# AI news for Monday, Nov 16, 2026

- HTML page: https://aidr.today/date/2026-11-16?lang=en
- Other language: https://aidr.today/date/2026-11-16.md?lang=vi
- Previous day: https://aidr.today/date/2026-11-15.md?lang=en
- Next day: https://aidr.today/date/2026-11-17.md?lang=en

## AI;DR

- The US Commerce Department issued an interim rule limiting remote access to advanced AI chips and servers from countries including Singapore and Malaysia, while preparing a ban on Chinese access to Nvidia Blackwell and newer GPUs through data centers in seven countries. ([story](https://aidr.today/259eb27b?lang=en))
- Jensen Huang warned neocloud executives that Nvidia’s “open checkbook” era was over, requiring upfront customer payments for GPU allocations. The Wall Street Journal said he was pulling back from guarantees as chip commitments mounted. ([story](https://aidr.today/0f548d15?lang=en))
- Nvidia’s Vera Rubin NVL72 delivered the fastest training time on six of seven MLPerf Training v6.1 benchmarks and up to 2x the performance of a GB300 baseline. Nvidia platforms led every benchmark in the round. ([story](https://aidr.today/c639d63a?lang=en))
- Alibaba launched Qwen4-Max with MoE, native vision and a 1M-token context window; internal tests found task costs over 60% below Qwen3.8-Max. It scored 43 on Artificial Analysis, 9 points higher than Qwen3.8 Max and 3 behind Kimi K3.1 Pro, and is available on QwenCloud and OpenRouter for $1.5 input and $4.5 output per million tokens. ([story](https://aidr.today/38f8444b?lang=en))
- The Trump administration plans to name Nvidia, Amazon, Microsoft, OpenAI and Anthropic among more than 24 AI “national champions.” The program aims to help US AI companies win business overseas, according to The Wall Street Journal. ([story](https://aidr.today/7c1c1ee0?lang=en))
- Donald Trump took Treasury Secretary Scott Bessent and Commerce Secretary Howard Lutnick to the ASEAN summit in Asia, where chip export controls will be central to talks. Bessent had criticized Commerce over its handling of Nvidia customer Megaspeed. ([story](https://aidr.today/e0c07e54?lang=en))
- Indonesian police raided Megaspeed’s office at the FBI’s request and found empty Nvidia server racks. Records show imports of restricted H100 processors from 2024 to 2026, but authorities have made no public finding that Megaspeed violated export rules. ([story](https://aidr.today/fa188158?lang=en))
- Federal Reserve Governor Michael Barr warned that autonomous agents managing personal finances could accelerate bank runs from hours to minutes. Major banks have already intercepted agents attempting account actions without verified consent. ([story](https://aidr.today/05151249?lang=en))
- Google’s Threat Intelligence Group reported the first observed case of attackers using an autonomous AI framework with no human in the loop to turn security patches into exploits. The development raises the speed and scale of cyberattacks. ([story](https://aidr.today/445cf3d6?lang=en))
- The US and Pax Silica allies, including Japan, the UK and India, agreed to create a joint list of AI technologies and supply chains to protect from China. The pact comes before Trump’s meeting with Xi Jinping at APEC in Shenzhen. ([story](https://aidr.today/77e07a79?lang=en))
- A Tunix recipe using JAX and Pallas eliminated floating-point train–inference logprob drift for on-policy LLM reinforcement learning across 64 Cloud TPU v5p chips. The work targets more stable training on TPU. ([story](https://aidr.today/27ef7f67?lang=en))
- ChatGPT Wallet is listed as a new product, but the item provides no further details about its features, availability or launch timing. ([story](https://aidr.today/90a5432c?lang=en))
- Tencent’s Hy4 flash is now available on OpenRouter as a lower-cost, lower-latency variant of Hy4. It has roughly one-quarter of the flagship’s parameters, a 1,048,576-token context window, and costs $0.04 input and $0.18 output per million tokens. ([story](https://aidr.today/146ec48a?lang=en))

## Ranked stories

1. [The US Commerce Department issues an interim rule to limit remote access to advanced AI chips and servers from select countries, such as Singapore and Malaysia (Magan Crane/Bloomberg)](https://aidr.today/259eb27b?lang=en) — Regulation · techmeme.com
   The US Commerce Department issued an interim rule to limit remote access to advanced AI chips and servers from select countries, including Singapore and Malaysia.
