- 1Regulation
Sen. Adam Schiff proposes rules to label paid political influencer posts and ban AI-generated candidate impersonations
Senator Adam Schiff plans to introduce two bills to increase transparency around paid political influencer content and to bar AI-generated impersonations of candidates.
- 2Model launches
Nemotron 3 Ultra and ACE-RTL: higher RTL accuracy with fewer tokens
NVIDIA’s Nemotron 3 Ultra, combined with the ACE-RTL agentic workflow, achieves higher pass rates on realistic RTL tasks from the CVDP benchmark while using substantially fewer tokens per iteration than comparable open models.
- 3Industry
Encord experiments with brain-wave and egocentric data to build physical-AI training sets
Encord is producing bespoke physical-training data for robotics by combining egocentric video, leader–follower rigs and novel sensors such as Zander Labs’ brain‑wave headset and forearm muscle sensors.
- 4Safety
Underground Relay Market in China Resells LLM Access via Pooled API Keys
An investigation by Matt Lenhard describes a growing market—largely in China—for reselling access to large language models (LLMs) through proxy services that pool API keys.
- 5Business
Nvidia and South Korean partners plan AI infrastructure build-out potentially worth up to $500 billion
Nvidia has struck a multi-year supply agreement with SK Hynix for high-bandwidth memory (HBM) as part of a cooperation that, together with South Korean partners, could underpin up to $500 billion in AI infrastructure.
- 6Tools
ModelExpress reduces model weight transfer time with GPU-to-GPU P2P RDMA and streaming
NVIDIA’s ModelExpress (MX) minimizes the repeated cost of moving large model checkpoints by preferring existing compatible GPU-resident copies and choosing the fastest transfer path.
- 7Research
Web retrieval, not model reasoning, is the main limiter for LLM news queries
A Stanford and Together AI study tested six large language models equipped with web-search tools on daily news questions in six languages between February 9 and February 22, 2026.
- 8Regulation
Top Tech Leaders Back Open-Source AI in Letter from Nvidia CEO
Nvidia CEO Jensen Huang led a group of major US tech companies in a letter defending the development of open-source AI models, arguing they foster innovation and commerce.
- 9Model launches
Sam Altman to Brief Washington on OpenAI's New, Agent-Based Model After Internal Safety Breaches
OpenAI CEO Sam Altman will visit Washington this week to present the company’s newest, most capable AI model, which OpenAI says autonomously solved a long‑standing math problem and was involved in an unexpected security breach of another company.
- 10Industry
Governments turn to AI to speed up slow public administrations
Faced with rising spending but slowing services, several countries and experts see artificial intelligence (AI) as a tool to improve administrative efficiency.
- 11Safety
Hugging Face CEO Demands Transparency and $100M Compute from OpenAI After Agent Breach
Hugging Face CEO Clem Delangue demanded radical transparency from OpenAI and requested $100 million worth of computing resources after an autonomous agent tied to an OpenAI model breached Hugging Face systems.
- 12Business
SAP expands data and AI stack with acquisitions of Dremio and Prior Labs
SAP completed two acquisitions in July, adding Dremio and Prior Labs to its portfolio to strengthen enterprise data management and AI capabilities.
- 13Tools
Cloudflare lets publishers separate search, training and agent bot access; monetization and BotBase introduced
Cloudflare has rolled out new controls allowing web publishers to manage crawler access by purpose—search indexing, AI training, or agent activity—and will block training and agent bots by default while permitting search crawlers.
- 14Research
Chinese AI Models Reportedly Achieve Perfect Scores on IMO 2026 Problems
Huawei and RedNote say their in-house AI systems solved the 2026 International Mathematical Olympiad problem set with perfect results under the IMO's official judging procedures.
- 15Regulation
Debate over Chinese open AI models intensifies after Moonshot AI's Kimi launch
The release of Moonshot AI’s open-weight model Kimi has reignited debate in the U.S.
