Top Stories
Jensen Huang’s GTC Keynote Today
Nvidia’s biggest event of the year begins today in San Jose, with CEO Jensen Huang delivering what’s expected to be a two-hour keynote at 2 p.m. ET (11 a.m. PT). The livestream is available at nvidia.com without registration.
Huang will cover the full AI stack: chips, software, models, and applications. Key expectations include details on the Vera Rubin architecture (Blackwell’s successor), updates on OpenClaw (described as “the fastest-growing open source project in history”), and updates on Nvidia’s robotics initiatives including Isaac and GR00T.
With 30,000 attendees from 190 countries expected, GTC 2026 is positioned around the “Agentic AI Inflection Point” - the shift from AI that responds to prompts to AI that takes autonomous action.
Source: NVIDIA GTC Blog
Claude Surges to #1 as Users Boycott ChatGPT Over Pentagon Deal
In one of the stranger turns in AI industry history, Anthropic’s Claude climbed to the top of Apple’s App Store just days after the Pentagon banned the company and labeled it a “supply chain risk.”
The catalyst: OpenAI’s announcement that it had secured a $200 million Pentagon contract after Trump’s administration blacklisted Anthropic. Users spread “Cancel ChatGPT” campaigns across Reddit and X, posting guides for deleting accounts and migrating to Claude. Anthropic’s annualized revenue has reportedly topped $2.5 billion from its coding agent alone.
The irony is thick. Anthropic drew red lines against using Claude for mass surveillance and autonomous weapons. The Pentagon wanted “all lawful purposes.” Negotiations collapsed. OpenAI stepped in. And now Anthropic is more popular than ever with consumers, even as it’s locked out of government work.
OpenAI and Google Employees Back Anthropic’s Pentagon Lawsuit
More than 30 employees from OpenAI and Google DeepMind filed a statement supporting Anthropic’s lawsuit against the Defense Department.
Anthropic is the first American company publicly labeled a “supply chain risk” - a designation historically reserved for foreign adversaries. In a leaked memo, Dario Amodei accused OpenAI’s leadership of “gaslighting” the public and criticized Altman for giving “dictator-style praise to Trump.”
The tech industry appears genuinely split on whether working with the military under current terms is acceptable.
Sources: TechCrunch, Time
Quick Hits
-
Meta delays “Avocado” again: Meta’s next-gen AI model is pushed to May 2026. While it beats Gemini 2.5, it lags Gemini 3.0 and the latest from OpenAI and Anthropic. Leadership has discussed temporarily licensing Gemini from Google. Trending Topics
-
Hugging Face CEO: “LLM bubble, not AI bubble”: Clem Delangue distinguishes between an AI bubble (which he doesn’t think exists) and an LLM bubble (which he does). He argues smaller, specialized models will make more sense going forward. TechCrunch
-
Gemini comes to Workspace: Google is rolling out Gemini-powered features for Docs, Sheets, Slides, and Drive. Beta starts today for AI Ultra and Pro subscribers. Google Blog
-
EU AI Act approaching full enforcement: The EU AI Act reaches full enforcement on August 2, 2026, though the Commission’s Digital Omnibus package may delay high-risk AI rules by over a year. 61 Data Protection Authorities issued a joint statement on AI-Generated Imagery. Blank Rome
Worth Watching
GTC keynote at 11 a.m. PT today is the main event. Watch for OpenClaw details, Vera Rubin architecture specs, and any signals about Nvidia’s relationship with the major AI labs (Jensen Huang recently said Nvidia is “pulling back” from OpenAI and Anthropic investments, raising questions about why).
The Anthropic situation remains volatile. A company simultaneously banned by the government and surging in consumer popularity is unprecedented territory. The lawsuit outcome could reshape how AI companies negotiate with defense agencies - and whether drawing ethical red lines is commercially viable.
The LLM bubble narrative is gaining traction. If Delangue is right about an LLM-specific bubble, we may see a shift toward smaller, specialized models and non-LLM AI applications in biology, chemistry, and robotics. Worth tracking whether GTC announcements reflect this diversification.