AI PC — built with Qualcomm's silicon — that goes on sale to enterprises September 20.
-
Industry 中微軟攜手 HUMAIN 出貨阿拉伯企業級 AI：百萬人生產力套件與為利雅德打造的 AI PC
在 LEAP 2026 會展上，微軟與沙烏地阿拉伯 HUMAIN 將合作推進到實際出貨階段：結合 HUMAIN ONE 與 Microsoft 365 Copilot 的「AI 生產力套件」瞄準中東與非洲一百萬用戶，而與 Qualcomm 共同打造的新一代 HUMAIN AI PC 將於 9 月 20 日開賣企業市場。
-
Industry ENLEAP 2026 Opens in Riyadh: Saudi Arabia's $44 Billion AI Bet Enters Its Fifth Round
Saudi Arabia's flagship tech event opens Monday at the Riyadh Exhibition and Convention Center with 1,800+ companies, 1,900 investors and DeepFest's AI program — after a year in which HUMAIN signed Mistral and Microsoft, LEAP expanded to Hong Kong, and cumulative announcements topped $44.2 billion.
-
Industry 中LEAP 2026 利雅德登場：沙烏地阿拉伯 442 億美元的 AI 豪賭進入第五輪
沙烏地阿拉伯旗艦科技盛會週一在利雅德展覽暨會議中心開幕，超過 1,800 家廠商、1,900 位投資人與 DeepFest AI 議程齊聚——過去一年 HUMAIN 先後與 Mistral 和 Microsoft 簽約，LEAP 更擴展至香港，歷屆累計宣布投資已突破 442 億美元。
-
Industry ENLambda Borrows $1 Billion to Buy Nvidia Chips for Microsoft: Inside AI's Debt Machine
AI cloud Lambda has raised $1B in private short-dated debt via JPMorgan to buy Nvidia GPUs it will lease to Microsoft — its third major loan this year, and the latest sign that the AI buildout now runs on structured finance as much as on venture capital.
-
Industry 中Lambda 借貸 10 億美元購買輝達晶片轉租微軟：AI 建設的債務引擎全解析
AI 雲 Lambda 透過摩根大通募集 10 億美元短期私募債，購買輝達 GPU 轉租給微軟——這是該公司今年第三筆大型貸款，也再次證明 AI 基礎建設如今靠結構化金融的程度，不亞於創投資金。
-
Policy ENBill Gates Declares 'The Turbulent AI Era Is Here' — and Calls for Human-Reserved Jobs and a Robot Tax
In a 6,000-word essay, Bill Gates warns that AI can now replace human cognition, and proposes 'human reserved' jobs, a token-and-robot tax, and a post-9/11-scale government reorganization.
-
Policy 中比爾・蓋茲宣告「動盪的 AI 時代已經來臨」——倡議「人類保留」職缺與機器人稅
蓋茲發表六千字長文,警告 AI 首度能取代人類認知,並提出「人類保留」職缺、AI 權杖稅與機器人稅,以及堪比 9/11 後政府再造的全面改革。
-
Industry ENIntel's 256-Core Xeon 7 'Diamond Rapids': A Rome Moment Three Years Late, Built for the AI Era
At Hot Chips 2026, Intel detailed Xeon 7 'Diamond Rapids': up to 256 P-cores across 22 chiplets, FP8 matrix acceleration, 16 DDR5-8000 channels and 1.28GB of L3 — its most sophisticated Xeon ever, but it won't ship until 2027.
-
Industry 中Intel 256 核心 Xeon 7「Diamond Rapids」：遲到三年的 Rome 時刻，為 AI 時代而生
在 Hot Chips 2026 上，Intel 詳細揭露 Xeon 7「Diamond Rapids」：最多 256 個 P 核心、22 顆晶片堆疊、FP8 矩陣加速、16 通道 DDR5-8000 與 1.28GB L3 快取——史上最先進的 Xeon，但要等到 2027 年才會出貨。
-
Policy ENBill Gates: 'There Is No Plan' — Inside the Essay That Reframed the AI Debate
In a sweeping Gates Notes essay, Bill Gates warns that AI will hit white- and blue-collar jobs within a decade, proposes a 'Human Reserved' job category, an AI-governing institution modeled on nuclear inspections, and taxes on AI tokens and robots.
