AI News: June 5, 2026

1. Nvidia Acquires Kumo AI to Strengthen Predictive Analytics

Nvidia. Nvidia acquired Kumo AI, a startup building foundation models for business predictions such as customer churn and credit-default risk without additional task-specific training. Kumo counted DoorDash, Reddit, and Sainsbury’s among its customers and had raised $37 million from Sequoia since its 2022 founding. The deal fits Nvidia’s strategy of assembling a full-stack AI ecosystem through dozens of startup acquisitions; financial terms were not disclosed. Source

2. EU Agrees to AI Act Amendments With Timeline Relief and New Prohibitions

European Commission. EU negotiators provisionally agreed to amend the AI Act, delaying high-risk system compliance deadlines by 12 to 16 months to give companies more time to operationalize requirements. New prohibitions on AI systems that generate non-consensual intimate imagery and child sexual abuse material were added, taking effect December 2, 2026. The EU AI Office also gained exclusive supervisory competence over general-purpose AI systems and models integrated into major digital platforms. Source

3. GE HealthCare Gets FDA Clearance for AI Radiation-Therapy Contouring

GE HealthCare. GE HealthCare received FDA clearance for MIM Contour ProtegeAI+ 2.0, a system that automates the manual contouring step in radiation-therapy planning for CT and MR imaging. The clearance includes a Predetermined Change Control Plan, letting the company roll out future model improvements across additional anatomical regions without full re-review. Radiation therapy is used in roughly 60% of cancer cases, making workflow automation here a meaningful clinical opportunity. Source

4. Cloudflare CEO Says the Web’s Future Is “Pay to Crawl”

Cloudflare. Cloudflare CEO Matthew Prince said bot traffic now exceeds human traffic on the internet, earlier than projected, driven largely by AI agents scraping and interacting with web content. Prince argued the web’s economics will require sites to charge fees for automated crawling, reshaping the relationship between publishers and AI companies. The remarks reflect a wider industry debate over how to compensate publishers whose content trains and informs AI systems. Source

5. Tech Leaders Press Congress on DNA Screening as AI Reaches Virology Competency

Tech industry leaders. A group of technology executives sent Congress a letter urging legislation to require synthetic-DNA order screening, citing AI systems that now demonstrate PhD-level virology laboratory competency. The executives warned that AI can coach amateur researchers through complex biological procedures, lowering the barrier to misuse of genetic synthesis. The push marks one of the most direct biosecurity-focused lobbying efforts from the AI industry to date. Source

6. Bain Study Finds Companies Miss AI Savings Targets

Bain. A Bain survey of 951 companies found that nearly 40% achieved less than 10% cost savings from AI despite targeting reductions of 11 to 20%, with only 7% deploying fully autonomous agents. Bain attributed the gap largely to workforce hesitation, organizational friction, and a tendency to keep humans in decision loops rather than letting agents operate independently. The study underscores the divide between enterprise AI ambition and actual deployment outcomes. Source

7. Google Adds AI-Search Opt-Out Controls Under UK Regulator Pressure

Google. Google introduced opt-out controls in Search Console that let site operators exclude their content from AI-powered search features reaching billions of monthly users. The UK Competition and Markets Authority pushed for the change, citing the limited leverage of publishers who depend on Google traffic and cannot practically exit the ecosystem. Critics noted that while the opt-out exists, most publishers have little real choice given Google’s position as a primary traffic source. Source