Reports
Report 2026-06-05 171 sources Input Tokens: 105,235 · Output Tokens: 4,535 · Total Tokens: 113,293

AI 戰情報告 (2026-06-05)

anthropicnvidiaopenaicloudflareapplemicrosoftalibabametavercel nemotron-3-ultragrok-imagine-1.5gemma-4claude-mythosqwen-3.5 agentcontext-windowfinopsspatial-aion-device-ailegaltechmultimodal

AI 戰情報告 (2026-06-05)

昨日 AI 重點新聞

Anthropic 秘密遞交 IPO 申請,年化收入飆升至 470 億美元 ref-1, ref-150

Anthropic 正式向 SEC 遞交 IPO 申請。根據數據,其 ARR 已從 2025 年底的 90 億美元 爆發性增長至目前的 470 億美元。公司強調由於模型開發與運算成本極其高昂,急需透過公開市場籌資,目前估值已逼近 1 兆美元。

NVIDIA 發布 Nemotron 3 Ultra,具備 100 萬 Context 並優化代理工作流 ref-28, ref-11, ref-151

NVIDIA 推出 550B 參數的 MoE 模型 Nemotron 3 Ultra,支援 100 萬個代幣 的上下文視窗。該模型專為「長時運作代理 (Long-running agents)」設計,宣稱推理速度提升 5 倍且成本降低 30%,目前已在 AWS SageMaker 同步上線。

Cloudflare 警告:AI 代理流量將於 2027 年正式超越人類流量 ref-102, ref-99, ref-105

Cloudflare 執行長指出,目前自動化機器人與人類流量佔比已達 57.5% 對 42.5%。隨著 AI 代理應用爆發,預計到 2027 年,由 AI 代理產生的網絡請求將全面蓋過人類手動產生的流量,這對網絡安全與計費模式將產生深遠影響。

Apple 批准首個商業 AI 代理 Poke 進入 Messages for Business 平台 ref-39, ref-142, ref-65

Apple 已正式核准新創公司 Poke 作為其商務簡訊平台上的首個 AI 代理。這讓用戶能透過 iMessage 簡單發送指令,由 Poke 代理執行複雜任務,標誌著 Apple 在 AI 代理與商業流程結合邁出關鍵一步。

Cursor 與 OpenAI 針對「代幣成本」引進治理與計費優化功能 ref-14, ref-107, ref-13

針對 Sam Altman 提到的「代幣成本危機」,Cursor 宣布下調 Teams 方案價格並新增企業支出控制功能。同時,OpenAI 也更新了 ChatGPT 的記憶體架構(Dreaming 計畫),以提升運算效率並強化上下文新鮮度。

垂直領域 AI 融資熱:LegalTech 與零售支付 AI 獲數億美元投資 ref-131, ref-91, ref-58

歐洲 LegalTech 領域正迎來爆發,包括 AttiFin、Bayshore 等獲數百萬歐元投資。此外,丹佛新創 Scotch 獲 2,000 萬美元融資,致力於將 AI 嵌入高碎片化、高監管的酒類零售支付系統。

NVIDIA 執行長黃仁勳訪問首爾,強化機器人與物理 AI 供應鏈 ref-111

NVIDIA 創辦人黃仁勳訪問南韓,與合作夥伴商討 AI 供應鏈 佈局。他強調 機器人 (Robotics) 將是韓國下一個重大投資領域,NVIDIA 將推動 Vera Rubin 架構與南韓製造業整合。

Vercel 更新服務條款,釐清 AI 代理自主行為的責任歸屬 ref-27

Vercel 更新條款以應對 Agentic Workflows 的普及。新條款明確定義了「AI 功能」與「第三方工具」的責任邊界,規範當 AI 代理自主執行部署或調用 API 時,用戶需承擔最終審核與產生成本的責任。