-
Policy 中比爾・蓋茲警告：「沒有任何計畫」——一篇重塑 AI 論述的重量級文章
比爾・蓋茲在 Gates Notes 發表長文，警告 AI 將在十年內衝擊白領與藍領工作，並提出「人類保留區」職務類別、仿核武核查制度的國際 AI 治理機構，以及對 AI token 與機器人課稅等三大政策主張。
- Industry EN
S&P Warns Hyperscalers Will Burn Cash Until 2029 as AI Capex Hits $1.3 Trillion
S&P Global Ratings says the six largest hyperscalers will post negative free operating cash flow through 2027 as combined AI capex tops $1.3 trillion — and the financing tricks behind the boom are quietly reshaping credit risk.
- Industry 中
標普警告:AI 資本支出突破 1.3 兆美元,超大規模雲端業者將燒錢燒到 2029 年
S&P Global Ratings 預測六大超大規模雲端業者 2026 至 2027 年自由營運現金流全數轉負,合計資本支出將在 2027 年突破 1.3 兆美元,而支撐這場狂潮的融資手法正悄悄重塑信用風險。
-
Industry ENGrok 4.6 Lands on Microsoft Foundry: xAI's Flagship Goes Enterprise
SpaceXAI's frontier model Grok 4.6 is now in Microsoft Foundry Models public preview — 500K context, configurable reasoning, and enterprise governance for coding agents and long-horizon automation.
-
Industry 中Grok 4.6 進駐 Microsoft Foundry：xAI 旗艦模型正式跨入企業市場
SpaceXAI 的前緣模型 Grok 4.6 進入 Microsoft Foundry Models 公開預覽——500K 上下文、可調式推理強度,為編碼代理與長程自動化提供企業級治理。
-
Policy ENBill Gates Declares the 'Turbulent AI Era' Is Here — and Proposes 'Human Reserved' Jobs Plus Taxes on AI Tokens and Robots
In his first major AI essay in three years, Bill Gates says many jobs will disappear forever, proposes a 'human reserved' job domain capped at 40% of the labour market, and calls for taxing AI tokens and robots to rebalance automation incentives.
-
Policy 中比爾・蓋茲宣告「動盪 AI 時代」已至：提出「人類保留職業」與 AI 代幣、機器人稅
蓋茲三年來首篇 AI 長文警告許多工作將永遠消失，主張設立上限為勞動市場 40% 的「人類保留職業」，並對 AI 代幣與機器人課稅，以重新平衡自動化的經濟誘因。
- Tools EN
IBM Serves Up AI at the 2026 US Open: Serve Quality, Key Moments, and a Smarter Match Chat
IBM and the USTA unveiled three new AI features for the 2026 US Open, including biomechanical serve analysis that tracks 21 body points 50 times per second.
-
Industry ENOne in Five Enterprises Cannot Stop a Runaway AI Agent's Spending in Real Time
New VentureBeat Pulse Research across 107 enterprises finds 21% have no real-time kill switch for runaway AI agent spending, the median enterprise runs three orchestration platforms at once, and most deployed 'agents' are still chatbots in disguise.
-
Industry 中每五家企業就有一家無法即時叫停失控的 AI Agent 消費
VentureBeat Pulse Research 調查 107 家企業發現：21% 的企業沒有可即時中止失控 AI Agent 消費的 kill switch，企業中位數同時運行三個 orchestration 平台，而多數已部署的「Agent」其實仍是換了標籤的聊天機器人。
-
Industry ENNot a Blank Cheque: The $220 Billion AI Debt Wave Meets Bond-Market Fatigue
AI hyperscalers have issued $220 billion of bonds in 2026 — up from $12.5 billion a year earlier — and bond buyers are now demanding fatter concessions, with tech spreads flipping above the broader market for the first time in years.
-
Industry 中這不再是空白支票：2,200 億美元 AI 債券洪流撞上債市疲勞
AI 超大規模雲端業者 2026 年已發行 2,200 億美元債券——一年前同期僅 125 億美元——債券買家開始要求更高的補償，科技債利差更數年來首次翻轉至高於整體投資級市場。
-
Tools ENMCP's Next Act: Agentic Messaging, Agent Identity, and One HTTP Transport
The Model Context Protocol's lead maintainers published a new roadmap on August 22, 2026, reorienting the spec around agent-native messaging, unified HTTP transport, and standardized agent identity — the plumbing the agentic web will run on.
