Reports
Report 2026-06-09 161 sources Input Tokens: 100,224 · Output Tokens: 4,424 · Total Tokens: 106,950

AI 戰情報告 (2026-06-09)

appleopenaianthropicmicrosoftgooglenvidiaxiaomiperplexity siri-aiafm-cloud-promythosgemma-4mimo-v2.5claude-4.8 agentiporushon-device-aimulti-agentsecuritycoding-assistant

AI 戰情報告 (2026-06-09)

昨日 AI 重點新聞

Apple WWDC 震撼發表 Siri AI 與 Apple Intelligence 次世代架構 ref-127, ref-2

Apple 宣布其語音助手正式進化為 Siri AI,具備強大的螢幕感知能力,能直接在 App 間執行複雜任務。此架構結合了 Private Cloud Compute 安全技術,並能調用授權自 Gemini 的雲端模型,標誌著 Apple 在 AI 代理領域的全面反擊。

OpenAI 緊隨 Anthropic 秘密遞交 IPO 申請,估值直指兆元 ref-13, ref-28

OpenAI 已正式向 SEC 提交 S-1 文件計畫上市。這發生在競爭對手 Anthropic 遞交申請僅一週後。雖然具體細節尚未公開,但外界預期其估值將挑戰 1 兆美元。Sam Altman 表示,此舉是為了籌措資金達成「個人 AGI」的遠大目標。

Anthropic 發表 Claude Code 動態工作流,實現「一人變一隊」 ref-85, ref-30

Anthropic 推出 Claude Code 重大升級,引入動態工作流。這項技術讓 Claude 能自主撰寫編排腳本,並在單一會話中分發多達 5 個並行代理 同時處理複雜專案(如自動建構 CLI 工具),將開發週期從季度縮減至數日。

Apple 坦承基礎模型 AFM Cloud Pro 係由 Gemini 蒸餾訓練而成 ref-1, ref-127

Apple 證實其 AFM Cloud Pro 模型是利用 Google Gemini 的輸出數據進行對齊與訓練,並在 Google Cloud 環境下的 Nvidia GPU 上運行。這揭示了即便在高度競爭的 AI 賽道,矽谷巨頭間的「算力與數據」依附關係依然緊密。

Microsoft Foundry 進入生產階段,主打企業級代理可靠性 ref-65, ref-34

微軟 在 Build 2026 大會宣布 Microsoft Foundry 全面升級,提供受管理的代理運行時 (Runtime)、治理標準與記憶體機制。這項服務旨在讓開發者能將實驗性代理無縫轉化為具備生產力、符合合規要求的企業應用。

小米 MiMo-V2.5 模型刷新速度紀錄,推理突破 1,000 TPS ref-0, ref-54

小米 發布其 1 兆參數模型 MiMo-V2.5-Pro-UltraSpeed,聲稱在標準 8-GPU 伺服器上實現了每秒 1,000 個代幣 的推理效率。小米計畫透過 API 開放此模型,挑戰 ChatGPT 與 Claude 的響應速度地位。

Google Gemma 4 12B 全球首款筆電端多模態代理模型 ref-1

Google 發表 Gemma 4 12B,這是一款專為本地筆電設計的無編碼器架構模型。它能完全在離線狀態下執行 Python 代碼、生成視覺洞察與自動化網頁開發,徹底降低了對高昂雲端 Token 的依賴。

微軟 GitHub 開源專案遭惡意蠕蟲入侵,竊取 AI 開發者憑證 ref-56, ref-141

逾 70 個微軟官方 GitHub 專案被植入惡意代碼,名為 Miasma 的蠕蟲利用 AI 編程代理讀取代碼時的漏洞,竊取開發者的系統憑證。此供應鏈攻擊事件引發了業界對 Agentic Coding 安全性的高度戒備。

Perplexity 宣示 2028 年 IPO 目標,並推動廣告獲利模式 ref-50

Perplexity CEO Aravind Srinivas 指出,該公司不急於跟風 OpenAI 的上市潮,將按原訂計畫於 2028 年 IPO。目前 Perplexity 將強化其「引用與搜尋」的核心優勢,並探索基於 AI 的精準廣告獲利路徑。