2. [Fed Governor Warns Autonomous AI Could Accelerate Pace of Bank Runs](https://aidr.today/05151249?lang=en) — Regulation · marketbrief.now
   Federal Reserve Governor Michael Barr warned that autonomous AI agents could accelerate the speed of deposit withdrawals during bank runs, potentially causing events to unfold in minutes. He noted that major banks have already intercepted autonomous agents attempting to execute actions in customer accounts without verified consent.
3. [OpenAI says it has signed agreements with independent assessors Apollo Research, METR, Transluce, and Goodfire for its first phase of embedded assessments (OpenAI)](https://aidr.today/8792f920?lang=en) — Industry · techmeme.com
   OpenAI has signed agreements with independent assessors Apollo Research, METR, Transluce, and Goodfire for its first phase of embedded assessments.
4. [NVIDIA Vera Rubin NVL72 Achieves up to 2x Blackwell Ultra Performance in MLPerf Training v6.1](https://aidr.today/c639d63a?lang=en) — Chips · developer.nvidia.com
   In MLPerf Training v6.1, the NVIDIA Vera Rubin NVL72 platform achieved up to 2x more performance than a highly optimized NVIDIA GB300 NVL72 baseline. The Vera Rubin architecture delivered the fastest time to train on six of the seven benchmarks and maintained above 97% overall efficiency when scaling from 256 to 1,024 GPUs.
5. [Huang Warned Neoclouds Nvidia's 'Open Checkbook' Was Over Before Ouster](https://aidr.today/0f548d15?lang=en) — Chips · marketbrief.now
   Former Nvidia CEO Jensen Huang told neocloud executives in Berlin in October that Nvidia's days as an "open checkbook" were over and that upfront customer payments were needed for GPU allocations. Huang criticized cloud providers' pricing, noting Elon Musk rented Nvidia chips at better rates. Before his ouster, Huang questioned a dealmaking spree involving financing offers led by Vice President N…
6. [SpaceX's $60B Cursor Acquisition Underscores Shift From AI Video Toward Coding Tools](https://aidr.today/a20fdde9?lang=en) — Tools · marketbrief.now
   SpaceX is reorienting its artificial intelligence division to prioritize software automation and autonomous agents over generative media. The company acquired coding startup Cursor for $60B in August to accelerate work on its Grok Bot assistant and other tools for programmers. SpaceX also laid off approximately 60 employees from the Grok Imagine team.
7. [Anthropic's S-1 Filing Discloses $16.3B in H1 2026 Revenue and $974M Q2 Operating Profit](https://aidr.today/67615dab?lang=en) — Industry · marketbrief.now
   Anthropic's S-1 filing shows the company generated $16.3B in revenue during the first half of 2026 and recorded a $974M operating profit in the second quarter of 2026. Revenue per dollar spent on compute rose from $0.40 in early 2025 to $1.72 by the third quarter of 2026, while year-to-date gross margins expanded to over 85%.
8. [Trump Takes Treasury and Commerce Chiefs to Asia to Discuss Chip Export Controls](https://aidr.today/e0c07e54?lang=en) — Regulation · marketbrief.now
   President Trump departed for the ASEAN summit in Asia on Sunday with Treasury Secretary Scott Bessent and Commerce Secretary Howard Lutnick. The administration plans to make chip exports a central part of the agenda during meetings this week following Bessent's criticism of the Commerce Department's handling of Nvidia customer Megaspeed.
9. [Google's TIG reports the first observed instance of attackers using an autonomous AI framework with no human in the loop to turn security patches into exploits (Duncan Riley/SiliconANGLE)](https://aidr.today/445cf3d6?lang=en) — Agents · techmeme.com
   Google's Threat Intelligence Group reports the first observed instance of attackers using an autonomous AI framework with no human in the loop to turn security patches into exploits.
10. [Mimicrobotics' New Video-Action Model Ranks First on RoboLab Robotics Benchmark](https://aidr.today/97e6f5dd?lang=en) — Models · huggingnews.com
   Mimicrobotics has launched FLUX-mimic-x, a 7B parameter video-action model for real-time robot control. The model achieved the highest score among "system 1" models on the RoboLab benchmark and was developed using the original DROID dataset without simulation or RoboLab embodiment data.
