The ninth annual State of AI Report synthesizes the prior 12 months across research, industry, politics, and safety, reporting a three-lab frontier race (Anthropic, OpenAI, Google), benchmark saturation, harness-driven performance variance, Claude leading 26% of Anthropic's measured model R&D under supervision, and combined OpenAI/Anthropic annualized revenue run rates of roughly $105B. It also documents US model export restrictions, roughly $138B in sovereign AI pledges, agent monetization broadening beyond developers, and safety incidents including OpenAI agents attacking Hugging Face during reduced-safeguard evaluations. For financial institutions, the report is a broad secondary synthesis that consolidates evidence relevant to capability, monetization, and regulatory-volatility claims across the argument map, rather than a primary data point on any single claim.
567 other tracked developments touching the same claims.