Stanford AI Index reveals acceleration, transparency crisis

The 2026 report confirms AI is moving faster than infrastructure and regulation can handle, while the lack of transparency from top labs makes independent verification nearly impossible.

• –Models now match or beat human experts in PhD-level tasks, with Anthropic currently leading the pack
• –The US dominates in infrastructure and compute, but China is practically tied in research output and robotics
• –Environmental costs are staggering, with global AI data centers drawing 29.6 gigawatts of power
• –Junior software developer employment has dropped 20% since 2022 despite a 26% boost in productivity
• –Leading benchmarks are officially "broken," with models increasingly trained on the tests themselves