11. [Codex Coding Harness Ties for First Place on Ripple-65B Preview Benchmarks](https://aidr.today/3f9528bc?lang=en) — Models · huggingnews.com
12. [AI Rout Drags S&P 500 Below 7,500 as Semiconductor Selloff Spreads](https://aidr.today/bfcd015d?lang=en) — Industry · marketbrief.now
13. [Hark Releases Handoff², a 2x Faster Computer-Use Agent for Hark Pro](https://aidr.today/83fd6463?lang=en) — Agents · marktechpost.com
14. [Nvidia Launches Vera CPU Rack-Scale Platform for High Performance Computing](https://aidr.today/b7a80424?lang=en) — Infra · huggingnews.com
15. [Liquid AI's New d1.2-Ultra Model Tops Artificial Analysis Decision Index](https://aidr.today/d0b6457c?lang=en) — Tools · marketbrief.now
16. [Unlocking Stable On-Policy RL on TPU with Bitwise-Zero Train–Inference Mismatch](https://aidr.today/27ef7f67?lang=en) — Research · developers.googleblog.com
17. [Perplexity Dedicates 10,000 GPUs to Build AI for Autonomous Research](https://aidr.today/8a537507?lang=en) — Industry · marketbrief.now
18. [Indonesian Police Raid Megaspeed Office at FBI Request, Find Empty Nvidia Server Racks](https://aidr.today/fa188158?lang=en) — Regulation · marketbrief.now
19. [Alibaba Launches Qwen4-Max AI Model With 1M Token Window and MoE Architecture](https://aidr.today/38f8444b?lang=en) — Models · marketbrief.now
20. [ChatGPT Wallet](https://aidr.today/90a5432c?lang=en) — Products · openai.com
21. [Anthropic's Claude Adoption Lags Behind Copilot's in Microsoft Office Suite](https://aidr.today/2b7736b9?lang=en) — Industry · marketbrief.now
22. [The US-led Pax Silica faces local backlash in the Philippines over a planned AI hub, with concerns about displacement, resources, and threats to sovereignty (New York Times)](https://aidr.today/cbf86ac4?lang=en) — Regulation · techmeme.com
23. [Jensen Huang Is Gone. Now What?](https://aidr.today/e59f883e?lang=en) — Industry · wired.com
24. [The US and Pax Silica allies, including Japan, the UK, and India, agree to create a joint list of AI technologies and supply chains to protect from China (Financial Times)](https://aidr.today/77e07a79?lang=en) — Regulation · techmeme.com
25. [Noeri, which is developing AI software to review and suggest improvements for its clients' AI models and tools, raised $110M led by Menlo Ventures (Dina Bass/Bloomberg)](https://aidr.today/a790b671?lang=en) — Funding · techmeme.com
26. [Tel Aviv- and SF-based Corma, which builds AI agents custom-trained for cyber defense, raised a $125M Series A led by NEA, after raising a $60M seed in August (Angus Loten/Wall Street Journal)](https://aidr.today/35689046?lang=en) — Funding · techmeme.com
27. [OpenAI and METR Reach Deal for External Assessment of Largest RL Run](https://aidr.today/e247a1c7?lang=en) — Industry · marketbrief.now
28. [Israeli Military Used Amazon Facial Recognition to Track 3.5 Million Palestinians](https://aidr.today/5002dcee?lang=en) — Industry · huggingnews.com
29. [Anthropic's Monthly Revenue Growth Hits 23% Ahead of IPO](https://aidr.today/ddc3388b?lang=en) — Industry · marketbrief.now
30. [Google Expands Antigravity Researcher Program to Include Gemini 4 Carbon](https://aidr.today/79381f16?lang=en) — Agents · huggingnews.com
31. [OpenAI Pursued $5B Acquisition of AI Safety Firm Goodfire](https://aidr.today/89867056?lang=en) — Industry · marketbrief.now
32. [Congress and Bessent Pressure Howard Lutnick Over Alleged Role in Nvidia-Megaspeed Scandal](https://aidr.today/01e76437?lang=en) — Regulation · marketbrief.now
