Anthropic has raised $65 billion in Series H funding at a $965 billion post-money valuation — vaulting past OpenAI's last reported valuation of $730 billion. The round, the largest in the history of the AI industry, was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, and makes Anthropic one of the most valuable private companies in the world within striking distance of a trillion-dollar valuation.
The company reported a $47 billion annualized revenue run-rate, underscoring surging enterprise demand for Claude. Simultaneously, Anthropic inked massive compute agreements: Amazon will install up to 5 gigawatts of new infrastructure, Google and Broadcom will deliver 5 gigawatts of next-generation TPU capacity, and SpaceX will provide GPU access at its Colossus facilities.
OpenAI strikes back — big Codex update
One day after Anthropic's announcement, OpenAI shipped a sweeping update to Codex, its autonomous development platform. Codex can now control desktop applications on Windows and macOS, generate images via gpt-image-1.5, schedule future tasks, and retain context across sessions through an optional Memory feature.
The move is a direct response to the runaway success of Anthropic's Claude Code and reflects an aggressive internal resource shift at OpenAI to close the gap. Codex can now run in the background alongside other tools, deploy multiple agents in parallel, and browse the web from within the app — including the ability to comment on pages to give precise instructions.
Microsoft consolidates Copilot into a super app
Fortune reported that Microsoft is building a super app combining GitHub Copilot, Copilot Chat, the Copilot Cowork tool, and a new agentic workflow capability internally called Autopilot — all under one roof. The project, operating under the slogan "Delivering one Copilot," is expected to launch by the end of summer and may be referenced at Microsoft's Build conference in San Francisco next week.
The move addresses a persistent customer pain point: Copilot products scattered across Excel, GitHub, Teams, and Windows. The super app will centralize them with a single interface, including a toggle between personal and enterprise accounts.
Illinois passes toughest US AI safety law
Governor JB Pritzker announced he will sign a new AI bill passed by the Illinois legislature. The law requires independent audits at AI companies, whistleblower protections, and transparency mechanisms — going further than recently enacted AI laws in New York and California.
The legislation joins a growing global trend toward AI regulation and positions Illinois as a major player in shaping the rules of engagement for the industry.
In other news
- Microsoft launched a preview of Copilot Health AI, capable of analyzing medical records.
- OpenAI is sunsetting the Canvas interface on GPT-5.5 models.
- Google is adding Gemini chat sharing via Google Drive starting June 3.
- Figma Make is now generally available — all users can build apps using natural language.
- An Amazon exec warned employees against "using AI for the sake of AI" after an internal leaderboard incentivized fake chatbot usage.
- OpenAI's Codex can now generate and iterate on images with gpt-image-1.5, and has new plugins for GitLab, Atlassian Rovo, and Microsoft Suite.
The bottom line
Two stormy days in AI: Anthropic and Microsoft are maneuvering for pole position, OpenAI is responding aggressively, and regulators are beginning to close the net. The combined valuation of the leading AI companies has long since crossed into the trillions, and the battle for users — and lawmakers — is just getting started.