AI 編程新星 Cursor 年化營收突破 40 億美元,傳 SpaceX 有意收購 ref-101, ref-136

據傳 Cursor 的營收在數月內呈指數級成長。最新消息指出,Elon Musk 計畫讓 SpaceX 在 Cursor 上市後進行收購,此舉將整合 AI 自動化編程技術與 SpaceX 的複雜系統工程,強化實體 AI 的競爭優勢。

Apple Safari 引進「描述即擴充」功能,實現 Vibe Coding 生態 ref-6

Apple 展示了 Safari 的新功能,用戶只需輸入描述(如:追蹤網頁食譜並存入筆記),AI 就會自動生成對應的瀏覽器擴充功能。這代表 Vibe Coding 正式從工程師專屬進入到大眾消費者的日常應用中。

Microsoft Discovery 平台 GA,推動 AI 代理深入科學 R&D ref-75

微軟宣布 Microsoft Discovery 平台在 Azure 正式上市。該平台專為科學研究團隊設計,能協調多個自主代理進行材料模擬與分子研發,此前已成功協助研發出效能提升 1,000 倍的 Majorana 2 量子晶片。


市場洞察

1. IPO 浪潮確立 AI 產業進入「成果收割與洗牌期」 ( ) ref-85, ref-30, ref-144

延續昨日 Anthropic 獲 350 億美元融資與估值逼近 1 兆美元的趨勢,今日 OpenAI 秘密遞交 IPO 申請,宣告了 AI 獨角獸正集體轉向公開市場。這不僅是為了尋求流動性,更是因應模型開發成本極速膨脹後的必然選擇。隨著 Perplexity 設定 2028 年長線計畫,市場將出現「兆元權威」與「垂直利基」的分水嶺。

2. 「代理編排」從對話進化至並行作業系統 ( ) ref-85, ref-30, ref-151

昨日 OpenAI 宣告「聊天機器人已死」的論點,在今日獲得技術驗證。Anthropic 的動態工作流將單一代理拆解為「總承包商 + 執行團隊」模式,與微軟 Foundry 的代理生產平台遙相呼應。這代表 AI 應用已從單純的 Q&A 轉向複雜的專案交付,多代理協作 (Multi-agent Orchestration) 正式取代單體模型成為企業採用的核心標準。

3. Apple 的「主權 AI 務實主義」與生態系鎖定 ( ) ref-1, ref-6, ref-86, ref-33

Apple 在 WWDC 展現了極高的策略靈活性:在本地端使用私有雲 (PCC) 保護隱私,但在雲端則務實地採用 Gemini 數據進行訓練。這種「描述即功能 (Prompt-to-App)」的 Safari 擴充功能策略,顯示 Apple 意圖透過 AI 代理大幅縮減與 Google 瀏覽器生態系的擴充功能差距,讓使用者透過 Vibe Coding 建立個性化工作流。

4. 開放平台下的資安災難與「安全邊界」重構 ( ) ref-157, ref-131, ref-65

微軟 GitHub 蠕蟲事件與 LiteLLM 的漏洞曝光,再次為「代理化開發」敲響警鐘。當 AI 代理被賦予自動讀取、修改代碼的權限時,現有的靜態安全檢查已無法阻擋。市場將衍生出針對 AI 代理行為的「動態審計」與「安全網關」需求,這與微軟 Foundry 強調的「治理高於性能」趨勢一致。