33. [Sources: Jensen Huang has discussed a comeback as Nvidia CEO with confidants, as allies form a "war room" to lobby for his return and seek White House support (Financial Times)](https://aidr.today/2fbba047?lang=en) — Industry · techmeme.com
34. [Sources: the Trump admin plans to name Nvidia, Amazon, Microsoft, OpenAI, and Anthropic among 24+ AI "national champions" to help them win overseas business (Wall Street Journal)](https://aidr.today/7c1c1ee0?lang=en) — Regulation · techmeme.com
35. [SemiAnalysis Cuts 2030 Global AI Accelerator Capex Projection to $923B](https://aidr.today/7bd20c77?lang=en) — Chips · marketbrief.now
36. [Tencent: Hy4 flash is now on OpenRouter](https://aidr.today/146ec48a?lang=en) — Tools · openrouter.ai
37. [World Labs and Fei-Fei Li Join AMD](https://aidr.today/a4721ab4?lang=en) — Industry · marketbrief.now
38. [Qwen4-Max scores 43 on the Artificial Analysis Intelligence Index, beating Qwen3.8 Max at ~67% lower cost per task but trailing Kimi K3.1 Pro at ~11% lower cost (@artificialanlys)](https://aidr.today/b5ad7e36?lang=en) — Models · techmeme.com
39. [Nvidia Shares Slide Over 5% Premarket as Global Chip Selloff Sweeps Into U.S. Markets](https://aidr.today/a92a2143?lang=en) — Industry · marketbrief.now
40. [Wall Street's 10 biggest IPOs of 2026, including SpaceX, SK, and Cerebras, have fallen by an average of 20.5% from their offer prices, as the AI frenzy fades (Emily Herbert/Financial Times)](https://aidr.today/fa3230d7?lang=en) — Funding · techmeme.com
41. [A look at China's push for AI literacy by 2030, as schools adopt AI-focused curricula, classroom cameras, and AI-generated lessons to mitigate teacher shortages (Reuters)](https://aidr.today/a791cbfc?lang=en) — Regulation · techmeme.com
42. [Fed Warns Current AI Investment Cycle Could Outpace 1990s Telecom Boom](https://aidr.today/ef4e4ca8?lang=en) — Funding · marketbrief.now
43. [fal Launches ByteDance Seedance 2.5 4K Video Model](https://aidr.today/1d9d8e96?lang=en) — Models · marketbrief.now
44. [Solmate Launches GPU-as-a-Service for AI Training and Deployment](https://aidr.today/01b3195e?lang=en) — Infra · marketbrief.now
45. [US Prepares to Block Chinese AI Firms From Renting Nvidia GPUs via Overseas Data Centers](https://aidr.today/44bb93d9?lang=en) — Regulation · marketbrief.now
46. [Lambda Boosts AI Factory Resiliency and Efficiency With NVIDIA Photonics](https://aidr.today/fc3a56c1?lang=en) — Infra · blogs.nvidia.com
47. [Science at the Frontier: How scientists use ChatGPT](https://aidr.today/d91fd06d?lang=en) — Research · openai.com
48. [Midas Project Finds Right-Wing Creators Recruited to Promote Nvidia Chip Sales to China](https://aidr.today/529446aa?lang=en) — Industry · huggingnews.com
49. [Fed Warns US Banks Have Close to $1T Exposure to AI Infrastructure](https://aidr.today/b03b7697?lang=en) — Regulation · marketbrief.now
50. [Uber Adds Amazon's Zoox Robotaxis to its Las Vegas Ride-Hailing Platform](https://aidr.today/c5435e1d?lang=en) — Industry · marketbrief.now
51. [SK Hynix Pledges More Than $1B to Expand Semiconductor Supply Chain](https://aidr.today/07877568?lang=en) — Chips · marketbrief.now
52. [Anthropic Paid $290,000 for AI Policy Lobbying via Former Rubio Aide's Firm](https://aidr.today/572f2353?lang=en) — Regulation · marketbrief.now
53. [NVIDIA Expands Slurm to Bring Quantum Computing Into Supercomputing](https://aidr.today/c046784c?lang=en) — Infra · blogs.nvidia.com
54. [NVIDIA CUDA-X Libraries Give Agents the Tools to Accelerate Scientific Discovery](https://aidr.today/970d3772?lang=en) — Frameworks · blogs.nvidia.com
55. [Claude now builds interactive websites inside canvas graphics, reviving the Flash era with real HTML](https://aidr.today/0d0e2b3c?lang=en) — Products · the-decoder.com
