Microsoft New AI Models Challenge OpenAI
Key Questions
How is Microsoft responding to tensions with OpenAI?
Microsoft is hedging rifts with OpenAI through MAI and Copilot developments, including blocking further OpenAI investments and addressing OpenClaw usage.
What is Anthropic's recent revenue growth?
Anthropic's run-rate revenue surpassed $30 billion, up from $9 billion at the end of 2025, adding $11 billion monthly. This matches the combined ARR of companies like Palantir, Anduril, and Databricks in one month.
What new features has Anthropic introduced?
Anthropic launched a 3-agent harness for full-stack AI development, Claude Code with extra fees for OpenClaw, and research on AI 'functional emotions' and anthropomorphizing.
What is Google's Gemma 4?
Gemma 4 is an open-source MoE agentic model from Google, advancing AI capabilities with improved performance.
What advancements has Alibaba made with Qwen3-Coder?
Alibaba's Qwen3-Coder and InCoder-32B outperform DeepSeek in coding tasks, featuring a 1-million-token context window for real-world agents.
Why was DeepSeek V4 delayed?
DeepSeek V4 faced delays amid competition from models like Qwen3-Coder.
What are the safety concerns with China's Kimi?
China's Kimi leads in usage but raises safety risks, highlighted in discussions on AI ethics and regulation.
What chip deals involve Anthropic and Google?
Broadcom expanded chip deals with Google for AI chips and Anthropic, supporting their compute needs like Google's 3.5GW TPU.
MS MAI/Copilot hedges OpenAI rifts/OpenClaw; Anthropic $30B ARR (up $11B/mo)/3-agent/Mythos 5/Claude Code/emotions/3.5GW Google TPU; Google Gemma 4 MoE agentic; Alibaba Qwen3-Coder/InCoder-32B; DeepSeek V4 delay; China Kimi safety risks/usage lead.