Grok Imagine 1.5 預覽版推出,實現高品質圖像轉影片功能 ref-26

xAI 發布 grok-imagine-video-1.5-preview,能將單張靜態圖片轉換為流暢、具備電影感的影片。該模型強調物理規律與光影的還原,支援最高 720p 解析度,並提供 API 供開發者整合。

美國政府討論入股 AI 領先企業以確保國家利益 ref-66, ref-8

消息指出美國官員已與主要 AI 公司商討政府持股的可能性,以確保關鍵技術主權。此舉與川普最新簽署的去規管化行政命令相呼應,反映出美國在 AI 領先行位上的國家級防禦佈局。

TSMC 預警 AI 晶片供不應求,短缺將持續數年 ref-96, ref-168, ref-45

TSMC 執行長魏哲家在股東會表示,AI 晶片需求極其暢旺,供應鏈瓶頸在未來數年難以完全解決。雖然承諾價格穩定,但也預警先進封裝與記憶體短缺將持續推高算力成本。

MIT 研發 Cartesian 空間 AI,透過 RFID 解決零售庫存定位難題 ref-138

MIT 衍生新創 Cartesian 成功將無線訊號與機器學習結合,解決零售商實時追蹤海量庫存的難題。目前已在 Inditex (ZARA) 全球 700 多家門市部署,顯現 Spatial AI 在實體商業轉型中的潛力。


市場洞察

1. 「代理流量經濟」的到來與網絡基礎設施的重構 ( ) ref-102, ref-99, ref-105

Cloudflare 的警告揭示了一個根本性的轉變:網絡流量的主體正從「人類」轉向「AI 代理」。延續昨日關於 Tokenomics 基金會 成立的報導,這不僅是流量佔比的變化,更意味著企業網頁與 API 必須具備「代理可讀性」與「機器友善架構」。未來,無法被 AI 代理高效抓取或理解的商務平台,將在 2027 年後的網絡經濟中被邊緣化。

2. Context Window 與推理成本的「平衡戰」 ( ) ref-107, ref-14, ref-151

NVIDIA Nemotron 3 Ultra 支援 100 萬上下文,但隨之而來的是 Sam Altman 所擔心的代幣成本危機。市場正出現分化趨勢:一方是如 NVIDIA 推動的超大上下文、長時代理架構;另一方是如 Cursor 與 OpenAI 正在強化的 FinOps 控制與記憶體優化。企業正從「追求最強模型」轉向「追求最高代幣效益 (Token Efficiency)」,這將迫使軟體廠在前端就進行模型調度的優化。

3. AI 治理、合規與「代理主權」的衝突 ( ) ref-27

Vercel 的條款更新與美國政府研究入股 AI 實驗室,共同指向了一個核心議題:AI 責任歸屬。當代理具備自主決策與支付能力時,法律上的「行為人」界定變得模糊。這種「代理主權」的抬頭,將促使企業在採購 AI 服務時,將「權限沙箱」與「行為審計」視為比模型性能更重要的指標。

