NVIDIA and data center company IREN announced a strategic partnership to build up to 5 gigawatts of AI infrastructure, anchored by a flagship 2GW campus in Sweetwater, Texas — with NVIDIA securing warrants to invest up to $2.1 billion in IREN.
A landmark UMass trial called SMART-AI showed that a real-time AI platform outperformed traditional biopsy at diagnosing bile duct cancer — a potential replacement for the current gold standard.
MIT and Mass General’s Mirai AI tool identified 4,000 high-risk cancer patients in routine mammograms that doctors had cleared as normal — but despite being tested across 100+ hospitals worldwide, insurance companies still won’t cover it.
Google’s spring 2026 Home update bakes Gemini 3.1 into its voice assistant, letting users issue a single multi-step command instead of repeating requests, while expanding automation to more smart home devices.
Microsoft launched Agent 365 as a generally available enterprise product on May 1, offering IT teams centralized visibility and control over every AI agent running across their organization — including shadow AI tools employees install on their own devices.
Multiple state AI laws took effect in 2026, with Connecticut’s AI bill, Maryland’s pricing transparency law, and Colorado’s AI Act creating compliance headaches as Congress twice rejected a federal moratorium.
A study published in Science found that an OpenAI reasoning model outperformed experienced physicians at diagnosing patients and managing their care using only electronic health records from a Boston emergency department.
OpenAI’s CEO has been targeted in two separate violent attacks — a Molotov cocktail thrown at his San Francisco home and shots fired at his property days later, as the anti-AI backlash turns dangerous.
The TAKE IT DOWN Act’s one-year grace period expires May 19, 2026, forcing all major social platforms to establish takedown systems for AI-generated intimate images and deepfakes within 48 hours of notification.
Elon Musk’s xAI quietly dropped Grok 4.3, featuring a 2-million-token context window, native video input, and a powerful voice cloning suite — all at aggressively lower prices than competing models.