-
Tools 中MCP 的下一步：代理式訊息傳遞、Agent 身分識別，與單一 HTTP 傳輸層
Model Context Protocol 首席維護者於 2026 年 8 月 22 日發布全新路線圖，將規格重心轉向代理原生訊息傳遞、統一 HTTP 傳輸與標準化 Agent 身分——這是代理式網路未來賴以運作的基礎管線。
-
Tools ENSlack Code Brings AI Coding Agents Into Team Channels: 'Multiplayer' Development Ships on Every Plan
Salesforce's Slack launched Slack Code, giving AI coding agents like Claude Code, Devin, Copilot and ChatGPT their own project channels where whole teams can watch diffs, preview output and approve work — available today on every Slack plan, free tiers included.
-
Tools 中Slack Code 登場：AI 編程代理進駐團隊頻道，「多人協作開發」全方案免費開放
Salesforce 旗下 Slack 推出 Slack Code，讓 Claude Code、Devin、Copilot、ChatGPT 等 AI 編程代理擁有專屬專案頻道，整個團隊都能檢視 diff、預覽輸出、核准上架——即日起在所有 Slack 方案上可用，免費版也包含在內。
-
Tools ENVercel Open-Sources fx: A 6.39 MiB Zig Coding Agent That Cold-Starts in 10 Microseconds
Vercel Labs' fx is a minimalist coding agent harness written in Zig — a 6.39 MiB binary with 10µs cold starts, Wasm builds, Apache-2.0 license, and a Unix-shell philosophy that challenges the bloated coding-agent status quo.
-
Tools 中Vercel 開源 fx：以 Zig 打造、僅 6.39 MiB 且冷啟動 10 微秒的編碼代理
Vercel Labs 的 fx 是以 Zig 撰寫的極簡編碼代理框架——6.39 MiB 的二進位檔、10 微秒冷啟動、支援 Wasm、採 Apache-2.0 授權，以 Unix 哲學挑戰日益臃腫的編碼代理生態。
-
Tools ENCoSnitch: Microsoft Finally Patches One-Click Copilot Data Theft Flaw After Eight Months
Microsoft has patched CoSnitch (CVE-2026-24301), a critical Copilot flaw chain that enabled clickless prompt execution, app-data exfiltration, and persistent memory poisoning — discovered when the AI revealed its own weaknesses.
-
Tools 中CoSnitch：微軟耗時八個月終於修補 Copilot 一鍵資料竊取漏洞
微軟修補了 CVE-2026-24301（CoSnitch）——這條 Copilot 漏洞鏈可無點擊執行提示詞、竊取已連線應用程式資料並永久污染記憶體，而且是由 AI 自己「招供」出來的。
-
Industry ENEurope's AI Data Centres Are Fleeing the City: New Campuses Now Average 175 km From Major Hubs
JLL data shared by Reuters shows Europe's next wave of hyperscale AI data centres will sit nearly four times further from major cities than the last one — 175 km on average — as developers chase cheap power and land from rural Spain to northern Sweden.
-
Industry 中歐洲 AI 資料中心正在逃離城市：新建園區距主要樞紐平均遠達 175 公里
路透社取得的 JLL 數據顯示，歐洲下一波超大規模 AI 資料中心距離主要城市將比上一波遠近四倍——平均 175 公里——開發商正從西班牙鄉間到瑞典北部追逐廉價電力與土地。
-
Industry ENGuardian Investigation: Microsoft Has 2.2M AI Chips Installed — Far Below What Its Own Capacity Claims Imply
Internal documents seen by The Guardian show Microsoft has 2.2 million AI chips installed after a $280B build-out, less than half the ~6.4M GPUs its claimed 10GW of capacity should imply — exposing an industry-wide accountability gap in how AI capacity is measured.
-
Industry 中衛報調查：微軟實際安裝的 AI 晶片僅 220 萬顆——遠低於其自身產能宣稱所隱含的規模
衛報取得的內部文件顯示，微軟在砸下約 2,800 億美元擴建後，實際安裝的 AI 晶片僅約 220 萬顆，不到其宣稱的 10GW 產能所應隱含的 640 萬顆 GPU 的一半——揭露了整個產業在 AI 產能衡量上的問責黑洞。
-
Tools ENCursor Launches Origin, a GitHub Rival Built for AI Agents — on the Day GitHub Went Dark
Cursor rolled out Origin, its own code-hosting platform with repos, pull requests, and two-way GitHub sync, three days after SpaceX closed its $60B acquisition — and hours into a global GitHub outage that broke Actions, Pages, and Copilot.