Sources 171 items
ref-0 Hacker News When AI Builds Itself: Our progress toward recursive self-improvement ref-1 Hacker News Anthropic's open-source framework for AI-powered vulnerability discovery ref-2 Hacker News Meta enables ADB on deprecated Portal devices [video] ref-3 Hacker News Open Code Review – An AI-powered code review CLI tool ref-4 Hacker News Do transformers need three projections? Systematic study of QKV variants ref-5 Hacker News KVarN: Native vLLM backend for KV-cache quantization by Huawei ref-6 Hacker News South Korean Forums Will Need to Scan Every Images with AI Censorship Tools ref-7 Hacker News 3D-printed book turns its own G-code into raised lettering ref-8 Hacker News Castor: CERN Advanced STORage Manager ref-9 Hacker News Fine-tuning an LLM to write docs like it's 1995 ref-10 Hacker News Magenta RealTime 2: Open and Local Live Music Models ref-11 Hacker News Latent Agents: A Post-Training Procedure for Internalized Multi-Agent Debate ref-12 Hacker News Satya Nadella 'Not Sure' Who Said Microsoft Wanted to Make Addictive AI ref-13 OpenAI Blog How Endava is redesigning software delivery around AI agents ref-14 OpenAI Blog Dreaming: Better memory for a more helpful ChatGPT ref-15 Google Developers Blog The latest updates to Google Pay ref-16 Google Developers Blog An important update: Transitioning Gemini CLI to Antigravity CLI ref-17 Google Developers Blog One Year of Innovation: Celebrating 100k Members in the Google Cloud x NVIDIA Developer Community ref-18 Google Developers Blog All the news from the Google I/O 2026 Developer keynote ref-19 Google Developers Blog Blazing fast on-device GenAI with LiteRT-LM ref-20 Google Developers Blog A Smarter Google AI Edge Gallery: MCP integration, notifications, and session continuity ref-21 Google Developers Blog Google Tensor SDK Beta with LiteRT ref-22 Hugging Face Blog Nemotron 3.5 Content Safety: Customizable Multimodal Safety for Global Enterprise AI ref-23 Hugging Face Blog EVA-Bench Data 2.0: 3 Domains, 121 Tools, 213 Scenarios ref-24 NVIDIA Blog Seoul Purpose: How NVIDIA and South Korea Are Building the Future of AI ref-25 NVIDIA Blog Forecast: Fun Ahead — 18 Games Join in June to Stream on GeForce NOW ref-26 xAI News Grok Imagine 1.5 Preview ref-27 Vercel Blog Updates to Legal Terms ref-28 AWS ML Blog NVIDIA Nemotron 3 Ultra now available on Amazon SageMaker JumpStart ref-29 The Rundown AI Ideogram and Reve rethink how AI images get made ref-30 Latent Space [AINews] not much happened today ref-31 Latent Space Reality: The Final Eval — Lukas Petersson and Axel Backlund of Andon Labs ref-32 Simon Willison AI enthusiasts are in a race against time, AI skeptics are in a race against entropy ref-33 Simon Willison Quoting Emanuel Maiberg, 404 Media ref-34 TechCrunch AI Mira Murati steps back into the spotlight, carefully ref-35 TechCrunch AI Ahead of its IPO, Anthropic’s Daniela Amodei shrugs off doubts about AI’s returns ref-36 TechCrunch AI Airbnb’s Brian Chesky plans to launch a new AI lab ref-37 TechCrunch AI Defense tech, AI, and fundraising take center stage at StrictlyVC Los Angeles on June 18 ref-38 TechCrunch AI Meta steals a tactic from Tesla and builds data centers in tents ref-39 TechCrunch AI Apple approves Poke as the first AI agent on its Messages for Business platform ref-40 TechCrunch AI Meta rolls out a new AI creator assistant on Facebook ref-41 TechCrunch AI What to expect from WWDC 2026: Siri’s highly anticipated revamp and Apple Intelligence updates ref-42 TechCrunch AI Is Silicon Valley ready to put robots in people’s homes? Hello Robot is. ref-43 TechCrunch AI Apple touts $1.4 trillion in App Store billings and sales, 90% without a commission ref-44 The Verge AI Kevin O’Leary agrees to downsize massive Utah data center ref-45 The Verge AI TSMC struggles to keep up with AI demand: ‘We can only support so much’ ref-46 The Verge AI Elon Musk is steamrolling Wall Street to become a trillionaire ref-47 The Verge AI Let us filter AI slop, you cowards ref-48 The Verge AI AI leaders call for tougher protections against AI-aided bioweapons ref-49 The Verge AI Amazon develops a warehouse robot that workers can speak to ref-50 Ars Technica Dashlane explains how attackers managed to download encrypted password vaults ref-51 MIT Technology Review AI How courts are coping with a flood of AI-generated lawsuits ref-52 TechMeme Starlink says that it has 12M+ "active customers" across 160+ countries, territories, and other markets, up from 9M in December 2025 (@starlink) ref-53 TechMeme Sources: data center developer Switch is in talks to raise billions of dollars from PE firms including Brookfield and KKR at a $50B+ valuation (The Information) ref-54 TechMeme Elon Musk petitioned the FTC in May to end its 2022 order restricting Twitter's data use, claiming Twitter no longer exists as X merged with xAI and then SpaceX (Ashley Belanger/Ars Technica) ref-55 TechMeme Analysis: Meta discreetly added code for an unreleased "NameTag" face-recognition system for its AI glasses over multiple updates to the Meta AI app this year (Wired) ref-56 TechMeme S&P Dow Jones rejected proposals to expedite S&P 500 eligibility for mega-cap IPOs such as SpaceX, which remains ineligible until one year after going public (Isabelle Lee/Bloomberg) ref-57 TechMeme An interview with Naomi Gleit, Meta's head of product who joined the company 20 years ago, on Zuckerberg's "unfair" reputation, AI agents' capabilities, more (Zoe Kleinman/BBC) ref-58 TechMeme Denver-based Scotch, which makes AI-powered payments tools for liquor retailers, raised a $20M Series A from VMG Partners, following a $10M seed in 2024 (Mary Ann Azevedo/Crunchbase News) ref-59 TechMeme Cloudflare CEO Matthew Prince says agentic traffic is "growing so fast that bots have now passed human traffic online for the first time" (Mark Tyson/Tom's Hardware) ref-60 TechMeme Anthropic calls for top AI labs to weigh slowing or temporarily pausing development, suggesting that self-improving AI systems may soon pose societal risks (Wall Street Journal) ref-61 TechMeme Sources: US officials have held talks with AI companies about the government acquiring shares; Sam Altman first pitched the concept directly to Trump in 2025 (NOTUS) ref-62 TechMeme OpenAI updates ChatGPT memory with a "more capable and compute-efficient" architecture and a summary page that lets users review and steer what it remembers (OpenAI) ref-63 TechMeme Quantinuum closed up 0.63% in its Nasdaq debut on Thursday, with a market value of $15.7B, after raising $1.68B in an upsized IPO (CNBC) ref-64 TechMeme Internal docs from lawsuits by 1,400 school districts show how social media companies targeted kids: Meta paid "teen ambassadors", Snap sent school-hour alerts (Jennifer Valentino-DeVries/New York Times) ref-65 TechMeme Poke, which lets users access AI agents via text message, becomes the first AI agent approved for Apple's Messages for Business platform (Sarah Perez/TechCrunch) ref-66 TechMeme Unsealed 2020 lawsuit: ex-IBM VP of threat intelligence alleges that IBM and AT&T concealed foreign cyber breaches to maintain eligibility for federal contracts (Bloomberg) ref-67 InfoQ AI/ML TypeORM Reaches 1.0 After Nearly a Decade, Signalling Renewed Maintenance ref-68 InfoQ AI/ML 30+ Updates per Second per Account: Uber Scales Ledger Processing with Batching ref-69 InfoQ AI/ML How a Culture of Data-Driven Conversations Can Support Platform Engineering ref-70 InfoQ AI/ML Presentation: Architecting a Centralized Platform for Data Deletion at Netflix ref-71 InfoQ AI/ML Article: Architectural Change Cases: A Practical Tool for Evolutionary Architectures ref-72 InfoQ AI/ML AWS Replaces Fat-Tree Data Center Networks with Random Graph