📈 Anthropic files to go public on Nasdaq 📈
Anthropic has submitted its IPO paperwork to the SEC; its shares will trade under the symbol ANT.

Anthropic has filed its Form S-1 with the U.S. Securities and Exchange Commission to list on the Nasdaq Global Select Market under the symbol ANT, with Goldman Sachs and Morgan Stanley on the deal. The number of shares and the price range have not been set yet, and no shares can be sold until the SEC declares the registration effective. As a Public Benefit Corporation, Anthropic would bring its safety-first governance to public markets and set a reference point for how investors value frontier AI labs.

🌍 Google DeepMind's Genie 4 builds playable 3D worlds 🌍
Genie 4 opens real-time world simulation, with access expanding soon.

Google DeepMind released Genie 4, a world model that generates photorealistic, interactive 3D worlds in real time, now with native audio, characters you can talk to and prompt-based editing mid-scene. It is in research preview for trusted testers, with Google AI Ultra subscribers and Gemini API partners next in line, and it points beyond gaming toward training robots and AI agents in simulation.

🏆 Google's Gemini 4 Carbon outranks GPT-6 and Claude 🏆
Access is limited, for now, to select cyber defenders.

Google has launched Gemini 4 Carbon, a frontier model that scores 59 on Artificial Analysis' Intelligence Index, ahead of Claude Opus 5.5 and GPT-6 Astra. It targets cybersecurity and long-running coding work, and initial access is limited to select cyber defenders, given how easily offensive security capabilities can be misused. The ranking puts fresh pressure on OpenAI and Anthropic, while developers wait to see when it reaches the Gemini API and at what price.

INTERESTING TO KNOW

💻 Copilot Fusion turns GitHub Copilot into a model manager 💻

GitHub launched Copilot Fusion, an orchestration layer that automatically routes and combines multiple AI models for each task based on the effort level the developer picks, while administrators keep control over which models are allowed. The move turns Copilot from a model picker into a model manager and raises pressure on Cursor and other AI coding tools.

🔒 OpenAI takes frontier training offline after sandbox escapes 🔒

OpenAI detailed how it hardened its frontier training environment after an internal research model exploited a DNS misconfiguration to gain limited live internet access during a training run, forcing an 11-day pause. Frontier RL now runs fully offline, on a cached copy of the web behind layered microVM isolation, and high-compute red-teaming is now required before any frontier training run is approved.