Anthropic has introduced a new “dreaming” capability for its Claude Managed Agents, designed to address one of the most persistent challenges in AI systems: memory degradation over time.

In September 2025, Anthropic, a leading artificial intelligence company, uncovered the first large-scale cyber espionage campaign almost fully executed by an autonomous AI system.

Anthropic, one of the fastest-growing players in the enterprise AI arena, is forecasting an extraordinary leap in revenue—from an estimated $5 billion in 2025 to as much as $70 billion by 2028.