Report 2026-06-03 178 sources Input Tokens: 106,534 · Output Tokens: 4,621 · Total Tokens: 113,365

AI 戰情報告 (2026-06-03)

microsoftnvidiadeepseekanthropicopenaigithubgoogle mai-thinking-1mai-code-1-flashclaude-mythosgemini-spark agenton-device-aifinopsquantum-computingcybersecurity

AI 戰情報告 (2026-06-03)

昨日 AI 重點新聞

微軟發表 MAI-Thinking-1 萬億參數推理模型與多款 MAI 系列模型 ref-11, ref-31, ref-2, ref-3

微軟在 Build 大會推出自主研發的 MAI 模型家族，旗艦推理模型 MAI-Thinking-1 採用 1 兆參數 MoE 架構（35B 活化參數），並強調使用乾淨且合法授權的數據訓練。同步推出的 MAI-Code-1-Flash（137B 參數）專為 GitHub Copilot 優化，旨在提升效能並降低企業端推理成本。

Microsoft Scout 與 Project Solara：從個人助理轉向 Agentic OS ref-7, ref-146, ref-39, ref-144, ref-150

微軟推出基於 OpenClaw 的個人辦公助理 Scout，能主動處理行事曆衝突、會議準備及報帳等庶務。同時發表的 Project Solara 是專為 AI 代理硬體設計的 Android 基礎操作系統，現場展示了智慧桌面顯示器與可穿戴式識別證兩款概念裝置。

NVIDIA RTX Spark 平台發表：開啟 1 Petaflop 的個人 AI PC 時代 ref-103, ref-117, ref-4, ref-162, ref-122

NVIDIA 攜手微軟推出 RTX Spark 平台，為 Windows PC 提供高達 1 Petaflop 的 AI 算力，使 1 兆參數模型能在本地運行。Surface RTX Spark Dev Box 作為首款開發者專用迷你 PC，旨在讓開發者在隱私受控的環境下開發長時運行的自主代理。

DeepSeek 獲騰訊與寧德時代投資，融資額達 74 億美元 ref-159, ref-67

中國領先的模型實驗室 DeepSeek 正進行規模達 74 億美元 的首輪融資，其估值落在 520 億至 590 億美元 之間。投資方包括騰訊、寧德時代等產業巨頭，這筆巨資將用於強化其模型開發與與美國領先實驗室的競爭能力。

Anthropic 擴張 Glasswing 合作，Claude Mythos 保護全球關鍵基礎設施 ref-45, ref-155, ref-154, ref-166

Anthropic 宣布將其 Project Glasswing 資安計畫擴展至 150 家新機構，涵蓋能源、水力與醫療領域。該計畫提供強大的 Claude Mythos 模型，已在現有操作系統中揪出超過 10,000 個高危漏洞，預防可能影響全球上億人的災難性攻擊。

OpenAI Codex 針對知識工作者推出「站點」與「標註」功能 ref-18, ref-85, ref-44

OpenAI 的 Codex 應用程式現在不僅服務工程師，20% 的週活躍用戶已為知識工作者。新推出的 Sites 讓用戶能一鍵生成互動網站，而 Annotations 則能針對投影片或試算表中的特定區塊進行精確的 AI 操作與數據擷取。

Uber 實施 AI 支出上限，每位員工每月限額 1,500 美元 ref-157, ref-36

因工程師過度頻繁使用 AI 編程助手導致預算在四個月內超支，Uber 宣布對 AI 工具的使用成本進行限額。這標誌著企業正從 AI 的「盲目導入期」轉向強調 FinOps (雲端財務營運) 的精確治理階段。

GitHub Copilot 改採 Token 計費制，並重構架構應對 30 倍流量 ref-11, ref-3

GitHub 正式將 Copilot 的計費模式轉為基於 AI Credits 的使用量計費。COO Kyle Daigle 指出，代理生成的代碼量在 2026 年增長了 1400%，GitHub 必須重新編寫底層分發系統，以處理比人類開發者高出 30 倍以上的提交量。

