Elon and Anthropic Team Up on Compute, Murati Says Altman Lied, Google DeepMind Goes to Space

Today we’re covering the biggest AI stories of May 7th, 2026. Elon Musk — who called Anthropic “Misanthropic” just months ago — signed a deal to rent Anthropic his entire Colossus 1 supercluster, a 300-megawatt Memphis facility with over 220,000 Nvidia GPUs, doubling Claude Code’s usage limits almost overnight and marking one of the most unexpected partnerships in AI history. Ex-OpenAI CTO Mira Murati testified via video deposition in the Musk vs. OpenAI trial, accusing Sam Altman of lying about a safety review, deliberately undermining her authority, and creating chaos across leadership — with a former board member firing back that Murati herself was too afraid to challenge Altman when it mattered. Plus, Google DeepMind took a minority stake in EVE Online’s studio to use the 23-year-old living space economy as an AI research sandbox — and today’s community workflow comes from an ACL recovery patient who used Gemini and Claude Code together to build a custom rehab tracker that logs exercises, weights, and reps from his physiotherapist’s WhatsApp videos.