Released in April 2026, Stanford HAI's annual AI Index has become the most-cited single reference point for AI policy and investment discussions worldwide — and this year's numbers explain why it gets that much attention.
The investment number
Global corporate AI investment reached $581.7 billion in the year covered, a 130% year-on-year increase. That's not steady growth — it's a more-than-doubling in a single year, the kind of trajectory that invites the "bubble" question as readily as it invites optimism.
Capability moved just as fast
On standard coding benchmarks, AI system performance rose from roughly 60% to near 100% in the space of a year — a capability jump that's arguably more consequential than the investment figure, since it's the thing the investment is actually betting on.
The number regulators are watching
Reported AI incidents rose 55% to 362 for the year — a reminder that capability and investment growth have been accompanied by a real, measurable rise in documented failures, misuse, and harms, not just headline progress.
The China angle
The report also tracks a closing performance gap between US and Chinese AI systems on key benchmarks — a trend with obvious implications for the export-control and trade-policy fights playing out in parallel (see: the H200 chip story).
Bubble or not — what the numbers actually support
A 130% year-on-year investment jump is the kind of figure that shows up in most historical accounts of a bubble in formation — but it's also consistent with a genuine platform shift where capability truly is compounding as fast as the capital suggests, given the coding-benchmark jump. The more useful number for business readers may be the third one: a 55% rise in incidents means the operational and legal risk of AI deployment is scaling right alongside the investment, which is the more immediate planning question for most companies right now, regardless of how the macro bubble debate eventually resolves.