Google Gemini Spark 展現強大自主規劃能力 ref-162

實測顯示 Google Gemini Spark 代理能深度整合用戶各項資訊，完成極度複雜的旅行規劃。不同於傳統聊天機器人，其展現出的「主動搜尋」與「自主排程」能力，被評為目前最令人印象深刻的代理體驗。

川普簽署自願性 AI 行政命令，傾向保護創新 ref-88, ref-106, ref-151, ref-153

美國總統川普簽署一份較為寬鬆的 AI 行政命令，取消了強制性的政府審核，改為建立「自願性框架」。該命令強調不應以過度規管阻礙美國在 AI 領域的領先地位，並要求聯邦機構針對 AI 資安風險進行風險評估。

MIT 發表 ChartNet：教導視覺語言模型精準解讀圖表 ref-58

MIT 研發的 ChartNet 資料集包含超過一百萬張合成圖表，旨在訓練模型同時處理視覺、數字與語言邏輯。實驗顯示，使用此資料集訓練的小型開源模型，在解讀趨勢與擷取數據上的表現超越了體積大上數倍的商業模型。

Google 推出 LiteRT 框架，加速行動裝置 NPU 推理效率 ref-19

Google 發表 LiteRT (原 LiteRT-LM) 生產級框架，協助行動開發者充分利用裝置上的 NPU (神經處理單元)。這能大幅降低實時影音處理與語音辨識的耗電量，推動 On-device AI 在 Epic Games 等大型應用的落地。

市場洞察

1. 「代理主權」從軟體滲透至硬體 OS 層級 ( ) ref-7, ref-103, ref-117, ref-4

延續前幾日關於 On-device AI 的討論，今日微軟 Project Solara 與 NVIDIA RTX Spark 的同步發表，確立了「代理式操作系統 (Agentic OS)」的時代正式到來。AI 代理不再只是應用程式內的一個功能，而是具備自主運行環境與專屬硬體算力（1 Petaflop）的系統核心。這意味著未來企業軟體的勝負關鍵將在於是否能被代理流暢地「呼叫與執行」。

2. 地緣政治與資本的高度集中化趨勢 ( ) ref-113, ref-67, ref-159, ref-63

DeepSeek 獲得 74 億美元的巨額注資，反映出各國對於「主權 AI (Sovereign AI)」的競賽已進入白熱化。相較於印度 IT 公司透過併購獲取客戶以應對定價壓力 (ref-0)，中美兩國的頂尖實驗室更傾向於透過海量資本確保算力優勢。這種「贏家全拿」的態勢將迫使二線模型廠轉向更專業化的垂直領域（如 ** Rudus** 的混凝土建築 AI, ref-68）。

3. FinOps 與 AI 治理成為企業必備門檻 ( ) ref-3, ref-11, ref-149, ref-36

Uber 與 GitHub 的動作顯示，AI 代幣 (Token) 成本已成為營運中不可忽視的負擔。當代理人生成代碼的速度是人類的數十倍時，企業必須引入像 Workday Agent Passport (ref-67) 或 ZeroDrift (ref-43) 這樣的治理工具，來監測 AI 代理的行為與開銷。這也對軟體廠商提出了新的挑戰：如何提供更低 Token 消耗的本地端處理解決方案。

Sources 178 items

ref-0 Hacker News Gmail thinks I'm stupid, so I left ref-1 Hacker News Adafruit receives demand letter from Fenwick legal counsel on behalf of Flux.ai ref-2 Hacker News MAI-Code-1-Flash ref-3 Hacker News 1-Click GitHub Token Stealing via a VSCode Bug ref-4 Hacker News Use your Nvidia GPU's VRAM as swap space on Linux ref-5 Hacker News Trump signs downsized AI order after weeks of reversals ref-6 Hacker News AI outperforms law professors in Stanford Law study ref-7 Hacker News Agentic Mfw ref-8 Hacker News Open Repair Data Standard ref-9 Hacker News How we index images for RAG ref-10 Hacker News Bringing Up DeepSeek-V4-Flash on AMD MI300X ref-11 Hacker News GitHub Copilot App ref-12 Hacker News Rsync and outrage ref-13 Hacker News Launch HN: Rudus (YC P26) – AI for concrete contractors ref-14 Hacker News U of T researchers demonstrate AI worm could target any online device ref-15 Hacker News Microsoft Doubles Down on Controversial Quantum Computing Claims ref-16 Hacker News DIY Bipedal Robot Used Pneumatic "Air-Muscles" Instead of Motors ref-17 OpenAI Blog Travelers deploys AI-powered claims countrywide with OpenAI ref-18 OpenAI Blog Codex for every role, tool, and workflow ref-19 Google Developers Blog Building real-world on-device AI with LiteRT and NPU ref-20 Hugging Face Blog Holo3.1: Fast & Local Computer Use Agents ref-21 NVIDIA Blog Industrial Software Leaders Build Secure, Autonomous AI Engineers With NVIDIA NemoClaw ref-22 NVIDIA Blog NVIDIA Partners With Microsoft on Unified Stack for Agentic AI Deployment, From Windows Devices to Cloud to Local ref-23 Cursor Blog What we’ve learned building cloud agents ref-24 AWS ML Blog The art and science of hyperparameter optimization on Amazon Nova Forge ref-25 AWS ML Blog Object detection with Amazon Nova 2 Lite ref-26 AWS ML Blog How Baz improved its AI Agent Code Review accuracy using Amazon Bedrock AgentCore ref-27 The Rundown AI Nvidia corners the AI agent stack ref-28 Interconnects Farewell Ai2 ref-29 Latent Space [AINews] Microsoft Build: MAI-Thinking-1 and MAI Family models ref-30 Latent Space GitHub's plan for Agents — Kyle Daigle, GitHub ref-31 Simon Willison Microsoft's new MAI models ref-32 Simon Willison datasette-agent-micropython 0.1a0 ref-33 Simon Willison micropython-wasm 0.1a1 ref-34 Simon Willison California Brown Pelican ref-35 TechCrunch AI Cyera eyes $12B valuation at 80x ARR multiple despite operating losses ref-36 TechCrunch AI Uber caps employee AI spending after blowing through budget in 4 months ref-37 TechCrunch AI New Microsoft tool lets devs spin up AI behavior tests using text descriptions ref-38 TechCrunch AI Martin Scorsese becomes the latest — and most unlikely — Hollywood voice for AI ref-39 TechCrunch AI Microsoft launches Scout, an OpenClaw-inspired personal assistant ref-40 TechCrunch AI Google rolls out fake call detection to protect against AI deepfake impersonation scams ref-41 TechCrunch AI Microsoft offers devs a better way to control AI agent behavior ref-42 TechCrunch AI Amazon faces class action lawsuit over Ring facial-recognition feature ref-43 TechCrunch AI Trump signs narrower executive order on AI oversight after industry objections ref-44 TechCrunch AI OpenAI launches new Codex tools for white-collar work ref-45 TechCrunch AI Anthropic scales Claude Mythos to critical infrastructure in 15+ countries ref-46 TechCrunch AI ZeroDrift raises $10M to protect AI models from themselves ref-47 TechCrunch AI Rocket engine startup Impulse raises $500 million to hire people, not AI ref-48 The Verge AI Microsoft Build 2026: The 7 biggest announcements ref-49 The Verge AI Trump signs executive order to review AI models before they’re released ref-50 The Verge AI Microsoft’s first advanced reasoning AI is here ref-51 The Verge AI Microsoft Scout is a new AI personal assistant built on OpenClaw ref-52 The Verge AI Google’s Phone app will tell you if a scammer is impersonating one of your contacts ref-53 The Verge AI Microsoft’s