Theory, Cutting Routers by 69% ref-73 The New Stack Cursor cuts prices and adds enterprise spend controls amid “tokenomics” reckoning ref-74 The New Stack Google Gemma 4 12B nearly matches 26B benchmarks — and runs on your laptop ref-75 The New Stack Snowflake thinks it knows what’s really slowing developers down ref-76 The New Stack Autonomous agents have met their biggest challenge yet: The database. ref-77 The New Stack Why agentic AI makes the ops platform the most important layer in the enterprise ref-78 The New Stack How to dramatically improve enterprise security alert tuning to battle cyberattacks ref-79 The New Stack Why the need for humans won’t disappear in the age of autonomous databases ref-80 The New Stack How to secure Kubernetes in the age of AI workloads ref-81 The New Stack Asana says its new AI “chief of staff” turns your Slack chaos into trackable work ref-82 The New Stack Nvidia’s best model is now live ref-83 The New Stack Mate Security’s Asaf Wiener made every backend engineer a model router. He’s right to. ref-84 Wired AI The AI IPO Race Heats Up, DOGE Whistleblower Sues Elon Musk, and Instagram Gets Hacked ref-85 Wired AI Jeff Bezos Is Funding a Wild Hunt for the Brain’s ‘Core Algorithm’ ref-86 Wired AI Alpha School’s Ritzy New York City Campus Costs $65,000 a Year—but Isn’t Actually a School ref-87 Wired AI Quantum Computing Is Having Its Public Market Moment ref-88 TechCrunch Startups Startup Battlefield is returning to Australia — here’s what happened the last time we came to Sydney ref-89 TechCrunch Startups Helion, the Sam Altman-backed fusion startup, raises $465M to build a power plant for Microsoft ref-90 Crunchbase News Active Startup Investors Didn’t Hold Back In May ref-91 Crunchbase News Exclusive: Scotch Raises $20M Series A To Disrupt Legacy Liquor Retail Tech With AI ref-92 Stratechery An Interview with Microsoft CEO Satya Nadella About Finding Core Competencies ref-93 Tom's Hardware Take your OpenClaw box back to the future with retro Mac Mini, Mac Studio docks — Wokyis tempts Nintendo and Apple lawyers, while adding a screen, ports, and style to your modern Mac ref-94 Tom's Hardware AMD's Frank Azor pushes back against claim that FSR 4.1 won't be ported to RDNA 3.5 GPUs — says 'no such decision' has been made ref-95 Tom's Hardware 8GB of RAM is back on laptops — companies are lowering memory offerings to make affordable notebooks during component crisis ref-96 Tom's Hardware TSMC CEO C.C. Wei says, ‘It will be a long time before we can meet customer demand’ — tells shareholders that he will keep prices stable, refrain from implementing price hikes ref-97 Tom's Hardware Qualcomm Roundtable Interview transcript — SVP of Compute and Gaming talks Snapdragon C, RTX Spark, and the agentic AI future ref-98 Tom's Hardware This RTX 5070 Ti gaming laptop is just $1,499 — Walmart fights AI price crisis with $500 discount ref-99 Tom's Hardware Nintendo Switch 2 with user-replaceable batteries coming to the EU — console maker confirms it will comply with regulations set to take effect from 2027 ref-100 Tom's Hardware Elon Musk's SpaceX secures 100% property tax exemption for planned $55 billion Terafab semiconductor factory in Texas — county approves 35-year deal worth hundreds of millions despite resident backlash ref-101 Tom's Hardware US tech layoffs record single-highest month in two years, and more than any other sector — nearly 40,000 get the axe, AI the most cited reason for layoffs ref-102 Tom's Hardware Silicon Motion increases sales of SSD controllers amid NAND shortage, but expects NAND shortages to get worse in 2027 — 'supply conditions will become even worse' ref-103 Tom's Hardware Hyte shows off Y50 chassis, aesthetic cable accessory kit, new fans, and updates Nexus Software — sub $100 Y50 brings value to Y-series, Nexus 3.0 goes web-based, now works on Mac, Linux, Windows, and