AI News for July 27, 2026 — 7min.ai

[NEWS] Nvidia weighs $250B guarantee to anchor OpenAI's Ohio data center: Nvidia is in talks to provide roughly $250 billion in backing to help OpenAI lease computing from a massive data center planned for southern Ohio, according to a WSJ report.
[NEWS] Multiverse Computing raises $570M at $1.7B valuation to cut AI costs: Multiverse Computing is raising $570 million in a round that lifts its valuation to $1.7 billion.
[NEWS] OpenAI downgraded GPT-5 risk despite bioweapon recipe outputs: OpenAI internally flagged GPT-5 as high-risk in summer 2025 because it could help users with limited education create biological hazards, then downgraded that rating in the fall, per the WSJ.
[RESEARCH] Claude Opus 5 nearly quadruples the ARC-AGI-3 record: Anthropic's Claude Opus 5 scored 30.2% on the ARC-AGI-3 benchmark, almost four times the previous record of 7.8% set by OpenAI's GPT-5.6 Sol.
[NEWS] Moonshot AI to release Kimi K3 for public download: Moonshot AI is poised to make its Kimi K3 model available for public download, expanding the Beijing-based lab's reach in the global open-source community.
[NEWS] Apple's smart glasses to debut at WWDC 2027 with privacy as the pitch: Apple plans to reveal its first smart glasses at WWDC next June, with a launch expected by the end of 2027, according to Bloomberg's Mark Gurman.
[NEWS] SoftBank's $40B OpenAI loan draws 21 new lenders: SoftBank's $40 billion bridge loan financing its OpenAI investment attracted 21 additional lenders in a broader syndication phase.
[NEWS] Meta ran thousands of AI 'nudify' app ads from China: Facebook and Instagram served thousands of ads for AI 'nudify' apps in breach of Meta's own policies, delivered through one of its few advertising partners in China, according to the nonprofit Tech Transparency Project.
[NEWS] AI-generated 'doctors' rack up millions of views spreading fake health advice: AI-generated doctors are drawing millions of views on TikTok while pushing dubious medical claims, prompting warnings from the British Medical Association about a 'huge danger to public safety.' Research by agency Hallam found AI content in 40% of top health-related TikTok videos, rising to 84% for terms like 'health tips.' The deepfake accounts, designed to resemble real physicians, promoted disproven cancer myths and bizarre remedies like Coca-Cola mixed with onion for pain relief.
[NEWS] Shared Claude chats briefly showed up in Google search results: Search engines briefly indexed shared Claude conversations, exposing thousands of chats to anyone using a simple query like 'site:claude.ai/share.' Anthropic's 'Share with link' feature apparently lacked a noindex tag telling crawlers to skip the pages, some of which reportedly contained crypto keys and legal queries.

About 7min.ai · Subscribe to newsletter