OpenAI and Anthropic conducted joint safety testing on their AI models, marking rare collaboration amid intense industry competition. The research revealed contrasting approaches to hallucination management, with Anthropic models refusing uncertain questions while OpenAI models attempted more answers. Both companies expressed interest in continued safety partnerships despite recent API access disputes.

Nvidia reported $46.7 billion in Q2 revenue, up 56% year-over-year, with AI-driven data center sales reaching $41.1 billion. Net income rose 59% to $26.4 billion. Blackwell chips accounted for $27 billion in sales. Despite strong overall performance, the company faces challenges in China, reporting no H20 chip sales to Chinese customers amid regulatory uncertainty.

Google has reduced managers overseeing small teams by 35% in the past year, with many transitioning to individual contributor roles. The company is pursuing efficiency through reduced bureaucracy, offering voluntary buyouts across ten product areas, with 3-5% of eligible employees accepting offers. CEO Sundar Pichai emphasized giving employees agency in workforce changes.

In 2025, 33 US AI companies have raised funding rounds of $100M or more, continuing momentum from 2024. Notable investments include OpenAI's $40B round valuing it at $300B, Anthropic's $3.5B round, and multiple startups raising more than one mega-round. Healthcare AI and infrastructure companies attracted significant capital, with many achieving unicorn status.

Maisa AI raised $25M for its platform tackling the 95% failure rate of generative AI pilots in enterprises. The startup's approach uses accountable AI agents and a deterministic system to limit hallucinations, targeting critical tasks in banking, manufacturing, and energy sectors. Their model-agnostic platform, Maisa Studio, deploys digital workers trained with natural language.