your phone ref-104 Tom's Hardware Best of Computex 2026: Innovating despite disruptions ref-105 Tom's Hardware ‘Bots have now passed human traffic online,’ Cloudflare boss laments — says agentic traffic wasn’t expected to eclipse real people until next year ref-106 Tom's Hardware AMD's Helios MI455X AI platform breaks cover, initial systems use UALink-over-Ethernet interconnects — AMD's Vera Rubin rival surfaces, but the downsides of Ethernet could hamstring performance ref-107 Tom's Hardware OpenAI CEO Sam Altman admits AI token costs are becoming 'a huge issue' — company seeks improved value as overspending becomes a meme ref-108 Tom's Hardware Production of DDR4 memory and motherboards is restarting amid unprecedented memory shortages — PC industry preparing for a world without DDR5 ref-109 Tom's Hardware $559 Nvidia RTX 5070 GPU deal is the cheapest model available — 1440p high-performance gaming at just $10 above MSRP ref-110 Tom's Hardware HP has slashed an astonishing $2,600 off this RTX 5080 gaming PC, nearly 50% off — get an epic Omen 35L rig with a 9900X3D, 64GB DDR5, and 4TB of SSD storage for just $2,899.99 ref-111 Tom's Hardware Frore shows off LiquidJet Nexus coldplate for Nvidia Vera Rubin, other AI accelerators — offers up claimed 10% token generation boost over rival liquid-cooling solutions ref-112 Tom's Hardware Microsoft CEO says new AI data centers use as little water annually as a restaurant — closed-loop cooling system aims to slash consumption from millions of gallons as AI infrastructure faces mounting environmental scrutiny ref-113 Tom's Hardware Surface Laptop Ultra targets 110W TDP for RTX Spark Superchip — Microsoft reveals power budget of its high-end 15" system in hands-on session ref-114 Tom's Hardware Tom's Hardware Unfiltered: Computex 2026, Day 3 — the heat bites as our team races across Taipei ref-115 Tom's Hardware Aviation enthusiast uses Raspberry Pi and ADS-B radio to create viral real-time airport tracker — open-source 'Skylight' intercepts aircraft signals and projects flight paths onto your ceiling ref-116 Tom's Hardware TeamGroup shows off external SSD with wireless ‘self-destruct’ function — T-Create Expert P35SG External SSD can be wiped with a single text message ref-117 Politico Tech Hegseth doubles down on Anthropic's security risk designation ref-118 Politico Tech House unveils AI draft that would preempt state laws ref-119 The Hacker News (Security) FIFA World Cup 2026 Scams Are Already Live: Fake Sites, Banking Malware, and Stolen Logins ref-120 The Hacker News (Security) PCPJack Hijacks 230 AWS, Google Cloud, and Azure Servers for Covert SMTP Relay Network ref-121 The Hacker News (Security) Cisco Patches CVE-2026-20230 in Unified CM as Exploit Code Goes Public ref-122 The Hacker News (Security) Claude Code GitHub Action Flaw Let One Malicious Issue Hijack Repositories ref-123 The Hacker News (Security) Agentic AI Is Transforming Defense, But Only Secure IT Infrastructure Will Maximize It ref-124 The Hacker News (Security) ThreatsDay Bulletin: AI Agents Gone Wrong, Sketchy C2 Tools, ClickFix Tricks, JS Backdoors & 20+ New Stories ref-125 The Hacker News (Security) China-Linked TA4922 Expands Phishing Attacks to U.K., Germany, Italy, and South Africa ref-126 The Hacker News (Security) FlutterShell Backdoor Spreads to macOS via Malicious Google and YouTube Ads ref-127 The Hacker News (Security) Fake Sites Mimicking Open-Source Tools Rank High on Google to Deliver Malware via TDS ref-128 The Hacker News (Security) Hackers Spied on a Stock Exchange Executive's Outlook Mailbox for Five Months ref-129 EU Startups Your brand is no longer what you say it is, it is what ChatGPT says it is ref-130 EU Startups Lugano’s GR3N closes €15.5 million Series B to build the world’s first microwave-assisted PET recycling plant in Spain ref-131 EU Startups 10 European