Nvidia is investing $1 billion in AI startup Poolside at a $12 billion valuation, while also paying $6 billion to license its technology and hiring over 100 of its engineers. This strategic move aims to accelerate the development of Nvidia's Nemotron open-weight AI models and establish a U.S.-based open AI ecosystem to compete with Chinese rivals like DeepSeek.
Sources: HuggingNews, Reddit AI
A leaked National Republican Senatorial Committee memo warns AI firms that public backlash against terrestrial data centers poses a significant electoral risk, particularly in battleground states like Ohio. In response to land and energy constraints, infrastructure builders are seeking massive capital, with Nscale planning a $3 billion IPO and Starcloud raising $250 million to build AI data centers in Earth's orbit.
Sources: HuggingNews, Reddit AI
Server manufacturers have notified major clients including Microsoft, Google, and Oracle of a 15% price increase on upcoming AI servers due to surging HBM memory costs from suppliers like Samsung, SK Hynix, and Micron. This hike on Vera Rubin and Blackwell platforms may complicate the industry's large-scale AI infrastructure build-out.
Sources: HuggingNews
A humanoid robot named Tianzhuo won the 100-meter sprint at the World Humanoid Robot Games in Beijing with a time of 9.39 seconds, surpassing Usain Bolt's 2009 human world record of 9.58 seconds. This milestone highlights rapid advancements in autonomous robotic locomotion and dynamic control.
Sources: HuggingNews
Anthropic has hired Amir Salek, former head of Google's Tensor Processing Unit program, to build custom in-house AI hardware and reduce reliance on external vendors like Nvidia and Google.
Sources: HuggingNews
Alibaba is offering 710 million new shares to raise HK$80 billion (approximately $10.2 billion) to fund the development and expansion of its full-stack AI capabilities and infrastructure. This transaction marks the third-largest primary follow-on stock offering globally for the year and the largest in the history of the Hong Kong exchange.
Sources: HuggingNews
Growing public opposition to AI infrastructure has led to over 500 local data center bans across the United States, including Texas halting up to 1,800 projects. This grassroots backlash has escalated into a key political issue for the 2026 midterm elections, prompting OpenAI to hire a community engagement lead in Ohio to help mitigate local resistance.
Sources: HuggingNews
Amazon introduced Bedrock AgentCore, a managed platform designed to build, connect, and govern AI agents at scale with centralized security controls, identity management, and gateways for Model Context Protocol (MCP) traffic.
Sources: AWS ML Blog
Z.ai has stealthily launched a new multimodal AI model with a 1M token context window, identified as a Flash variant of the GLM family. The model achieves a 63% score on the DeepSWE coding benchmark, matching Claude Opus 4.8 and DeepSeek V4 Pro.
Sources: HuggingNews
SpaceXAI is offering on-site onboarding for corporate teams in San Francisco as part of a limited weekend launch, bringing its remote autonomous Grok Bot agent to enterprise workflows.
Sources: HuggingNews
Waymo has rolled out its new Ojai robotaxi fleet in San Francisco, Los Angeles, and Phoenix, featuring integrated Google Gemini AI assistants for hands-free in-cabin controls and navigation. This represents the company's most advanced in-cabin experience to date.
Sources: HuggingNews
An internal probe by Super Micro found no senior management involvement in a $2.5 billion illegal export-diversion scheme involving Nvidia chips to China, resulting in the termination of two employees and a contractor.
Sources: HuggingNews
Developers report that the new Qwen 3.8 27B open-source model matches frontier capabilities from previous years, showing superior OCR quality compared to Gemini 3.5 Flash Lite and sparking serious enterprise discussions about local hardware adoption.
Sources: Reddit AI
Alibaba has released the Qwen3.8-27B model, which the local LLM community reports significantly improves tool-calling reliability and agentic coding capabilities compared to its predecessor.
Sources: Reddit AI
Alibaba's open-source AI model, Qwen, has achieved a major milestone by surpassing three billion downloads globally. Concurrently, Chinese AI startup Moonshot AI has launched Kimi K3, which is recognized as the largest open-weight model released to date.
Sources: Michael Parekh AI
Tim Scarfe speaks with Ilia Shumailov and Alexander Panfilov about their paper, Stealing Reasoning Traces from Proprietary LLM APIs.The core bug sounds deceptively simple: providers return encrypted…
Relaunch week of AI in the AM brings together highlights from four live mornings and nine guests, centered on who checks frontier AI, how wide the gap is between lab-internal systems and public…
Models keep absorbing the harness into their weights — soon, it will be a harness for human attention rather than for the model.
Simile’s CEO about his journey from the viral Generative Agents to creating 8 Billion Digital Twins of every living human... and why it’s gone from fun exploration to very serious business.
Matt Pocock tells us about his /wayfinder skill, for greenfield projects or for when the way forward is unclear.
Astrophysicist Adam Becker, author of "What Is Real?", joins Tim Scarfe to take apart the futures Silicon Valley keeps selling: the 2045 singularity, mind uploading, Mars colonies, and the AI…
Max Hodak, co-founder and CEO of Science Corporation, joins Sarah Guo to discuss the future of vision, brain-computer interfaces, and the human experience. Max explains how Science’s PRIMA retinal…
Glean CEO Arvind Jain explains why model routing helps control AI costs for organizations, and how human feedback loops at scale improve its routing systems.
Rich Sutton, who helped pioneer reinforcement learning and wrote the seminal AI essay The Bitter Lesson, has now cofounded Oak Lab with his former student Khurram Javed. Their goal: to build agents…