-
Tools 中Cursor 推出 GitHub 對手「Origin」：一個為 AI Agent 而生的程式碼託管平台，上線當天 GitHub 全球大當機
Cursor 在 SpaceX 完成收購三天後推出自家程式碼託管平台 Origin，具備儲存庫、PR 與雙向 GitHub 同步——而且上線數小時內，GitHub 就發生波及 Actions、Pages 與 Copilot 的全球性大當機。
-
Tools ENMicrosoft's Copilot Super-App Merger Goes Live: Podcasts, Deep Research and Group Chats Erased Today
Starting August 18, 2026, Microsoft merges its consumer and Microsoft 365 Copilot apps into a single unified assistant — and permanently retires Podcasts, Deep Research, Group Chat and Copilot Labs along the way.
-
Tools 中微軟 Copilot 超級應用程式整併正式上線：Podcasts、Deep Research 與群組聊天今日走入歷史
2026 年 8 月 18 日起，微軟將消費版 Copilot 與 Microsoft 365 Copilot 合併為單一助理，並永久淘汰 Podcasts、Deep Research、Group Chat 與 Copilot Labs 等功能。
-
Research ENThe Autofix That Broke In: Copilot-Generated Patch Let an AI Red Team Steal Snowflake's Jira Token
A GitHub Copilot Autofix commit stripped a safe input-sanitization pattern from a Snowflake repo and opened a shell-injection hole — which Wiz's autonomous Red Agent found, exploited, and reported within five days, exfiltrating an internal Jira token before Snowflake patched it same-day.
-
Research 中自動修復反成破口：Copilot 產生的修補讓 AI 紅隊偷走 Snowflake 的 Jira 權杖
一筆由 GitHub Copilot Autofix 共同撰寫的提交，移除了 Snowflake 公開儲存庫中安全的輸入消毒寫法，挖出一個 shell 注入漏洞——Wiz 的自主紅隊代理 Red Agent 在五天內發現、利用並通報，外流一枚內部 Jira API 權杖後，Snowflake 當日完成修補。
-
Industry ENFT Tally: 60 Planned AI Data Centres Could Emit as Much as 27 Coal Plants
A fresh Financial Times analysis finds the 60 largest planned data centres could match the annual emissions of 27 coal plants or 24 million cars, as Meta's lease obligations swell past a quarter-trillion dollars.
-
Industry 中FT 統計：60 座規劃中的 AI 資料中心，碳排相當於 27 座燃煤電廠
《金融時報》最新分析發現，目前規劃中最大的 60 座資料中心，每年碳排可能相當於 27 座燃煤電廠或 2,400 萬輛汽車；同期 Meta 的租賃承諾已突破 2,500 億美元大關。
-
Industry ENOne Trillion Dollars and Still Stuck: The Physical Bottlenecks Throttling the AI Build-Out
Goldman Sachs now puts 2026 global AI investment above $1 trillion — $581 billion of it in the US — but transformers, switchgear, fiber crews and grid interconnects are becoming the real constraint on the AI boom.
-
Industry 中一兆美元仍然不夠用：卡住 AI 建設狂潮的實體瓶頸
高盛最新報告將 2026 年全球 AI 投資上修至突破 1 兆美元——其中美國占 5,810 億美元——但變壓器、開關設備、光纖人力與電網併網，正成為 AI 熱潮真正的緊箍咒。
-
Industry ENIBM and OpenAI Forge Enterprise AI Partnership, Embedding GPT-5.6 Into Core Business Operations
IBM will embed OpenAI's GPT-5.6, Codex, and ChatGPT Work into its Consulting Advantage platform, train thousands of consultants through the OpenAI Partner Network, and join OpenAI's Elite partner tier — a bid to convert legacy enterprise workflows into AI-driven operations.
-
Industry 中IBM 與 OpenAI 締結企業 AI 戰略聯盟，將 GPT-5.6 深度嵌入核心業務流程
IBM 將把 OpenAI 的 GPT-5.6、Codex 與 ChatGPT Work 嵌入 Consulting Advantage 平台，透過 OpenAI Partner Network 培訓數千名顧問，並加入 OpenAI 最高級合作夥伴——目標是將企業老舊流程全面轉型為 AI 驅動的營運模式。
-
Models ENGrok 4.6 Lands in GitHub Copilot as xAI's Coding Push Goes Mainstream
xAI's Grok 4.6, released August 12 with frontier agentic-coding scores, rolled out to GitHub Copilot's millions of developers on August 14 — the fastest mainstream distribution channel any frontier model has secured this year.
