DEDUPEDTop 10 today
-
The US barred civilian federal agencies from using DeepSeek.
The Trump administration told civilian agencies to stop using products or services from the Chinese AI startup, citing national security risks tied to its operations in China.
Full story -
Amazon launched TorchNeuron.
An open-source layer that runs PyTorch code on Amazon's Trainium chips with no rewriting, in public beta. It goes after Nvidia's real moat, which is software.
Full story -
Tenstorrent raised $1.5 billion at a $6.2 billion valuation.
The Nvidia challenger took the money to stay independent after acquisition interest earlier this year.
Full story -
MiniMax released M3.1 Flash.
$0.15 per million input tokens for a large mixture-of-experts agent model. It also confirmed the stealth model Space Bunny was an early build.
Full story -
Anthropic says junior hires fell from 27% to 9% of its technical hires.
Its own workforce study also says staff agents logged far more hours than the people running them.
Full story -
Anthropic will put 5% of its equity into a public benefit fund.
More could follow as its valuation grows, and the fund is meant to address AI job disruption. Stockholders still need to approve.
Full story -
Microsoft added Guardian Agents to Project Perception.
Agents are now treated as potential insiders, supervised continuously, with per-agent permission controls.
Full story -
World launched World ID for banks.
A way for banks to verify that a human, or an authorized AI agent, is on the other end.
Full story -
Replit made OpenAI its default provider over Anthropic.
Claude stays selectable for users, so this is about cost and partnership.
Full story -
China added 40 technologies, including large AI models, to its export ban and restriction list.
The ban on DeepSeek and Beijing's own export curbs now point the same direction: AI is being treated like a strategic export.
Full story
Also ranked today
-
OpenAI Logs 4 More Rogue Agent Cases, 2 Critical
-
Nvidia launches new open models to run agents in the enterprise and on PCs
Nvidia has released new open-weight models including the 120B Nemotron 3.5 Super VL and lightweight Bolt models designed for edge computing, local AI hardware, and always-on enterprise agents.
-
Deepseek And Huawei Are Collaborating On New AI Chip Designs To Challenge Nvidia
DeepSeek is working closely with Huawei Technologies to shape the design of future AI chips, two people familiar with the matter said, in a collaboration that could help China build frontier AI models without Nvidia’s technology.
-
How China's proposed AI export controls could affect open-weight ecosystem
-
Nvidia wins fast-growing AI startup Modal for new central processors
Modal plans to run its Sandboxes product on Nvidia’s Arm-based Vera CPU, its first step away from x86 chips. In Modal’s own tests, Vera handled up to twice as many tasks as an Intel server when packed with containers.
-
A Lone Hacker Used AI Agents to Breach More Than 100 Companies
-
NVIDIA Open Sources Run:ai Under Apache 2.0 for Kubernetes GPU Orchestration
NVIDIA has released the full NVIDIA Run:ai platform as open source software. The source code ships under the Apache 2.0 license on GitHub.
-
India's Cybersecurity Officials Say AI Agents Accessed Live IT and Communications Systems
-
Nvidia partners with TSMC to help Samsung produce high-speed AI memory NVHBM
-
UMG and Sony ask court to force DistroKid to hand over AI music data for Suno case
-
Meta removes AI training clause after backlash over web scraping concerns
-
Inception Releases Mercury 3, Its Flagship Diffusion LLM
How this list works
We read 1,000+ AI sources every day. Stories about the same event get merged into one, then ranked on impact, how new it is, how many independent outlets covered it, and how useful it is to you. "Everyone covered this" means most of the AI press ran it. "Almost nobody caught this" means it is a quiet signal worth knowing early. Updated every morning. RSS feed.