Anthropic Eyes Trillion-Dollar IPO After $965 Billion Valuation
The AI safety company founded by former OpenAI researchers is on a trajectory that few predicted even a year ago. Anthropic has raised $65 billion in Series H funding led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, valuing the company at a staggering $965 billion — surpassing OpenAI and setting the stage for what could become the first trillion-dollar AI IPO.
The company confidentially filed IPO paperwork with the SEC on June 1, with a public listing on the Nasdaq or NYSE expected as early as October 2026. Investment bankers now consider a debut above the $1 trillion mark the base case, assuming markets cooperate. Anthropic's run-rate revenue reportedly crossed $47 billion earlier this year, underscoring the explosive commercial adoption of its Claude model family.
The filing comes as Anthropic continues to strengthen its research bench. John Jumper, the Nobel Prize-winning creator of AlphaFold, announced his departure from Google DeepMind to join the Claude maker — giving Anthropic a level of scientific legitimacy that no benchmark score could provide. Anthropic has also scheduled an AI for Science event for June 30, hinting at expanded ambitions beyond language models.
GPT-5.6 Launch Window Collapses — Now Expected in July
OpenAI's much-anticipated GPT-5.6 will not ship in June after all. The late-June launch window that prediction markets had priced at 83% probability has collapsed to roughly 18%, with the model now most likely arriving in July 2026.
OpenAI's chief scientist Jakub Pachocki reportedly told staff that GPT-5.6 would represent a "meaningful improvement" over GPT-5.5, with expected features including a 1.5-million-token context window — a dramatic jump from GPT-5.5's 400K — along with improvements on FrontierMath benchmarks, UI generation, and long-horizon coding tasks. An alignment fix is also said to be included in the release.
The delay comes at an awkward time for OpenAI, which has also confidentially filed for an IPO. With Anthropic's Fable 5 leading on software engineering benchmarks and Chinese lab Zhipu AI releasing GLM-5.2 with a 62.1 SWE-bench Pro score (vs. GPT-5.5's 58.6), the pressure to deliver GPT-5.6 is mounting.
Google DeepMind's Talent Hemorrhage: Four Senior Departures in Six Days
What started as a shockwave has become a full-blown crisis for Google DeepMind. Following the headline-grabbing departures of Noam Shazeer to OpenAI and John Jumper to Anthropic, two more senior Gemini researchers — Jonas Adler and Alexander Pritzel — are also expected to join Anthropic, bringing the total to four high-profile exits in just six days.
The departures have rattled investors. Alphabet shares experienced their worst single-day drop in over a year, falling approximately 5% on concerns about Google's ability to retain top-tier AI talent. Shazeer's exit stung particularly hard: Google had paid $2.7 billion to acqui-hire him and part of his Character.AI team barely two years ago.
DeepMind CEO Demis Hassabis sought to calm fears, saying that movement between leading labs was expected in the current market and arguing that Google still maintains the largest and broadest research team in the industry. But with Gemini 3.5 Pro still awaiting its full release and competitors strengthening their rosters, the optics are unfavorable.
Congress Proposes the Great American AI Act — First Federal AI Framework
In a rare bipartisan effort, Representatives Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a discussion draft of the Great American Artificial Intelligence Act (GAAIA), which would create the first comprehensive federal framework for governing AI in the United States.
The 250-page bill is organized into four titles covering frontier AI governance, workforce protection, cybersecurity, and international cooperation. Key provisions include requiring frontier AI developers to disclose model information and submit to third-party audits, creating an AI Workforce Research Hub within the Department of Labor, and authorizing $100 million per year for the Center for AI Standards and Innovation.
Perhaps most notably, the bill takes a novel approach to federalism: AI development would be regulated at the federal level, while deployment and use regulation remains with the states. This comes as Colorado's own AI law — originally set to take effect June 30 — has been amended and delayed to January 2027, signaling that even state-level regulators are still working out how to govern this technology.
OpenAI Expands GPT-5.5-Cyber for Defenders
OpenAI continued its push into cybersecurity with an expanded rollout of GPT-5.5-Cyber, a specialized model designed to help defenders identify and patch vulnerabilities. The model achieved an 85.6% score on CyberGym, outpacing the standard GPT-5.5 (81.8%) and even Anthropic's Mythos 5 on certain benchmarks.
Access remains gated through OpenAI's Trusted Access for Cyber program, available to vetted security organizations including Akamai, Cisco, Cloudflare, CrowdStrike, Fortinet, Oracle, Palo Alto Networks, and Zscaler. OpenAI also launched the Daybreak Cyber Partner Program and is funding Patch the Planet, an initiative helping open-source maintainers remediate vulnerabilities identified with AI-assisted tools.
The expanded capabilities include deeper analysis across large codebases, identification of security-relevant components, and automated development and testing of software patches — a clear signal that domain-specific AI models are becoming a serious market category.