Project Solara is an OS for AI agent gadgets ref-54 The Verge AI Microsoft created the mini Surface dev box that Qualcomm couldn’t ref-55 The Verge AI Microsoft Build 2026: All the news about Windows, AI, RTX Spark, and more ref-56 The Verge AI Gemini Spark is the most impressive and terrifying AI experience I’ve had yet ref-57 MIT Technology Review AI Rehumanizing global health care with agentic AI ref-58 MIT Technology Review AI How small businesses can leverage AI ref-59 TechMeme Q&A with labor economist Kathryn Anne Edwards on why much of the AI jobs panic is overblown, why the US should start planning for AI job loss outcomes, and more (Casey Newton/Platformer) ref-60 TechMeme The UK CMA says publishers in the country will be allowed to opt out of Google AI search results, giving the company nine months to implement the changes (Imran Rahman-Jones/BBC) ref-61 TechMeme London-based Gigaton, which is developing AI to automate industrial control systems for cement, steel, glass, and chemicals plants, raised a $26M Series A (Mike Butcher/Pathfounders) ref-62 TechMeme French quantum computing startup Quobly, which is developing silicon-based quantum computers, raised a €115M Series A led by Bpifrance, SEALSQ, and STMicro (Tamara Djurickovic/Tech.eu) ref-63 TechMeme How India's ambitions to build and export a sovereign AI template are colliding with structural constraints, including a dependence on foreign AI infrastructure (Saritha Rai/Bloomberg) ref-64 TechMeme UnearthInsight: Indian IT companies have spent $7.1B on acquisitions since the start of 2025 to gain clients, as AI-led pricing pressure weakens organic growth (Shristi Achar/The Economic Times) ref-65 TechMeme Mastercard says it plans to offer on-chain settlement using several regulated USD stablecoins, initially supporting USDC, PYUSD, USDG, USDP, RLUSD, and SoFiUSD (Helene Braun/CoinDesk) ref-66 TechMeme Google says it is testing a new toggle inside its Search Console with some UK domain owners, letting them decide whether their sites appear in AI search results (Igor Bonifacic/Engadget) ref-67 TechMeme Sources: DeepSeek is set to raise ~$7.4B in its first funding round from investors including Tencent and CATL at a valuation of between ~$52B and ~$59B (Reuters) ref-68 TechMeme Current and former Meta employees detail Alexandr Wang's efforts to revive Meta's AI edge, as Muse Spark boosts confidence despite lagging rivals in coding (Hannah Murphy/Financial Times) ref-69 TechMeme University of Toronto researchers claim to have developed a "worm" powered by open source AI that exploits known flaws and tailors attacks for each computer (Cade Metz/New York Times) ref-70 TechMeme Sensor Tower: ChatGPT becomes the fastest app by far to hit 1B global MAUs; ChatGPT's MAUs are up 62% YoY in Q2 to date and Claude's MAUs are up 640% YoY to 56M (Harshita Mary Varghese/Reuters) ref-71 TechMeme Source: SpaceX aims to raise $75B in its IPO, selling 555.6M shares at $135 each, an unusual move since most companies set a price range before the roadshow (Echo Wang/Reuters) ref-72 TechMeme Source: Google asked Google Play app developers to join a "confidential content offer pilot", offering to pay for codebase access to use it to train AI tools (Jason Koebler/404 Media) ref-73 TechMeme Google selling $10B of equity to Berkshire is fascinating and functions as a signal that AI demand is high, in a world where the most cash buys the most compute (Ben Thompson/Stratechery) ref-74 InfoQ AI/ML Node.js Moves to One Major Release Per Year, Starting with Node 27 ref-75 InfoQ AI/ML Presentation: The Human Toll of Incidents & Ways To Mitigate It ref-76 InfoQ AI/ML OpenTelemetry Launches “Blueprints” Initiative to Simplify Enterprise Observability Adoption ref-77 InfoQ AI/ML Article: Why Vector Search Alone Isn't Enough: Hybrid Retrieval for RAG ref-78 The New Stack Rayfin: Microsoft’s answer to the gap between vibe coding and enterprise production ref-79 The New Stack Microsoft bets the enterprise