The CEOs of OpenAI, Anthropic, and Google DeepMind are all heading to the G7 together: as models launch, IPOs file, and the first US AI law counts down to June 30.
All three AI giants head to the G7, together, for the first time
Sam Altman (OpenAI), Dario Amodei (Anthropic), and Demis Hassabis (Google DeepMind) will all attend the G7 summit in Évian-les-Bains, France on June 15–17. It marks the first time leaders of all three major frontier AI labs will appear at the G7. The agenda includes voluntary AI safety commitments, youth protection online, and warnings about AI-enabled cyber and biological threats, a joint concern the three CEOs have already put in writing to the US Congress.
Why it matters: When the world's seven richest democracies sit down and the AI industry sends its three biggest players, something substantive is being decided. Even "voluntary" G7 commitments shape what laws look like next.
Anthropic's Claude Fable 5 is live, its most powerful model yet
Anthropic launched Claude Fable 5, the first publicly available model in its new Mythos class, on June 9. It is already accessible inside Claude Code and the Claude App for Pro and Max subscribers, and will remain subscriber-only until June 22. Pricing sits at $10 per million input tokens and $50 per million output, premium, but significantly cheaper than the earlier Mythos Preview. Benchmarks position it above all previous Claude models.
Why it matters: If you use Claude regularly, you already have access to something meaningfully better. If you have been on the fence about a paid plan, this is the clearest reason yet to try one.
OpenAI files for IPO, following Anthropic into the public markets
OpenAI submitted a confidential draft registration to the US Securities and Exchange Commission on June 8, setting the stage for a public offering. It follows Anthropic's earlier confidential filing. OpenAI was last valued at $852 billion. Both companies are still burning significant compute costs, which makes the race to public markets a bet on sustained investor confidence in frontier AI growth.
Why it matters: Once AI labs are publicly traded, quarterly earnings reports and shareholder pressure will shape their decisions alongside safety considerations. The incentive structure of the whole industry is about to change.
Colorado's AI law takes effect June 30, the first of its kind in the US
Colorado's AI Act, the first comprehensive state-level AI law in the United States, becomes enforceable in 17 days. It requires companies that use AI for significant decisions (hiring, lending, healthcare) to conduct impact assessments, implement bias-prevention measures, and run security risk management programmes. Other states including California, New York, and Illinois have their own AI laws either already in force or taking effect by 2027.
Why it matters: If your employer, bank, or health provider uses AI to make decisions about you, Colorado just gave you legal rights. The rest of the US is watching to see how enforcement plays out over the next few months.
Equal AI's spam-filtering assistant hits 1 million monthly users in India
Equal AI, an AI-powered call assistant built specifically for the Indian market, has crossed 1 million monthly active users. Its core function is filtering spam and scam calls before they reach the user, a problem disproportionately large in India, where phone fraud is widespread. The milestone comes as India's AI Governance Guidelines (published earlier in 2026) begin shaping how AI products must operate domestically, and as the India AI Impact Summit concluded its latest showcase of AI for farming, skilling, and local language applications.
Why it matters: This is practical AI solving a daily Indian problem at real scale, not a pilot. It also shows the consumer AI market in India is large enough to build for, not just adapt Western products toward.
Gemini 3.5 Pro is almost here, with a 2-million-token context window
Google announced Gemini 3.5 Pro at I/O 2026 but only its smaller sibling, Flash, went live in May. Pro remains in internal testing and limited Vertex AI preview, with Sundar Pichai indicating a June arrival. When it ships, it will carry a 2-million-token context window (double Flash's capacity) making it capable of processing entire codebases, long legal documents, or hours of meeting transcripts in a single session. Gemini 3.5 Live Translate is already available in 70+ languages.
Why it matters: A 2-million-token window is a qualitative shift, not just a number. Tasks that required splitting content across sessions will become single-pass. Developers and researchers should watch for the Pro launch this month.
Two AI giants going public: at the same time
Anthropic and OpenAI are both on track to go public. That has never happened before. When frontier AI labs answer to public shareholders, the pressure to show quarterly growth can conflict with long-term safety investments. Watch how each company talks about safety commitments in its S-1 prospectus, those disclosures will reveal what risks they consider material enough to tell investors about.
Google trains 6 million educators for free + AI biosafety letter to Congress
Google launched its AI Educator Series on May 13, offering free AI literacy training to all 6 million K-12 and higher-ed teachers in the US, with new modules each month. Separately, OpenAI, Anthropic, and Google jointly wrote to Congress urging stricter regulations on synthetic DNA and AI-assisted biological threats, a rare moment of alignment among competitors, and one that signals how seriously frontier labs view the biosecurity risk.