Groq's $650M comeback, AI agents go infinite loop, and DeepMind bets on Hollywood
Groq re-emerges with $650M after Nvidia's not-acqui-hire, agentic AI enters the 'loop' era, and OpenAI hunts open source bugs.
Subscribe free →Groq Confirms $650M Raise and a Fresh Start After Nvidia's $20B Not-Acqui-Hire
After Nvidia's headline-grabbing $20B not-acqui-hire deal — where Nvidia licensed Groq's talent and tech without actually buying the company — Groq has confirmed it raised $650M and is rebuilding its executive bench. The AI chipmaker is doubling down on its neocloud business, positioning itself as an inference speed play against Nvidia's own hardware dominance. Getting ghosted by Nvidia would kill most startups; Groq is treating it as a pivot opportunity.
Why it matters: A $650M raise signals serious investor conviction that there's still room for a non-Nvidia inference infrastructure play, even after the weird quasi-acquisition saga.
AI Agents Are Going Infinite — The "Loop" Is the Next Frontier
The next evolution of agentic AI isn't just agents completing tasks — it's swarms of agents running continuously in the background with no endpoint. This "loop" architecture authorizes AI to keep working, iterating, and acting without waiting for human prompts between steps. We've moved from chatbots to agents to, apparently, AI that just never stops.
Why it matters: Persistent background agent loops raise the stakes on AI oversight — once you authorize the loop, the hard question becomes how and when you turn it off.
Google DeepMind Drops $75M on AI Filmmaking Tools With A24
Google DeepMind is partnering with prestige film studio A24 in a $75M deal to develop AI tools specifically for Hollywood filmmaking. This isn't a vague "AI in entertainment" press release — it's DeepMind backing one of the most creatively credible studios in the business to build real production tools. If anyone can make AI filmmaking tools that artists actually want to use, A24's involvement is a meaningful signal.
Why it matters: DeepMind putting real money behind a respected creative studio is the most credible attempt yet to land AI in Hollywood without triggering an immediate backlash from the creative community.
OpenAI Launches Initiative to Hunt and Patch Open Source Security Bugs
OpenAI announced a new program using its AI to scan open source projects for vulnerabilities and help patch them. The initiative is framed as a contribution to the open source ecosystem's security posture — using the same class of models that OpenAI builds commercially to do defensive security work for free. Details on scope and which projects are covered remain thin.
Why it matters: It's a smart piece of goodwill engineering — OpenAI using its own tools to secure the open source infrastructure that the entire AI ecosystem depends on.
Anthropic's Claude Might Ask to See Your Passport
Anthropic quietly updated Claude's privacy policy to allow the chatbot to request identity verification — including a passport or driver's license — "in certain circumstances." The company hasn't fully spelled out what triggers this, but it signals a move toward age-gating and identity-linked AI interactions. This is the first major consumer AI assistant to float document-level ID checks.
Why it matters: If identity verification becomes normalized in AI assistants, it fundamentally changes the anonymous, frictionless nature of chatbot access and sets a precedent every other lab will have to respond to.
Nvidia's "Water-Saving" Cooling Announcement Misses AI's Actual Water Problem
Nvidia unveiled a new data center cooling system that reduces water consumption inside the facility — but a closer look reveals it does nothing about AI's largest water draw: the fossil fuel power plants generating the electricity in the first place. The announcement is getting framed as a sustainability win, but the math doesn't hold up once you follow the water upstream.
Why it matters: The gap between Nvidia's PR framing and the actual environmental footprint of AI infrastructure is a preview of the greenwashing battles coming as AI energy demand keeps scaling.
Quick Hits
- →Microsoft locked in a 20-year natural gas power deal with Chevron to fuel new data centers — clean energy goals, meet hyperscaler appetite. TechCrunch
- →TechCrunch's running tracker of 2026 tech layoffs where companies explicitly blamed AI keeps growing — worth bookmarking. TechCrunch
- →Midjourney's pivot to medical ultrasound body scanners is raising eyebrows — The Verge found the clinical evidence "lacking." The Verge
- →Fika Jobs raised $4M for a "LinkedIn meets TikTok" hiring platform where AI agents conduct the first-round interview. TechCrunch
AI TLDR