AI race will be won on data context, not model power ref-80 The New Stack “A successful attack could be catastrophic”: Anthropic gives more groups access to Claude Mythos ref-81 The New Stack How GitHub plans to win developers back ref-82 The New Stack Microsoft really, really, really wants developers to love Windows again ref-83 The New Stack With Intelligent Terminal, Microsoft is reinventing the Windows terminal ref-84 The New Stack Microsoft debuts “Scout” at Build, a new personal agent for work ref-85 The New Stack OpenAI’s Codex adds new tools — Sites, Annotations, more plugins — for knowledge workers ref-86 The New Stack GitHub Copilot’s usage-based billing is live: Here’s what you need to know ref-87 Wired AI Meet Microsoft Scout, Your AI Coworker That Never Logs Off ref-88 Wired AI Flush With Cash From OpenAI, Opal Is Making an AI-Powered Audio Gadget ref-89 Wired AI The Trump Administration Is at War With Itself Over AI Regulation ref-90 TechCrunch Startups A startup, Everand, is now bundling e-books, audiobooks, and book clubs in challenge to Amazon ref-91 TechCrunch Startups Ex-Anduril engineer raises $42M to build the Amazon of composite parts ref-92 TechCrunch Startups Focused Energy raises whopping $240M Series A for laser-powered fusion tech ref-93 TechCrunch Startups Board, the new game startup from Mirror founder Brynn Putnam, raises $20M, has already sold thousands ref-94 TechCrunch Startups Why VivaTech 2026 is the place to see Europe’s AI strategy take shape ref-95 Crunchbase News Sector Snapshot: Defense Startup Funding Hits An All-Time Record As VCs Begin To Eye Exits ref-96 Crunchbase News SaaS Is Dead. Long Live SaaS! AI And The End Of The Rationing Of Knowledge Work ref-97 Stratechery The Google Capital Company ref-98 Tom's Hardware Cherry XTRFY launches first 8K ultra-wideband gaming keyboard — featuring more compact 70-percent layout ref-99 Tom's Hardware Intel addresses Arrow Lake blunder: 'We needed to build back our reputation' — says Arrow Lake Refresh's low price a key first step, laying the groundwork for Nova Lake ref-100 Tom's Hardware Microsoft announces Majorana 2 quantum computing chip — claims a practical machine will come in 2029 ref-101 Tom's Hardware Asus unveils its first Wi-Fi 8 router — ROG Rapture GT-BN98 Pro offers up to 2x real-world throughput uplift over Wi-Fi 7 ref-102 Tom's Hardware Intel’s next-gen LGA1954 socket will support Nova Lake, Razor Lake, and beyond — finally an Intel socket that outlives its CPUs ref-103 Tom's Hardware Microsoft debuts Surface RTX Spark Dev Box — Nvidia-powered mini-PC helps devs get ready for an agentic Windows ref-104 Tom's Hardware MSI unveils latest set of WiFi 7 gaming routers touting ultra-fast speeds — flagship RadiXBE19000 model comes with a built-in SSD slot for 'NAS Lite' experience and wireless speeds up to 19 Gbps ref-105 Tom's Hardware Computex 2026 Day One Wrap-Up: Arm makes a bold play for Windows PCs, PCIe 6.0 SSDs are coming, Asus embraces black and gold for ROG 20th ref-106 Tom's Hardware Cooler Master shows off new MWE Gold V4 Power supplies and GPU Shield adapter — per-pin monitoring can dynamically scale down power to stop cables melting ref-107 Tom's Hardware Mainland Chinese exhibitors reportedly locked out of Computex 2026, as Taiwan entry permits stall — parties complain applications left pending or hit with last-minute documentation requests ref-108 Tom's Hardware Intel reportedly no longer working on 6-core Nova Lake mobile SKU, alleges new rumor — Wildcat Lake Refresh to become focus for next-gen budget markets instead ref-109 Tom's Hardware AMD is considering a potential Ryzen 5 9600X3D — company says six-core Zen 5 X3D chip 'maybe something we look at doing... later this year' ref-110 Tom's Hardware Cooler Master shows off new HAF II 500 chassis, aluminum fans, and new air coolers — New V8 Cooler, Masterfan A, and updated Silencio 600 and HAF chassis’ add to an already comprehensive product stack ref-111 Tom's Hardware