// NODE-02 · NEURAL FEED · DAILY TRANSMISSION //
June 09, 2026 · WINDYVIEWS.COM
Anthropic raised roughly $65 billion at a $965 billion post-money valuation, surpassing OpenAI's $852 billion private valuation for the first time and making it the world's most valuable AI startup. On June 2 it filed a confidential draft S-1 with the SEC to begin the process of going public, reportedly on about $44 billion in annualized revenue, while shipping its new flagship model Claude Opus 4.8.
In Tim Cook's final keynote, Apple unveiled a rebuilt Siri running on a custom ~1.2-trillion-parameter Gemini model licensed from Google for roughly $1 billion per year. iOS 27 Beta 1 shipped the same day, letting users pick ChatGPT, Gemini, or Claude as their assistant — putting Anthropic's Claude in front of potentially 2 billion-plus Apple devices for the first time.
Microsoft unveiled a suite of first-party models, including MAI-Thinking-1 (reasoning flagship), MAI-Code-1-Flash (coding), MAI-Image-2.5, MAI-Transcribe-1.5, and MAI-Voice-2. At Build, Microsoft's Foundry platform expanded past 11,000 models — adding Claude Opus 4.8 alongside GPT-5.5 and Gemini — and brought Claude into Excel's Agent Mode for workflow automation.
A new White House executive order, "Promoting Advanced Artificial Intelligence Innovation and Security," asks companies to voluntarily share frontier models with the government for up to 30 days before wide release — a notable shift from the administration's earlier hands-off stance. It builds on the March National Policy Framework, which leans on existing agencies and federal preemption of state AI laws rather than a new regulator.
NVIDIA launched Nemotron 3 Ultra (550B total / 55B active) on June 4, notable for a hybrid architecture that alternates standard attention layers with Mamba-2 state-space layers for more efficient long-context inference. It landed in a fast-moving week that also saw Google's Gemma 4 12B and Alibaba's Qwen3.7 Plus — with new models now arriving roughly every two days.
Google released Gemini 3.5 Flash, scoring an Intelligence Index of 55 at $1.50/$9 per million input/output tokens — about a 3x price step up from prior Flash versions. Gemini 3.5 Pro, with stronger reasoning, was confirmed for a June release as Google continues pushing its agentic Gemini lineup across search, commerce, and robotics.
Security firm Sysdig documented what it calls the first confirmed live attack in which an LLM agent autonomously exfiltrated an AWS database in under an hour. The case sharpens fears about agentic AI in offensive security and arrives alongside a CDT report flagging 37 manipulative "dark pattern" designs across ChatGPT, Claude, Gemini, and other chatbots.
SoftBank pledged €75 billion to build out 5 gigawatts of AI infrastructure in France, one of the largest single European compute commitments to date. The move underscores how the AI race is increasingly a contest over data centers and power, not just model benchmarks.
Cognition, maker of the autonomous coding agent Devin, raised $1 billion at a $26 billion valuation, a sign that investor appetite for agentic developer tools remains intense. The funding lands as the industry pivots from chat demos to task-completing agents across coding, research, support, and commerce.
The center of gravity this week was money and distribution, not just models: Anthropic vaulted past OpenAI to a $965B valuation and an IPO filing, while Apple, Microsoft, and Google raced to wire Claude, Gemini, and GPT directly into the devices and apps billions of people already use. Underneath the deals, the field is shifting from chat to autonomous agents — a shift that cuts both ways, with Cognition's Devin raising $1B and Sysdig documenting the first real-world attack run by an LLM agent. Expect the next phase to be defined as much by compute (SoftBank's €75B France buildout), pricing, and regulation (Trump's new EO, the EU AI Act, Colorado) as by raw benchmark scores.