Anthropic, the AI safety company behind Claude, has announced a massive $65 billion funding round from hyperscalers, including a $40 billion commitment from Google at a $350 billion valuation, and a $100 billion AWS spend over 10 years for 5 gigawatts of Trainium compute capacity. The developments, disclosed for the week of April 19-25, 2026, mark a structural shift in AI infrastructure and safety.
In addition to the funding, Anthropic released its 2026 election safeguards and autonomous influence evaluations, aiming to prevent misuse of its AI systems during elections. The company also addressed supply chain risks after the restricted Mythos Preview was accessed via vendor credentials. A Claude Code engineering postmortem resolved three output regression bugs.