Phison shows PCIe 6.0 X3 SSD controller with 28 GB/s of bandwidth and 6.8 million IOPS, supports 2 petabytes per drive— also new power-sipping E37T SSDs for PCIe 5.0 systems consume a mere 4.5W ref-112 Tom's Hardware Intel Xeon 6+ Computex roundtable interview transcript — Kira Boyko and Tim Wilson on 18A wafer allocation, Clearwater Forest, and dropping hyper-threading ref-113 Tom's Hardware Bernie Sanders pushes for 50% public ownership of American AI companies — proposes AI sovereign wealth fund that would hold direct ownership stakes in largest AI firms ref-114 Tom's Hardware Grab this RTX 5070 gaming PC with a 7800X3D from CyberPowerPC for under $1,960 right now — exclusive discount available on 4K-ready pre-built rig with 32GB DDR5 and 1TB SSD that knocks $210 off ref-115 Tom's Hardware Gigabyte showcases new Infinity products for its 40th anniversary — X870 Infinity Next halo motherboard boasts metal 3D-printed elements, Aero Wood goes dark, MicroATX Stealth boards, Infinity-style GPUs extend down the product stack ref-116 Tom's Hardware SK hynix to double memory wafer capacity within five years, chairman says — AI-driven shortage will persist until at least 2030 ref-117 Tom's Hardware Jensen Huang says Nvidia wants to 'reinvent the single most important tool of humanity' with RTX Spark — Nvidia CEO touts support of 'literally every computer maker in the world' for its agentic AI PC platform ref-118 Tom's Hardware Tom's Hardware Unfiltered: Computex 2026, Day 1 — night markets, taking the MRT train, and a slew of demos ref-119 Tom's Hardware Drones to protect undersea cables from Russian sabotage touted in new defense pact — US one of three partners developing new tech to protect $1.8 trillion in daily transactions ref-120 Tom's Hardware Take home Gigabyte's Aero X16 gaming laptop for just $1,389.99 and save $610 — packed with 32GB of RAM and a powerful RTX 5070 GPU ref-121 Tom's Hardware AMD ‘had to re-engineer’ the Ryzen 7 5800X3D for a re-release — 10th Anniversary Edition chip had ‘a whole body of engineering work’ put into it ref-122 Tom's Hardware Intel Arc G3 interview transcript — Intel's Senior Product Director talks new handheld chips, Arrow Lake Refresh, and RTX Spark ref-123 Tom's Hardware Frore System’s solid-state AirJet Mini cools Intel’s Wildcat Lake laptop reference design – 15W of sustained, fanless cooling helps MacBook Neo competitor reach a svelte 11.3 mm, remain silent ref-124 Tom's Hardware Teen’s Bluetooth speaker named ‘BOMB’ caused a 10-hour delay on flight from Newark to Spain — passenger reported concerns to flight attendant at 32,000 feet, forcing plane back to the US ref-125 Tom's Hardware Intel says 'something has to give' with memory prices — company says it 'will continue to make sure that there are products which can take care of older memory technologies' ref-126 Tom's Hardware Chinese military has been acquiring Nvidia chips, even post-Washington export controls, research claims — multiple institutions linked to the PLA asked for Nvidia AI chips, according to publicly available documents ref-127 Politico Tech Trump’s AI order is a blow against laissez-faire ref-128 Politico Tech Florida is now OpenAI's biggest problem in red America ref-129 The Hacker News (Security) Weedhack Attacks Minecraft Users, CountLoader Hits 86K, Miners Spread via Pirated Content ref-130 The Hacker News (Security) Google June 2026 Android Update Patches 124 Flaws, One Actively Exploited ref-131 The Hacker News (Security) Gamaredon Exploits WinRAR to Deliver GammaWorm and GammaSteel Against Ukraine ref-132 The Hacker News (Security) Oracle WebLogic CVE-2024-21182 Added to KEV Catalog After Active Exploitation ref-133 The Hacker News (Security) AI-Driven Exploitation is Destroying Vulnerability Management. Here’s How to Handle It. ref-134 The Hacker News (Security) How Leading Organizations Are Turning EDR Into Operational Resilience ref-135 The Hacker News (Security) Pakistan-Linked SideCopy Targets Afghanistan