Sources 161 items
ref-0 Hacker News MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second ref-1 Hacker News Apple reveals new AI architecture built around Google Gemini models ref-2 Hacker News Siri AI ref-3 Hacker News xAI is looking more like a datacentre REIT than a frontier lab ref-4 Hacker News Surveillance is not safety: A statement on the UK's latest threat to privacy [pdf] ref-5 Hacker News AI is slowing down ref-6 Hacker News Apple Core AI Framework ref-7 Hacker News Ask HN: What are tools you have made for yourself since the advent of AI? ref-8 Hacker News Show HN: Gitdot – A better GitHub. Open-source, written in Rust ref-9 Hacker News Replies to comments on my "LLMs are eroding my career" post ref-10 Hacker News Launch HN: Intuned (YC S22) – Build and run reliable browser automations as code ref-11 Hacker News Apple bets cheaper AI will woo small developers ref-12 Hacker News Show HN: Command Center, the AI coding env for people who care about quality ref-13 OpenAI Blog Confidential submission of draft S-1 to the SEC ref-14 Google Developers Blog Accelerating on-device AI: A look at Arm and Google AI Edge optimization ref-15 Google Developers Blog Announcing Genkit Middleware: Intercept, extend, and harden your agentic apps ref-16 AWS ML Blog Unlocking AI flexibility in Europe: A guide to cross-region inference for EU data processing and model access ref-17 AWS ML Blog It’s safe to close your laptop now: Hosting coding agents on Amazon Bedrock AgentCore ref-18 AWS ML Blog Better decisions at scale: How mathematical optimization delivers where intuition fails ref-19 AWS ML Blog End-to-end encrypted ML inference with Amazon SageMaker AI and FHE ref-20 AWS ML Blog Amazon Quick ARNs: Cross-account migration and namespace permissions ref-21 AWS ML Blog Evaluate your Amazon Nova Sonic voice agent at scale, no microphone required ref-22 Import AI Import AI 460: Reward hacking society, RSI data from Anthropic; and RL-based quadcopter racing ref-23 The Rundown AI Washington wants a piece of OpenAI ref-24 Latent Space [AINews] FrontierCode: Benchmarking for Code Quality over Slop ref-25 Simon Willison Siri AI at WWDC 2026 ref-26 TechCrunch AI Why Apple’s slow-and-steady AI bet is starting to look pretty smart ref-27 TechCrunch AI Mercor’s Brendan Foody calls out Sequoia, accusing it of ‘dual-pricing’ valuation tricks ref-28 TechCrunch AI As OpenAI files for IPO, Sam Altman’s eye-scanning company is doing layoffs, report says ref-29 TechCrunch AI Apple’s WWDC AI demos looked more real after $250M false ad settlement ref-30 TechCrunch AI OpenAI files confidentially for IPO, following Anthropic ref-31 TechCrunch AI Apple plays catch-up at WWDC ref-32 TechCrunch AI WWDC 2026: Everything announced on Siri AI, iOS 27, Apple Intelligence and more ref-33 TechCrunch AI Apple just taught your iPhone to finish your sentences, your photos, and your workflows ref-34 TechCrunch AI Apple will let you build workflows using AI in its new Shortcuts app ref-35 TechCrunch AI Apple’s Image Playground doesn’t suck anymore ref-36 TechCrunch AI Apple’s Photos app is getting new AI editing features ref-37 TechCrunch AI Apple gives Siri its own dedicated app ref-38 TechCrunch AI Apple is fixing the headache of splitting the bill with its new Siri in Camera feature ref-39 TechCrunch AI Apple’s long-awaited AI Siri overhaul is finally here ref-40 TechCrunch AI Amazon now lets you design custom merch using AI ref-41 TechCrunch AI WWDC 2026: What to expect, from Siri’s highly anticipated revamp to Apple Intelligence and iOS 27 ref-42 The Verge AI Apple is using AI to fix Safari’s extension problem ref-43 The Verge AI OpenAI files for IPO, following Anthropic ref-44 The Verge AI Amazon is launching AI-generated custom merch ref-45 The Verge AI Apple announces Siri AI and its next generation of Apple Intelligence ref-46 The Verge AI NotebookLM’s Gemini 3.5 upgrade adds a cloud computer and help finding sources ref-47 The Verge AI Microsoft’s AI chief says superintelligence is near, but won’t take your job ref-48 The Verge AI WWDC 2026: How to watch and what to expect ref-49 Ars Technica For the 2nd time in weeks, Microsoft packages laced with credential stealer ref-50 TechMeme In an interview, Perplexity CEO Aravind Srinivas says that the company is planning its IPO for 2028, regardless of how well the IPOs for