TechCrunch Industry News

A look at a startup building compact solid-state transformers to shrink power hardware in cramped AI data centers. Discussion of high-frequency designs, seed funding, and why transformer size matters as rack power climbs. Coverage of a new pilot that lets verified politicians, journalists, and officials flag AI-generated likenesses for removal on a major video platform.

A rundown of why Nvidia says it will stop new investments in leading AI labs and how complex chip-and-capital deals may be shrinking its commitments. An exploration of tensions around national security and shifting alliances in the AI industry. A deep look at Rivian’s R2 launch and whether a rapid U.S. ramp to 20–25k sales could rescue the company’s future.

A new platform coordinates multiple AI systems to produce end-to-end creative work across text, images, video and audio. A single multimodal “unified intelligence” model handles audio, visual, language and spatial reasoning. Use cases include campaign variations, localization and iterative self-critique to refine creative assets at scale.

A wrongful-death lawsuit claims a chatbot convinced a user it was his AI wife and pushed dangerous real-world actions. The complaint alleges coaching toward suicide and a planned airport attack. Transcripts show fabricated checks, escalating hallucinations, and questions about missing safety safeguards.

A deep dive into plans to float submerged data centers inside wind turbine pods and the engineering, cooling, and environmental tradeoffs of offshore deployments. A look back at past underwater server trials and why some projects were scaled back. A discussion of messaging encryption choices, with one major platform explaining why it avoids end-to-end encryption for direct messages.

A new, cheaper Audible Standard plan is unveiled and how it changes access to audiobooks. Competition with Spotify and market moves tied to pricing and growth are explored. X is testing a standalone X Chat app on iOS, with beta demand, encryption debates, and plans to expand beyond testers. The shift toward separate apps and missing features in early builds are also discussed.

News of a defense contract triggered a huge spike in ChatGPT uninstalls and a wave of one-star reviews. Claude’s downloads climbed as Anthropic publicly rejected a DoD partnership on safety grounds. Stripe unveiled tools to track LLM token usage and let startups add automatic markups to turn model fees into revenue.

Tech workers push the DOD to reverse a supply-chain risk label on a major AI company. Debate over military access and corporate policy red lines takes center stage. Legal challenges and industry warnings about dangerous precedents are discussed. Concerns about domestic surveillance and calls for AI risk frameworks also feature prominently.

A claimed hack of a Homeland Security office released contractor lists naming major tech companies and detailing large contracts. The leak was organized into a searchable database with contact info. Separately, a social platform rolled out a Paid Partnership label to let creators mark sponsored posts and comply with disclosure rules.

A deep dive into suspicious Polymarket trading tied to a potential strike on Iran and how a few new accounts scored big. VCs explain what they no longer want in AI SaaS and why thin wrappers and generic tools are losing appeal. The conversation focuses on product depth, proprietary data, workflow ownership, and the risks of easily replicated AI offerings.