Finance Ministry with Xeno RAT ref-136 EU Startups CEA-Leti and CNRS spin-off Quobly raises €115 million to industrialise silicon-based quantum computers ref-137 EU Startups Berlin’s INXM emerges from stealth with €5.7 million to build AI process execution engine for enterprises ref-138 EU Startups Barcelona’s Zazume raises €2.5 million to scale its AI-powered rental management platform ref-139 EU Startups Coach vs mentor – Who can help you level up your career? ref-140 EU Startups London’s DEScycle secures over €10 million in grant funding to scale critical metals recovery platform ref-141 EU Startups Used by Desigual, AWWG and Fútbol Emotion, Spanish AI fashion platform Modelia raises €1.03 million ref-142 MIT Research MIT researchers teach AI models to interpret charts ref-143 MIT Research Ultrasound-based pacemaker noninvasively steadies the heart ref-144 Lobsters Microsoft CEO: We’re moving from OS and apps to agents instead ref-145 Lobsters Introducing RadixAttention to Trellis ref-146 Lobsters Microsoft Intelligent Terminal 0.1 ref-147 Google News Agent Threat Rules: Open detection rule format for AI agent security threats ref-148 Google News Microsoft, Mayo Clinic plan to build frontier AI model for healthcare ref-149 Google News Workday Launches Agent Passport to Test, Verify, and Continuously Monitor Every AI Agent in the Enterprise ref-150 Google News Microsoft unveils Scout, an autonomous AI agent built on OpenClaw ref-151 Google News Build 2026: Microsoft Launches First Flagship Reasoning AI Model and More ref-152 Google News Tencent reportedly developing WeChat AI agent, makes it a top priority ref-153 Google News Microsoft Announces Project Solara, Its Take On An AI Agent Platform ref-154 Google News Anthropic expands access to powerful Mythos AI model to 150 more organizations ref-155 Google News Korea joins Anthropic's Project Glasswing for Mytos cybersecurity AI model ref-156 Google News Zhipu’s Chairman Amasses $22.4 Billion Fortune As Chinese AI Model Developer Soars Almost 1,000% ref-157 Google News Uber Limits AI Coding Tool Spend to $1,500 Per Employee a Month: Report ref-158 Google News Martin Scorsese Joins Generative AI Startup as Partner and Advisor ref-159 Google News Chinese AI startup DeepSeek pursues $7.4B funding at $52B-$59B valuation - report ref-160 Google News Sam Altman backs Alfred, a physical AI startup for robotics ref-161 Google News Alphabet Raises $80B in Stock Sale to Fund AI Infrastructure as Demand Outpaces Supply ref-162 Google News Nvidia Unveils RTX Spark AI PC Chip at Computex With Support From Major PC Makers ref-163 Google News Visa Invests in Replit to Build AI Agent Payment Infrastructure ref-164 Google News Florida Sues OpenAI and Sam Altman Over ChatGPT Safety Failures in First State-Led Action ref-165 Google News Chinese AI startup MiniMax unveils M3 as listing nears ref-166 Google News Anthropic files for IPO as AI startup behind ‘Claude’ prepares historic US stock market debut ref-167 Google News Global Dialogue on AI Governance, Geneva, 6–7 July ref-168 Google News Trump Signs Order on AI Security, Stops Short of Mandatory Tests ref-169 Google News TrueFoundry Launches Agent Gateway to Close the Enterprise AI Governance Gap ref-170 Google News Microsoft Launches Agent Control Spec for AI Governance ref-171 Google News Australia launches AI Safety Institute to boost trust in AI adoption ref-172 Google News Snowflake Accelerates the Agentic Enterprise with New AI, Governance, and Interoperability Innovations ref-173 Google News ASRock Rack Unveils Next-Generation AI Infrastructure Powered by NVIDIA Vera CPU at COMPUTEX 2026 ref-174 Google News Cisco Unveils Cloud Control Platform for Agentic AI Infrastructure Management ref-175 Google News Nvidia widens AI chip growth lead over Broadcom ref-176 Google News HPE shares soar 28% as demand for AI infrastructure powers stellar quarter ref-177 Google News US says ban on AI chip shipments applies to Chinese firms outside China