Stanford HAI's 2026 AI Index consolidates cross-domain data showing frontier AI capability continuing to accelerate: SWE-bench Verified coding performance rose from 60% to near 100% in one year, AI agents jumped from 12% to ~66% task success on OSWorld, and several models now meet or exceed human baselines on PhD-level science and competition mathematics. The US-China frontier model gap has effectively closed, with DeepSeek-R1 briefly matching the top US model in February 2025 and Anthropic's leading model ahead by just 2.7% as of March 2026. US private AI investment reached $285.9B in 2025 (23x China's reported private figure), but US ability to attract global AI talent has dropped 89% since 2017. For financial institutions, the report provides the most comprehensive public data point on the simultaneous acceleration of capability, narrowing of competitive moats between US and Chinese labs, and the persistent gap between enterprise adoption (88%) and demonstrated responsible-AI governance, all of which bear directly on competitive positioning and risk assessments across AI-exposed portfolios. Also reported by: https://iapp.org/news/a/a-view-from-dc-can-ai-governance-catch-up-to-innovation · https://www.brookings.edu/events/data-gap-ai-policy-stanford-ai-index/
589 other tracked developments touching the same claims.