AI News Roundup — April 12, 2026
Utah becomes the first US state to authorize AI to renew prescriptions, NVIDIA expands open physical and healthcare AI models, and Yann LeCun's new AMI Labs closes a record $1.03 billion European seed round. Oracle also plans a $50 billion AI infrastructure raise as early benchmarks show Meta's Muse Spark trailing its rivals.
Utah Becomes First State to Let AI Renew Prescriptions
Utah signed legislation granting AI systems the authority to renew drug prescriptions — the first US state to move AI from diagnostic assistance into actual treatment decisions. The law marks a landmark expansion of AI into direct patient care and is expected to inspire similar legislation across other states.
NVIDIA Releases New Physical AI and Healthcare Models
NVIDIA expanded its open model families to power the next wave of agentic, physical, and healthcare AI. New Isaac GR00T models enable robots to understand natural language and execute complex multistep tasks, while new Cosmos world models improve generalization efficiency. CMR Surgical and Johnson & Johnson MedTech are already adopting the models for robotic surgical training and validation.
Anthropic Claude Mythos Preview Scores Near-Perfect on Key Benchmarks
Anthropic revealed that its unreleased Claude Mythos Preview model scores 93.9% on SWE-bench Verified and 94.6% on GPQA Diamond — among the highest ever recorded. The model has also identified thousands of zero-day vulnerabilities across every major operating system and browser, and is being made available to select cybersecurity partners through Project Glasswing.
Yann LeCun's AMI Labs Raises $1.03 Billion Seed Round — Largest in European History
Advanced Machine Intelligence Labs, founded by Turing Award winner and former Meta chief AI scientist Yann LeCun, closed a $1.03 billion seed round at a $3.5 billion valuation — the largest seed funding in European history. The lab is pursuing a distinctly different architectural philosophy from US-centric frontier AI labs and signals a major new competitor in the global AI race.
Oracle Plans to Raise $50 Billion for AI Infrastructure Expansion
Oracle announced plans to raise up to $50 billion to fund a massive build-out of global data centers designed specifically for generative AI and autonomous agents. The move positions Oracle as a major third-party AI cloud contender alongside Microsoft Azure, Google Cloud, and AWS in the intensifying infrastructure arms race.
OpenAI Publishes Child Safety Blueprint for the AI Era
OpenAI released a Child Safety Blueprint addressing child exploitation threats reshaped by AI. The proposal rests on three pillars: modernizing the legal framework, improving reporting and coordination between platforms and law enforcement, and strengthening safety-by-design requirements for AI systems. The release follows the company's earlier OpenAI Safety Fellowship announcement.
Meta Muse Spark Lags Behind Rivals in First Benchmark Tests
Early independent benchmark results show Meta's new Muse Spark model trailing Anthropic's Opus, OpenAI's GPT, and Google's Gemini on core reasoning, coding, and advanced AI tasks. Despite Meta's $115–135 billion capex commitment for 2026, analysts say the company faces a steep catch-up curve in model quality against entrenched leaders.