AI News Roundup — June 6, 2026
Trump, Sanders, and Sam Altman converged on the idea of public ownership stakes in AI companies, while Congress dropped a 269-page Great American AI Act seeking to preempt state AI laws for three years. Anthropic filed a confidential S-1 for its IPO at a $965B valuation, OpenAI launched ChatGPT Dreaming V3 for automatic memory synthesis, and Colorado's landmark AI anti-discrimination law is 25 days from taking effect.
Trump, Sanders, and Altman All Calling for Public Ownership Stakes in AI
In a surprising political convergence, President Trump, Senator Bernie Sanders, and OpenAI CEO Sam Altman are all separately advocating for the American public to hold equity stakes in AI companies. Sanders proposed a 50% public ownership threshold. Trump described a 'partnership with the American public' on Air Force One and said AI executives would visit the White House next week. Altman met with Sanders at the senator's request and said he shares the general idea but can't support the 50% figure. The convergence reflects growing bipartisan anxiety that AI's wealth is concentrating among a tiny group while ordinary Americans bear the disruption costs.
Congress Drops 269-Page 'Great American AI Act' — Three-Year Preemption of State Laws
Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA) unveiled a 269-page discussion draft of the Great American Artificial Intelligence Act — the most comprehensive federal AI framework ever introduced. The headline provision: a three-year freeze on all state AI laws for frontier AI models. Companies over $500M in annual revenue must publish Frontier AI Frameworks, report safety incidents to the federal government, allow cybersecurity audits, and support a $100M/year Center for AI Standards and Innovation. Labor unions (AFL-CIO, AFT, Association of Flight Attendants) issued a joint rejection: 'Hard no. This is a giveaway to the AI industry.'
Anthropic Files Confidential IPO S-1 at $965 Billion Valuation — $47 Billion Annual Run Rate
Anthropic filed a confidential draft S-1 with the SEC on June 1, formally starting the IPO clock. Revenue run rate hit approximately $47 billion in May 2026, up from ~$10B a year prior — roughly a 5x annual growth rate. The $65B Series H pushes post-money valuation to $965 billion, with a trillion-dollar debut as the analyst base case. A notable S-1 line item: Anthropic pays SpaceX $1.25 billion per month for compute through May 2029 ($15B/year to a single vendor). The S&P 500 declined to fast-track MegaCap IPOs — SpaceX, OpenAI, and Anthropic must still wait 12 months post-IPO for index inclusion.
ChatGPT Dreaming V3: Background Memory Synthesis Launches
OpenAI launched Dreaming V3, its most significant memory upgrade since the original ChatGPT rollout. Instead of requiring users to explicitly ask ChatGPT to remember things, Dreaming V3 runs a background synthesis process after each conversation ends — automatically cataloguing preferences, constraints, ongoing projects, and time-sensitive context. OpenAI reduced compute requirements for memory synthesis by ~5x, making the feature viable for free users. Privacy researchers flagged that 96% of ChatGPT memories in a recent study were created unilaterally — EU AI Act transparency rules taking effect in August 2026 will scrutinize this.
Colorado AI Act Takes Effect in 25 Days — Federal Bill Would Freeze It
Colorado's Consumer Protections for Artificial Intelligence Act — the first comprehensive state AI law in the US — is scheduled to take effect June 30, 2026. The law requires anti-discrimination protections in employment, education, financial services, healthcare, housing, and legal services for AI systems. The Great American AI Act's three-year state preemption provision, if passed, would freeze Colorado's law before it applies. But that's a discussion draft, not a passed bill — enterprises serving Colorado residents need compliance frameworks now.
China Dominates Humanoid Robot Production — But Demand Is the Problem
China accounted for ~85% of global humanoid robot shipments in 2025, with AGIBOT and Unitree each shipping over 5,000 units vs. a few hundred for US rivals like Figure AI and Tesla. Morgan Stanley projects China's humanoid sales to double to ~28,000 units in 2026. But demand is the constraint: most robots remain performative rather than functional, costs ($26,000-$99,000/unit) are prohibitive for mass deployment, and use cases are still limited to structured industrial environments.
OpenAI GPT-5.5-Cyber Expands to EU — Beats Anthropic to European Security Market
OpenAI granted the EU limited access to GPT-5.5-Cyber for vetted cybersecurity teams, EU businesses, governments, and EU institutions including the EU AI Office. Anthropic has not yet granted EU access to Mythos, giving OpenAI a tactical advantage in the race for European government and enterprise security contracts. OpenAI's EU expansion is part of its OpenAI for Countries initiative; Anthropic's Glasswing is blocked by data sovereignty concerns that a Milan office alone cannot resolve.