technology
U.S. AI Patents Granted (Quarterly)
Latest
4,001patents
Methodology
## What this measures
The number of U.S. utility patents granted in a quarter that carry any CPC symbol in G06N, the class for computing based on specific computational models: machine learning, neural networks and knowledge-based systems. Every applicant counts, U.S. or foreign.
## Why it matters
Grants trace the AI research of two to four years earlier and set who owns the methods that products will be built on.
## How to read it
The series starts in 2010 Q1 at 106. The latest reading, for 2026 Q3, is 4,000. The high is 4,577 (2022 Q4) and the low 106 (2010 Q1).
## What it does not say
G06N is the core AI class; patents that apply AI inside another field (image recognition in G06V, speech in G10L) count only if they also carry a G06N symbol. A patent counts once however many symbols it carries. CPC is the current classification, so older patents count where they sit today. Counts come from the Open Data Portal, which matches USPTO's published annual totals to within 0.5% from 2010.
## Source
Source: United States Patent and Trademark Office, www.uspto.gov (Open Data Portal, patent file wrapper search counts). Public domain.
Read this signal inside the lattice.
Palanor weighs every signal against the world your organization is watching. Methodology is public; integration is the platform.
See it in Palanor →