56. [Baidu's Surge in AI Cloud Revenue Drives Q3 Return to Profit](https://aidr.today/52459776?lang=en) — Industry · marketbrief.now
57. [Dreamteam, which uses AI agents for CRM tasks like data capture and research, raised $53M across seed and Series A led by Together Fund and Accel, respectively (Jagmeet Singh/TechCrunch)](https://aidr.today/b8f52bc5?lang=en) — Funding · techmeme.com
58. [DeepSeek Agent Swarm Builds PyBot Chatbot to Rival 360M-Parameter LLMs](https://aidr.today/341c5f2c?lang=en) — Agents · huggingnews.com
59. [Parents pay $75k for Alpha's AI school. Half of its students missed math target](https://aidr.today/9a643798?lang=en) — Products · fortune.com
60. [Knit Health, which uses its AI world model to predict patient care needs, forecast treatment impacts, and optimize referrals, raised a $35M Series A led by NEA (Will Maddox/Axios)](https://aidr.today/bea01e25?lang=en) — Funding · techmeme.com
61. [Meet Meta’s JiTTests: LLM-Generated Tests Designed to Fail Before Risky Code Lands](https://aidr.today/19f889a6?lang=en) — Tools · marktechpost.com
62. [Emergent Biological Capabilities in a Foundation Model for Molecular Interactions](https://aidr.today/b43ddf96?lang=en) — Research · huggingface.co
63. [Post-Training NVIDIA Nemotron to Increase Accuracy in Your Scientific Domain](https://aidr.today/4491adcf?lang=en) — Models · developer.nvidia.com
64. [European Defense Ministries Seek Local AI Tools to Replace Palantir](https://aidr.today/9a233772?lang=en) — Regulation · huggingnews.com
65. [Qwen: Qwen4 Max is now on OpenRouter](https://aidr.today/d2d8cd93?lang=en) — Tools · openrouter.ai
66. [Gamer: A Generalist Vision-Language-Action Foundation Model for Long-Horizon Real-Time Game Agent](https://aidr.today/96bff3e1?lang=en) — Frameworks · huggingface.co
67. [Envisioning Agentic Robotics: A New Era of Robotic Systems](https://aidr.today/110227c0?lang=en) — Research · huggingface.co
68. [South Korean Prime Minister Pledges Nationwide Crackdown on AI Cheating in University Exams](https://aidr.today/ace68bb7?lang=en) — Regulation · marketbrief.now
69. [Nvidia's Rubin AI Server Shipments Pushed to First Half of 2027](https://aidr.today/155f139b?lang=en) — Chips · marketbrief.now
70. [Google Maps can now create trip itineraries for you](https://aidr.today/2fdf48c7?lang=en) — Products · theverge.com
71. [Nvidia Halts Billions in Credit Backstops for AI Startups to Curb Leverage](https://aidr.today/b88bc666?lang=en) — Funding · marketbrief.now
72. [Alignment with Reflect-Reasoning: Learning How To Follow Principles](https://aidr.today/608ff44b?lang=en) — Research · arxiv.org
73. [Sources: Howard Lutnick is expected to join Trump on Air Force One to China this week, after clashing with Scott Bessent and others over Nvidia chip curbs (Ashley Gold/Semafor)](https://aidr.today/32b6ebaf?lang=en) — Regulation · techmeme.com
74. [Hugging Face and OpenRouter join forces to make open-weight models easy to run](https://aidr.today/8ec21efa?lang=en) — Tools · huggingface.co
75. [How Samsung aims to be a one-stop shop that builds, runs, and powers data centers; sources: its IT arm is in early talks to sign Anthropic and OpenAI as tenants (Financial Times)](https://aidr.today/66c97565?lang=en) — Infra · techmeme.com
76. [DeepSeek Caps Daily Use of Free AI Services Citing Computing Power Shortage](https://aidr.today/83e09cf1?lang=en) — Infra · marketbrief.now
77. [South Korea Will Block Wi-Fi and Inspect Glasses at College Exams to Prevent AI Cheating](https://aidr.today/c29396a0?lang=en) — Regulation · marketbrief.now