Anthropic and OpenAI go (Arjun Kharpal/CNBC) ref-51 TechMeme The US FCC waives its deadline for Amazon to deploy half of its Leo satellites by July; Amazon is still required to launch all 3,232 satellites by July 30, 2029 (Michael Kan/PCMag) ref-52 TechMeme Apple overhauls the App Store with subscription options for groups, businesses, and schools, retention messaging to reduce churn, bundled IAP reviews, and more (Andrew Orr/AppleInsider) ref-53 TechMeme The first developer beta of iOS 27 offers the clearest signs yet of Apple's upcoming foldable iPhone, with references to folding hardware and flexible displays (Mark Gurman/Bloomberg) ref-54 TechMeme Xiaomi claims MiMo-V2.5-Pro-UltraSpeed tops 1K tokens/second, a first at the 1T-parameter scale, using a standard 8-GPU commodity node; API trial starts June 9 (Jose Antonio Lanz/Decrypt) ref-55 TechMeme Sam Altman and Jakub Pachocki claim OpenAI is "entering the third phase", aiming to automate AI research, boost the economy, and give everyone a personal AGI (OpenAI) ref-56 TechMeme Microsoft disabled 70+ of its repos on GitHub, including Azure-related tools like azure-functions-host, after hackers added credential-stealing malware to them (Zack Whittaker/TechCrunch) ref-57 TechMeme Google and Nvidia are helping Apple with Apple Foundation Model Cloud Pro, which Apple says is comparable to Gemini frontier models and runs on Nvidia GPUs (Kif Leswing/CNBC) ref-58 TechMeme Infinity Constellation, which builds AI tools and software businesses for the professional services market, raised a $24M Series A; it previously raised $17M (Lucinda Shen/Axios) ref-59 TechMeme Apple unveils an Apple Intelligence feature to automatically change compromised passwords, using agentic AI, and save them to the Passwords app (James Pero/Gizmodo) ref-60 TechMeme The Trump administration relaunches efforts to block state AI laws; Sen. Blackburn is leading negotiations and pushing KOSA as part of an AI preemption package (Axios) ref-61 TechMeme Meta puts $115M toward a "workforce academy", a free 5-week program to train Americans to build data centers, offering job guarantees at Meta construction sites (Wall Street Journal) ref-62 TechMeme Source: Cursor, which is prepping for an expected acquisition by SpaceX, passed $4B in annualized revenue in the last week, up from $3B in April and $2B in Feb. (Richard Nieva/Forbes) ref-63 TechMeme OpenAI confidentially files for an IPO, says it has "not decided on timing yet", as "there are things we want to do that are likely easier as a private company" (Ashley Capoot/CNBC) ref-64 TechMeme Google lowers the price of its Google AI Plus plan to $4.99 per month, down from $7.99, and doubles the included storage to 400GB (Abner Li/9to5Google) ref-65 InfoQ AI/ML Microsoft Foundry Adds Runtime, Tooling, and Governance for Production Agents ref-66 InfoQ AI/ML AWS Releases Next Generation of Amazon OpenSearch Serverless ref-67 InfoQ AI/ML Pinterest Uses Content Fingerprints for URL Deduplication Across Millions of Domains ref-68 InfoQ AI/ML Java News Roundup: JDK 27 in Rampdown, JDK 28 Expert Group, GlassFish, Infinispan, Kotlin ref-69 InfoQ AI/ML Gemma 4 12B Enables On-Device, Multimodal Agentic Workflows with an Encoder-free Architecture ref-70 InfoQ AI/ML Celebrating 20 Years of InfoQ ref-71 InfoQ AI/ML Article: Artificial Intelligence-Driven Phishing: How Phishing Technique Is Evolving and Implemented ref-72 InfoQ AI/ML Podcast: From MCP and Vibe Coding to Harness Engineering: How Did AI Native Engineering Evolve in One Year ref-73 InfoQ AI/ML Presentation: Mitigating Geopolitical Risks with Local-First Software and atproto ref-74 InfoQ AI/ML Presentation: Beyond Speed Limits: Exploring the Performance Power of Valkey ref-75 InfoQ AI/ML Microsoft Discovery Reaches GA on Azure, Powering the Agentic AI behind Majorana 2 Quantum Chip ref-76 InfoQ AI/ML Article: The Technology Adoption Curve, Twenty Years On ref-77 The New Stack How AI is solving the memory crunch it created ref-78 The New Stack Microsoft’s pitch to enterprises: Ditch Azure Repos for GitHub, despite its rocky reliability record ref-79 The New Stack Claude Code’s biggest upgrade yet ran 5 agents at once — here’s what happened ref-80 The New Stack Why Anthropic just doubled Claude Cowork limits at no charge ref-81 The New Stack For years, Apache Cassandra handed this work to