The $65B Series H closed at a $965B valuation, run-rate revenue crossed $47 billion, and Anthropic said Mythos-level models are coming to all customers in weeks.
Anthropic raised $65 billion in a Series H led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia, valuing the company at $965 billion post-money (Fortune). Run-rate revenue has crossed $47 billion, up from $30B in April, $14B in February, and $9B at year-end 2025 (Simon Willison). In the same announcement window, Bloomberg reported Anthropic told customers it has made “swift progress” on safety safeguards and plans to release Mythos-level models to all customers in “coming weeks” (Bloomberg), the first public timeline on capabilities the company previously said could exploit vulnerabilities in every major OS and browser.
The two facts connect. A company raising capital at a $965B valuation is also a company telling its customers the model it withheld for safety reasons is about to become generally available. Those are not separate stories.
Round led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia. Post-money: $965 billion, surpassing OpenAI as the most valuable private AI company. Run-rate revenue: $47 billion. Revenue trajectory: $9B ARR at end of 2025, $14B at Series G close in February, $30B in April, $47B now. Three-times jump in three months, attributed to Claude Code.
Anthropic stated it has made “swift progress” developing safety safeguards sufficient to release Mythos-class models broadly. Mythos has been restricted to Glasswing partners since April on the basis that it could identify and exploit vulnerabilities in every major OS and browser when directed. No specific date given. Reported by Bloomberg, Insurance Journal, and Fortune.
Plugins in .claude/skills load automatically without marketplace requirement. claude plugin init <name> scaffolds new plugins. EnterWorktree switches between Claude-managed worktrees mid-session. WSL improvements: image paste, screenshot paste, Windows Explorer drag-and-drop. tool_decision telemetry includes tool_parameters when OTEL_LOG_TOOL_DETAILS=1. Removed startup sandbox banner. Bug fixes: worktree isolation, background session reliability, terminal rendering, IDE stop button for subagents, /model picker version hints.
How to use: claude update or reinstall.
Hotfix: Opus 4.8 thinking blocks were being modified in Claude Code, causing API errors. Fixed.
No model, API, SDK, apps, or research releases in today’s window beyond the CODE entries above. Opus 4.8, its SDK support, and the API mid-conversation system blocks are covered in yesterday’s sweep. docs.anthropic.com returned HTTP 403 throughout this sweep: API release notes, Claude Apps, system prompts, and research pages all unverified via primary source. “Coming weeks” on Mythos is not a date.
Daily digest at 9 PM ET. Weekly magazine every Friday morning. Six labs, one feed. No spam, one-click unsubscribe.