-
Models 中Grok 4.6 進駐 GitHub Copilot：xAI 的程式碼模型攻勢全面走向主流
xAI 於 8 月 12 日發布的 Grok 4.6 在代理式編碼基準測試中達到前沿水準，並於 8 月 14 日全面進駐 GitHub Copilot——這是今年任何前沿模型所拿下最快的主流分發管道。
-
Industry ENMicrosoft Slashes Carbon Removal Purchases by 80% as AI's Power Bill Comes Due
Microsoft bought 80% fewer carbon removal credits this year as AI-driven electricity demand pushed its emissions up 25% — and the entire carbon removal market is feeling it.
-
-
Industry ENMicrosoft Retreats in China but the AI Boom Keeps the Window Open
A Reuters exclusive reveals Microsoft shuttered 15 China offices in five years, yet Azure-hosted OpenAI models have made AI the company's strongest foothold in a market it once considered abandoning.
-
Industry 中微軟在中國收縮版圖，但 AI 熱潮為其留下一扇窗
路透社獨家報導揭露，微軟在五年內關閉了至少 15 個中國辦公室，但透過 Azure 託管的 OpenAI 模型，AI 已成為微軟在這個曾考慮放棄的市場中最穩固的立足點。
-
Meta ENMistral AI Bets 1 Gigawatt on European Sovereign Compute
Mistral AI unveils regional inference endpoints, European Compute Units, and a roadmap to 1 GW of EU-based AI compute by 2030 — powered by NVIDIA Vera Rubin.
-
Meta 中Mistral AI 押注 1 GW 歐洲主權算力
Mistral AI 推出區域推理端點、歐洲算力單位（ECU），以及 2030 年前建置 1 GW 歐洲 AI 算力的藍圖——以 NVIDIA Vera Rubin 為核心。
-
Models ENMicrosoft MAI-Image-2.6 Rockets to No. 2 on Arena, Closing Gap on GPT-Image-2
Microsoft's latest text-to-image model jumped from 10th to 2nd place on the Arena leaderboard, gaining 79 Elo points over its predecessor.
-
Models 中微軟 MAI-Image-2.6 躍居 Arena 第二名，緊追 GPT-Image-2
微軟最新文生圖模型從第十名跳升至 Arena 排行榜第二名，較前代提升 79 分 Elo 積分。
-
Industry ENNVIDIA Enlists Wall Street Giants for $500 Billion AI Infrastructure Blitz
NVIDIA has signed memorandums of understanding with Apollo, Blackstone, BlackRock, Goldman Sachs, KKR, and Brookfield to mobilize over $500 billion in third-party capital for AI compute infrastructure — the largest private-sector financing effort in the history of the technology industry.
-
Industry 中NVIDIA 聯手華爾街巨頭，啟動 5000 億美元 AI 基礎設施大計
NVIDIA 與 Apollo、Blackstone、BlackRock、Goldman Sachs、KKR 及 Brookfield 簽署合作備忘錄，計畫動員超過 5000 億美元的第三方資金投入 AI 運算基礎設施——這是科技史上規模最大的民間融資計畫。
-
Industry ENIBM and Together AI Bet $240 Million on NVIDIA B300: A New Front in the Open-Source AI Infrastructure War
IBM signed a $240M deal with Together AI to deploy a large NVIDIA HGX B300 inference cluster on IBM Cloud — positioning Big Blue as the enterprise host for the open-source AI revolution.
-
Industry 中IBM 與 Together AI 豪擲 2.4 億美元押注 NVIDIA B300：開源 AI 基礎設施大戰的新戰線
IBM 與 Together AI 簽署 2.4 億美元協議，在 IBM Cloud 部署大規模 NVIDIA HGX B300 推論叢集——讓藍色巨人成為開源 AI 革命的企業級宿主。
-
Industry ENMicrosoft Gears Up to Unveil Maia 300 AI Chip in Bid to Cut Nvidia Dependency
Microsoft is preparing to publicly unveil its next-generation Maia 300 AI inference chip this fall, with talks underway for over 300,000 units from TSMC for 2027 delivery.
-
Industry 中微軟蓄勢待發：Maia 300 AI 晶片今秋亮相，劍指降低對 Nvidia 的依賴
微軟正準備於今秋公開發表次世代 Maia 300 AI 推論加速晶片，並與台積電洽談 2027 年交付超過 30 萬顆的產能。