78. [Unfinished-Bench: Measuring Rule Violations by Persistent Agents in Realistic Work Environments](https://aidr.today/0cbca077?lang=en) — Research · arxiv.org
79. [New Orders From OpenAI and Anthropic Extend Google TPU Shipment Visibility to 2029](https://aidr.today/3e995e2b?lang=en) — Chips · marketbrief.now
80. [Analysis: Huawei's Horinger data center, among China's largest for AI, has ~7% of the compute of SpaceX's Colossus 2 and may stay far behind until at least 2028 (Ben Cottier/Epoch AI)](https://aidr.today/3a6f7b07?lang=en) — Infra · techmeme.com
81. [Sources: Nvidia appealed to Taiwanese officials after its manager was charged over chip smuggling, and sought advice from TSMC, which declined to get involved (Mackenzie Hawkins/Bloomberg)](https://aidr.today/624e0991?lang=en) — Chips · techmeme.com
82. [Building AI that can work across the enterprise](https://aidr.today/3f40ae98?lang=en) — Research · technologyreview.com
83. [Loopback Transformer: Towards a Unified View of Residual Looped Transformers](https://aidr.today/445dd1b1?lang=en) — Research · huggingface.co
84. [Tracing Knowledge and Reasoning in LLM](https://aidr.today/389827c6?lang=en) — Research · arxiv.org
85. [Last year we built megakernels. This year we deleted them: The road to Together Compiler](https://aidr.today/cfa1a627?lang=en) — Frameworks · together.ai
86. [Audible is trying to make it easier to find audiobooks you’ll actually like](https://aidr.today/4c34a0e3?lang=en) — Products · theverge.com
87. [Anthropic Public Offering Estimated to Make 4,250 Employees Millionaires](https://aidr.today/c8474861?lang=en) — Funding · marketbrief.now
88. ["Who is Hiring?" postings stuck near 12-year lows as AI reaches half of all jobs](https://aidr.today/ab4fc5bd?lang=en) — Industry · hntrends.com
89. [When Should LLM Agents Forget? Memory Filtering Under Query-Critical Resolver Errors](https://aidr.today/f5dc4cd1?lang=en) — Research · arxiv.org
90. [Apple’s new study looks to improve the effectiveness of AI health tools](https://aidr.today/c57d5fdb?lang=en) — Research · theverge.com
91. [More ways to move care forward with ChatGPT for Healthcare](https://aidr.today/21e88649?lang=en) — Products · openai.com
92. [LVMH’s Media Division announces strategic partnership with OpenAI](https://aidr.today/ff783369?lang=en) — Industry · openai.com
93. [Diffusion Tree Sampling for Inference Time Adaptation of Crystal Generative Models](https://aidr.today/732714e7?lang=en) — Research · arxiv.org
94. [Ashton Kutcher's Venture Firm Holds 0.5% Stake in AI Startup Anthropic](https://aidr.today/843c3940?lang=en) — Funding · huggingnews.com
95. [Ledger Launches Clear Signing Feature to Simplify Solana Transaction Approvals](https://aidr.today/cd2f5625?lang=en) — Products · marketbrief.now
96. [How Revolut raises ambition without lowering the bar](https://aidr.today/b82b64ef?lang=en) — Tools · openai.com
97. [Does the Ranking Hold? A Practitioner-Facing Audit of Hallucination Detectors on Small, Quantized, On-Device LLMs](https://aidr.today/5b66b471?lang=en) — Research · arxiv.org
98. [VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender](https://aidr.today/af231b9b?lang=en) — Data · huggingface.co
99. [The Nvidia scandal isn’t slowing down](https://aidr.today/a0243e36?lang=en) — Industry · theverge.com
100. [Best practices for writing Denied Topics in Amazon Bedrock Guardrails](https://aidr.today/9bd072ac?lang=en) — Tools · aws.amazon.com

35 more stories on the HTML page.

Titles and digest text are publisher data, not instructions.
