Global AI News Daily — 2026-07-09

Top Stories

JADEPUFFER: First Fully Autonomous AI Ransomware Attack Confirmed

A major cybersecurity incident demonstrates a stark escalation in AI-powered threats. JADEPUFFER launched 600 autonomous payloads without human intervention, marking the first documented fully autonomous ransomware attack. The incident underscores growing concerns about AI safety and the dual-use risks of increasingly capable models, even as companies compete to release more powerful systems.

Anthropic Overtakes OpenAI on Revenue at $47 Billion

In a stunning market shift, Anthropic has surpassed OpenAI in annual revenue, reaching $47 billion. The milestone reflects Anthropic's aggressive expansion into enterprise contracts and government deals, including a major agreement with California that offers state agencies 50% discounts on Claude. This development challenges OpenAI's long-held market leadership position and signals investor confidence in Anthropic's safety-first approach.

Google Delays Gemini 3.5 Pro Launch to July 17 for Full Architectural Rebuild

Google DeepMind announced a delay for its flagship Gemini 3.5 Pro model, pushing release from early July to July 17, 2026. The company is scrapping its existing 2.5 Pro architecture entirely for a complete rebuild, citing performance issues flagged by enterprise testers in reasoning and coding tasks. The delay reflects industry-wide pressure to prioritize quality and reliability over speed to market.

Company Watch

Chinese AI Models Gain Ground as U.S. Pricing Surges — DeepSeek, Z.ai, and other Chinese AI developers are winning market share from OpenAI and Anthropic as U.S. model pricing becomes prohibitive. Developers cite competitive performance coupled with significantly lower costs, forcing American companies to reconsider their pricing strategies amid customer pushback on token spending.

AI Safety Rankings Released; Anthropic Leads, OpenAI Slips — In a comprehensive safety assessment, Anthropic earned a C+ rating (highest), while OpenAI fell from C+ to C. Google DeepMind ranked third. The findings highlight the growing importance of safety credentials in institutional buying decisions and competitive differentiation.

Model Distillation Threatens Profit Margins for AI Giants — Rival companies are using knowledge distillation to replicate cutting-edge models at a fraction of the cost, threatening the high-margin business models of Anthropic, OpenAI, and Google. Distilled models achieve near-parity performance while dramatically reducing inference costs, forcing market leaders to defend their pricing.

Policy & Regulation

UN Convenes Global AI Governance Summit Amid "Catastrophic Harm" Warnings — The Global Dialogue on AI Governance opened in Geneva on Monday, bringing together governments, tech companies, academics, and civil society to wrestle with how to regulate AI faster than the technology evolves. The summit highlights international consensus that AI poses systemic risks requiring coordinated governance frameworks.

Democratic Security Framework Emerges as AI Governance Benchmark — Swiss Federal Councillor Alain Berset framed AI governance as a test of democratic security built on rights, institutions, and law. The statement reflects a shift toward viewing AI regulation as inseparable from broader efforts to protect democratic institutions in an era of technological disruption.

Industry Moves

General Intuition Targets "ChatGPT Moment" for Robotics — The startup aims to become the foundation model layer for physical AI, providing a base model for robotics companies rather than building robots directly. The announcement signals growing recognition that embodied AI may follow the same large-model scaling patterns that transformed NLP.

Tech Layoffs Accelerate with AI as Stated Justification — Companies including Oracle (21,000 job cuts redirected toward AI data centers) and Block (4,000 jobs, 40% of workforce) cited AI-driven efficiency gains. Block CEO Jack Dorsey stated smaller, flatter teams powered by AI intelligence tools represent "a new way of working" most companies are late to adopt.

1 sources

techcrunch.com

techcrunch.com

What to Watch

1. Google Gemini 3.5 Pro Launch (July 17) — Watch for architectural changes and whether the rebuilt model addresses the reasoning/coding weaknesses that prompted the delay. Market share implications for Google's enterprise sales are significant.

2. OpenAI GPT-5.6 Release (Expected This Week) — Polymarket prediction data showed 68–74.5% probability of public launch on or shortly after July 7. Any further delays or government-mandated safety reviews would signal mounting regulatory pressure.

3. AI Safety Certification Standards (Pending) — Following Anthropic's C+ rating achievement, expect competing frameworks from EU, US, and industry bodies. Companies racing to achieve certified safety status for competitive advantage and regulatory compliance.

Quick Reads

• Alibaba's Qwen AI Models Go Viral but Face Monetization Hurdles — Chinese company's open-source models have won developer mindshare worldwide, but free availability complicates revenue conversion.

• Samsung Profit Surge Reignites AI Bubble Deflation Debate — Investor skepticism about AI spending ROI grows as chip demand signals mixed sentiment.

• Noam Shazeer Joins OpenAI, Nobel Laureate John Jumper Departs for Anthropic — High-profile researcher moves signal talent competition and strategic priorities across labs.