your team — 6.0 takes it back ref-82 The New Stack “A dangerous combination”: The 2 factors that can “corrupt” AI agent workflows ref-83 The New Stack With Foundry, Microsoft bets the enterprise AI battle is about reliability, not capability ref-84 The New Stack Microsoft unlocks Visual Studio for developers left behind by its own AI ref-85 Wired AI OpenAI Confidentially Files for IPO on the Heels of SpaceX and Anthropic ref-86 Wired AI Apple’s New Siri AI Is Ready to Get Personal ref-87 Wired AI The UK Is Betting on a Billion-Dollar AI Supercomputer to Kick Its Addiction to US Tech ref-88 Wired AI Meta Deletes Face-Recognition System From Its Smart Glasses App After WIRED Report ref-89 Wired AI Momfluencers Are Pitching AI as a Better ‘Coparent’ Than Men ref-90 TechCrunch Startups Zepto’s IPO filing reveals fast growth, bigger losses, and a valuation question nobody’s answered yet ref-91 TechCrunch Startups How to apply to Startup Battlefield 2026, what you need ahead of today’s June 8 deadline ref-92 Crunchbase News How Bigger ACVs Are Bringing Direct Sales Back To Vertical AI ref-93 Stratechery Google Buys Compute From SpaceX, Broadcom’s Outlook, Apple’s AI Politics ref-94 Tom's Hardware Levelplay shows off magnetically attached fans that are reversible, connect via pogo pins and USB-C – plus an AIO that trades a screen for a big knob ref-95 Tom's Hardware Chinese startup claims photonic chip production without DUV lithography, says nanoimprint process cuts costs by 90% — 8-inch wafers produced without conventional optical lithography ref-96 Tom's Hardware Apple demonstrates cross-platform Siri upgrades in macOS 27 Golden Gate at WWDC — update brings Liquid Glass improvements and unifies AI strategy ref-97 Tom's Hardware Intel expands new game-boosting iBOT software with seven more games, up to a 27% improvement — Team Blue claims 12% average jump in newly-supported titles ref-98 Tom's Hardware Most new U.S. AI data centers are being built in drought zones — two-thirds of 809 planned projects set for areas with water shortages ref-99 Tom's Hardware Farmer donates land for a park, city sells it for data center development — $10 gift became $10M for city government, with $30M tax expected over next decade ref-100 Tom's Hardware Demand for data center CPUs has surged, and AI agents are responsible – why the CPU to GPU ratio is more important than ever for hyperscalers ref-101 Tom's Hardware Disgruntled ASML employees threaten to boycott Elon Musk conference appearance — staff express ire at political involvement and 'Nazi sympathies' ref-102 Tom's Hardware Travlfi JourneyGo 5G mobile hotspot review – Affordably priced, but lacking in performance and features ref-103 Tom's Hardware Four suspects identified in Finland undersea cable damage investigation — criminal case referred to prosecutors for consideration of charges ref-104 Tom's Hardware Bag a huge $308 saving on a two-year ExpressVPN Advanced sub and get four bonus months on top for free — 78% discount gets you this fast no-logs VPN service with support for 12 simultaneous connections for just $83 ref-105 Tom's Hardware Nvidia and SK hynix ink multi-year memory co-development and supply agreement — seeks to address extended development cycles ref-106 Tom's Hardware Intel introduced ‘the first processor in the x86 series and the first 8086 microprocessor’ on this day in 1978 — CPU was designed as a temporary substitute for the delayed iAPX 432 project ref-107 Tom's Hardware Executives are cutting jobs for an AI future that hasn't fully arrived yet, even as productivity gains remain difficult to prove — data neither confirms nor refutes an AI unemployment apocalypse ref-108 Tom's Hardware Amazon is offering a $324 Nvidia RTX 5060 GPU deal in a lightning sale, making it the cheapest model available — 1080p gaming on a budget ref-109 Politico Tech Trump’s AI meeting looks iffy — but the fight over tech profits is very much on ref-110 The Hacker News (Security) LiteLLM Flaw CVE-2026-42271 Exploited in the Wild, Chains to Unauthenticated RCE ref-111 The Hacker News (Security) One-Character Linux Kernel Flaw Enables Local Root Access, Exploits Now Public ref-112 The Hacker News (Security) Meta Blocks NSO Group's New WhatsApp Phishing Attack, Files Contempt Order ref-113 The Hacker News (Security) Critical Check Point VPN Flaw Exploited to Bypass Passwords in IKEv1 Setups ref-114 The Hacker News (Security) AI Phishing Is Crushing SOCs with Alert Volume: How to Reduce Tier 1 Overload ref-115 The Hacker News (Security) ⚡ Weekly Recap: Instagram Account Hacks, Android Zero-Day, GitHub Worm and More ref-116 The Hacker News (Security) The Hardest Fork ref-117 The Hacker News (Security) VerdantBamboo Deploys BSD Variant of BRICKSTORM on Linux Appliances ref-118 EU Startups Antwerp-based Pitchdrive closes €60 million Fund IV to back AI-native early-stage startups ref-119 EU Startups Milan-based Bending Spoons files for Nasdaq IPO – is Europe losing another tech contender? ref-120 EU Startups How Studocu uses NLP and AI to solve learning challenges at scale (Sponsored) ref-121 EU Startups UK-based NewOrbit raises €16 million to build commercial satellites for very low Earth orbit (VLEO) ref-122 EU Startups Ghent’s Companion.energy raises €7.8 million to help large enterprises manage energy in real time ref-123 MIT Research 3D-printed devices could streamline the production of drug-delivery microparticles ref-124 MIT Research Innovative projects explore ways to deal with extreme heat ref-125 MIT Research MIT astronomers discover the earliest known flickering quasar ref-126 Lobsters Self-hosting email the hard way from your own routable IPv4 block up ref-127 Lobsters Expanding Private Cloud Compute - Apple Security Research ref-128 Google News Meta's Muse Spark AI Model Replaces LLama 4 On Its Smart Glasses ref-129 Google News Apple to run top AI model on Nvidia chips and Google Cloud ref-130 Google News AWS Unfurls Open Source AI Agent to Enable Better AI Coding Outcomes ref-131 Google News Exclusive: Github and Hugging Face founders back AI agent startup Zaro’s $5.1m raise ref-132 Google News NotebookLM Is Getting Google's Latest Gemini AI Model ref-133 Google News MetaMask Launches AI Agent Wallet With Built-In Security Controls ref-134 Google News Anthropic to release public version of Mythos AI model tomorrow ref-135 Google News Cosine builds UK’s first sovereign frontier AI model ref-136 Google News AI coding startup Cursor, courted by SpaceX, picks London as European hub ref-137 Google News Apple Confirms AFM Cloud Is Its Own AI Model, Trained With Gemini Outputs ref-138 Google News OpenAI Plans Major ChatGPT Overhaul As AI Agent Race Intensifies ref-139 Google News At WWDC, Apple Tries To Make Siri An AI Agent ref-140 Google News AI model provides brain health insights from MRI scans ref-141 Google News New Relic Introduces AI Coding Observability to Bring Critical Visibility and Governance to AI Coding Assistants ref-142 Google News Google quietly installs 4GB AI model through Chrome; here’s how to remove it ref-143 Google News Tencent Lays Out Its Full AI Agent Strategy With WorkBuddy Enterprise and Agent Suite ref-144 Google News OpenAI Files Confidentially For IPO A Week After AI Rival Anthropic ref-145 Google News Martin Scorsese Backs AI Startup to Storyboard His Next Film ref-146 Google News Mistral AI buys Austrian physics AI startup in industrial push ref-147 Google News NVIDIA CEO Predicts Naver's Tenfold Growth via AI Partnership ref-148 Google News Nvidia’s $400 Million Kumo AI Acquisition Targets Enterprise Predictions ref-149 Google News Nvidias Huang, SKs Chey solidify AI partnership over chicken and somaek in Seoul ref-150 Google News Broadcom’s Perfection Premium: When a 143% AI Revenue Surge Triggers a 13% Wipeout ref-151 Google News Anthropic becomes most valuable AI startup, beats OpenAI with near $1T valuation ref-152 Google News Warren seeks admin leadership to testify on AI policy ref-153 Google News Germany to set up AI safety institute ref-154 Google News South Africa unveils draft AI policy, proposes new institutions and incentives ref-155 Google News Tech giants warn AI safety gaps could hand bioweapons to bad actors ref-156 Google News White House AI adviser Sriram Krishnan quits ref-157 Google News Stock Market Today, June 8: Intel Surges on Reported Google AI Chip Foundry Order ref-158 Google News Analysis: ASICs are coming for Nvidia's GPU dominance — and it could happen next year, says DIGITIMES analyst ref-159 Google News $35 Billion Bet on AI Infrastructure: Apollo and Blackstone Craft Custom Chip Financing Deal for Anthropic ref-160 Google News AI Chip Rout Wipes Out $1.3 Trillion in Semiconductor Value