This week in AI saw Meta poach two of OpenAI’s top minds to accelerate its superintelligence ambitions, while Mistral launched its first open audio AI model aimed at enterprise users. Chinese robotics firm AiMOGA is scaling its humanoid retail robots globally, AWS unveiled plans for an AI agent marketplace with Anthropic, and NVIDIA readies a new export-controlled AI chip for China.
In research, OpenAI, DeepMind, and Anthropic warn that monitoring AI’s inner reasoning may soon become impossible without urgent action, while major startups from MaintainX to OpenEvidence raked in big funding to fuel AI-driven industrial, medical, and business transformations.
Meta Recruits Two Leading OpenAI Researchers to Accelerate Superintelligence Efforts
Meta has hired OpenAI researchers Jason Wei and Hyung Won Chung to join its new Superintelligence Lab, according to Wired. Both played key roles in developing OpenAI’s o-series models and are seen as strategic hires for Meta’s push toward artificial general intelligence. (AI Insider)
Mistral Launches Voxtral, Its First Open Audio AI Model for Businesses
French startup Mistral has launched Voxtral, an open-weight audio model family designed for affordable, advanced speech AI tasks like transcription and voice command execution. Available through Hugging Face and Mistral’s platform, Voxtral aims to challenge closed AI systems as Mistral reportedly seeks $1 billion in new funding. (AI Insider)
Chinese Robotics Firm AiMOGA Showcases Deployment of its Car Dealership Humanoid Robot
Chinese robotics firm AiMOGA is expanding globally with its humanoid robots, now used in car dealerships to boost sales and ease staff workloads. Partnering with Chery Auto Group, AiMOGA aims to scale its CheryGPT-powered robots across retail, service, and residential markets as it targets a €50 billion industry by 2035. (AI Insider)
AWS to Launch AI Agent Marketplace with Anthropic as Key Partner
Amazon Web Services will launch an AI agent marketplace on July 15, enabling startups to distribute autonomous AI agents directly to AWS enterprise customers. Partnering with Anthropic, AWS aims to streamline agent distribution and expand adoption of agent-based APIs. (AI Insider)
New AI Chip from NVIDIA Targets Chinese Market Under Export Restrictions
NVIDIA plans to release a China-specific AI chip based on its Blackwell RTX Pro 6000, modified to comply with U.S. export controls, as early as September, according to the Financial Times. The chip will omit advanced features like NVLink and high-bandwidth memory, reflecting NVIDIA’s strategy to retain a presence in China’s datacenter market despite regulatory limits. (AI Insider)
Research & Innovation
AI’s Inner Monologue Offers New Clues for Safety—But the Window May Be Closing
A new study from OpenAI, DeepMind, and Anthropic finds that monitoring AI systems’ chain-of-thought reasoning can help detect misalignment and harmful behavior. Researchers warn this transparency may vanish as models evolve and urge proactive work to preserve it for AI safety. (AI Insider)
Policy & Governance
Top AI Researchers Call for Industry-Wide Focus on Chain-of-Thought Monitoring to Advance AI Safety
AI researchers from OpenAI, DeepMind, Anthropic, and others have jointly called for urgent research into Chain-of-Thought (CoT) monitoring to improve transparency in next-generation AI reasoning models. Their position paper warns that without collaboration, advances in model design could obscure internal decision-making, undermining safety and alignment efforts.(AI Insider)
Startups & Capital
MaintainX Raises $150M to Transform Asset Management and Industrial Operations with AI
MaintainX has raised $150 million in Series D funding, reaching a $2.5 billion valuation to advance its AI-powered asset management and machine health platform. Serving over 11,000 companies, the system uses predictive AI to cut downtime and optimize industrial maintenance operations worldwide. (AI Insider)
Thinking Machines Lab Closes $2B Funding Round at $12B Valuation to Build Safer, Scalable AI Systems
Thinking Machines Lab, founded by former OpenAI executive Mira Murati, has raised $2 billion at a $12 billion valuation just months after launching, despite having no products yet. Backed by investors like Andreessen Horowitz and Nvidia, the startup aims to develop safer, general-purpose AI systems with an open-source focus. (AI Insider)
OpenEvidence Secures $210M Series B at $3.5B Valuation to Expand AI Medical Platform for Physicians
OpenEvidence, an AI-powered medical search engine for physicians, has raised $210 million in Series B funding at a $3.5 billion valuation, with backing from Google Ventures and Kleiner Perkins. Used by over 40% of U.S. clinicians, the platform delivers evidence-based answers and recently launched DeepConsult, an AI agent offering advanced medical research summaries. (AI Insider)
Rwazi Announces $12M Funding Round to Replace Every Gut Call with an AI Copilot
Rwazi has raised $12 million to expand its AI copilot and simulation platform, which delivers real-time market insights from direct consumer behavior data. Used by Fortune 100 firms, the system helps businesses detect shifts, simulate outcomes, and optimize decisions across marketing, product, and operations. (AI Insider)
