AI News for June 2, 2026 — 7min.ai

[NEWS] Anthropic confidentially files for IPO at $965B valuation, racing OpenAI to public markets: Anthropic submitted a draft S-1 to the SEC on Monday, kicking off what could be the largest AI IPO in history.
[NEWS] Alphabet to raise $80B in equity for AI buildout, with $10B from Berkshire Hathaway: Google's parent disclosed Monday it will raise $80B in one of the largest equity offerings ever, including a $10B share sale to Berkshire Hathaway.
[NEWS] Florida sues OpenAI and Sam Altman over ChatGPT-linked violence and child safety: Florida Attorney General James Uthmeier filed an 83-page complaint accusing OpenAI of suppressing internal safety warnings and aggressively marketing ChatGPT despite knowing it could facilitate self-harm, violence, and child harm.
[RESEARCH] MiniMax M3 launches as open-weight model topping GPT-5.5 and Gemini 3.1 Pro at a fraction of the cost: Chinese AI startup MiniMax released M3 on Sunday, pairing a 1M-token context window, native multimodality, and frontier-tier coding performance at API pricing of $0.30 per 1M input tokens and $1.20 per 1M output tokens — roughly 5-10% the cost of leading proprietary models.
[TOOLS] Nvidia enters PC market with RTX Spark Arm chip, toppling Intel and AMD stocks at Computex: Nvidia CEO Jensen Huang unveiled RTX Spark at Computex Taipei — the company's first consumer laptop and desktop chip, combining a 20-core Arm CPU co-developed with MediaTek, up to 6,144 Blackwell GPU cores, and 128GB unified LPDDR5X memory.
[NEWS] Meta AI support chatbot exploited to hijack Obama White House Instagram and other high-profile accounts: Hackers used Meta's AI support bot to take over notable Instagram accounts including the Obama White House page and the US Space Force Chief Master Sergeant by simply asking the chatbot to swap the email address tied to a target profile.
[RESEARCH] OpenAI AI model disproves 80-year-old Erdős conjecture — Fields Medal winner calls it a "milestone in AI mathematics": OpenAI announced in mid-May that an internal AI model disproved the Erdős unit distance conjecture, a discrete geometry problem that had stumped mathematicians for 80 years.
[RESEARCH] Nvidia launches Cosmos 3 world model and Alpamayo 2 driving model at GTC Taipei: Nvidia unveiled Cosmos 3, a vision-language-action omnimodel that processes text, images, video, audio, and action data in a single system for robots, autonomous vehicles, and surveillance.
[NEWS] OpenAI frontier models and Codex go generally available on AWS: OpenAI made its frontier models and Codex coding agent generally available on AWS, giving enterprises a way to consume OpenAI through familiar AWS environments, controls, and procurement workflows.
[NEWS] Sam Altman declares robotics OpenAI's next frontier, backs stealth startup Alfred: Altman posted on X that robotics is OpenAI's next focus, with short-term work aimed at robots supporting skilled infrastructure workers and a long-term vision of "everyone having a personal robot doing anything they need." OpenAI is hiring across hardware, ML, systems, and operations roles, with the effort growing out of Aditya Ramesh's world simulation team.

About 7min.ai · Subscribe to newsletter