AI companies transforming the LegalTech sector ref-132 EU Startups London’s Fuse Energy secures €25 million Series B extension as it plans 32,000 sq ft London HQ ref-133 EU Startups London’s Airspeed raises €17.2 million Series A to build AI-powered execution layer for revenue teams ref-134 EU Startups The EU-Startups Podcast | Interview with Frédéric Mazzella, founder of BlaBlaCar ref-135 EU Startups Dutch startup New Dawn Bio raises €2.1 million to grow wood from tree stem cells in bioreactors ref-136 EU Startups German VC Merantix Capital closes €103 million fund for early-stage AI startups ref-137 MIT Research Startup helps retailers track their products in real-time ref-138 MIT Research Developing innovative alternatives to conventional carbon capture methods ref-139 MIT Research NSF renews support for MIT-led AI and physics institute, expanding a new model for discovery ref-140 Google News Supabase raises $500M as AI coding tools drive ‘phenomenal’ growth ref-141 Google News Meta Rolls Out AI Agent for Enterprises Globally ref-142 Google News Apple approves Poke as the first AI agent on its Messages for Business platform ref-143 Google News Microsoft (NASDAQ: MSFT) To Launch Homegrown AI Model Suite At Annual Build Developer Conference ref-144 Google News Google Cloud-IBM Team For Multi-Billion Dollar AI Agent Gemini Enterprise Push ref-145 Google News Alibaba opens Qwen to KFC, Luckin Coffee, and airlines in AI agent battle ref-146 Google News Developers Are Still Waiting for Access to Meta's Latest AI Model ref-147 Google News Anthropic built what may be its most powerful AI model ever, but the company decided not to publicly release it. Why? Alistair Barr reports. ref-148 Google News Microsoft Scout: An always-on Personal AI Agent for Microsoft 365 ref-149 Google News Hugging Face Transformers RCE flaw enables stealthy compromise via AI model configs ref-150 Google News Anthropic files for IPO, citing high AI model development costs ref-151 Google News NVIDIA Could Seal a Major Alliance With Apple After Launching Nemotron 3 Ultra AI model ref-152 Google News AI agent internet traffic could surpass human traffic by 2027, Cloudflare CEO warns ref-153 Google News Tencent’s WeChat AI Agent Edges Closer to Launch ref-154 Google News Meta Platforms Weighs $200 Monthly Subscription for AI Agent Product- The Information ref-155 Google News AI startup Flourish reportedly raises $500M round backed by Jeff Bezos ref-156 Google News Microsoft's AI Revenue Run Rate Just Crossed $37 Billion. Is It the Best AI Stock to Buy Now? ref-157 Google News Salesforce (NYSE: CRM) Agentforce Platform Crosses $1B ARR As AI Revenue Surge Reshapes Enterprise Software Race ref-158 Google News Palantir Expands Google Cloud AI Partnership ref-159 Google News Nvidia (NVDA) Expands AI Capabilities with Kumo AI Acquisition ref-160 Google News U.S. and Japan Launch $1B AI Partnership to Propel Scientific Ad ref-161 Google News Goldman Sachs expects SpaceX’s AI revenue to surge 100 times by 2030 ref-162 Google News Alphabet Expands AI Funding Plan to $84.75 Billion ref-163 Google News Martin Scorsese Joins AI Image Startup Black Forest Labs as Partner and Adviser ref-164 Google News Bipartisan ‘Great American AI Act’ draft proposes new federal AI governance framework ref-165 Google News Trump’s Executive Order frames U.S. AI policy around deregulation, cybersecurity, and voluntary federal coordination ref-166 Google News OpenAI diverges from White House on AI safety rules ref-167 Google News Warren invites Nvidia CEO Jensen Huang to Senate hearing on China AI chip sales ref-168 Google News TSMC CEO says AI chip demand will outpace supply for years ref-169 Google News Foxconn, Intel Team Up to Develop AI Infrastructure ref-170 Google News Broadcom bets big on AI infrastructure as networking and XPU demand hits ‘insatiable’ levels