Stanford released the AI Index 2026 — over 400 pages of data, analysis, and visualizations on the state of the AI industry. A few numbers stand out.
The Trust Gap
Only 33 percent of Americans expect AI to improve their jobs. Globally, it’s 40 percent. In China, 83 percent of respondents say AI’s benefits outweigh its drawbacks — in the US, that number is just 39 percent.
Trust in government AI regulation sits at 31 percent in the US — the lowest among all surveyed nations. There’s a chasm between Silicon Valley’s excitement and the public’s skepticism.
China Is Catching Up — Fast
The gap between the top US model (Anthropic’s Mythos) and China’s best performer is down to just 2.7 percent. In 2023, the US lead ranged from 17 to 31 percentage points on major benchmarks. The technical gap is shrinking rapidly.
But the US still leads significantly in investment ($285.9 billion vs. China’s $12.4 billion), GPU capacity (roughly 75 percent globally), and talent (42 percent of global AI researchers, 60 percent of the elite tier). China dominates in patents (69.7 percent of all AI patent grants) and industrial robots (51.1 percent of global installations).
The Transparency Crisis
A concerning trend: 80 out of 95 notable models in 2025 were released without training code. Leading developers stopped disclosing training data sizes and parameter counts for their top systems. The industry is becoming increasingly opaque.
Beyond US vs. China
The Index also shows the AI race is no longer a two-horse competition. South Korea leads in patent density per capita, 44 nations now maintain state-backed supercomputing clusters, and Europe prioritizes regulatory architecture over model production.
My Take
The trust gap is the biggest problem in this report. When 67 percent of the population doesn’t believe AI will benefit their work, that will create political headwinds — no matter how good the models get. And the shrinking technical gap with China shows that US dominance is far less secure than many in Silicon Valley assume.
Sources: