Executive summary
Observed facts: the reviewed sample of 1,515 repositories shows 12.74% identified as having AI built in (12.74%), 4.42% identified as AI-using (4.42%), and 15.91% identified as either (15.91%). The margin of error for the sample is 1.68 percentage points. Within the sample, 193 repos were labeled built, 67 using, 241 either, and 19 both.
Interpretation: the prevalence of projects with AI-built components in this sample is non-trivial relative to the rest of tracked activity, while direct AI-using signals remain lower. The margin of error (±1.68 points) indicates that reported percentages should be treated as estimates with moderate precision; the sample size is 1,515 repositories.
- Sample size: 1,515 repositories (sample).
- AI-built: 12.74% (of sampled repos).
- AI-using: 4.42% (of sampled repos).
- Either built-or-using: 15.91% (of sampled repos).
- Margin of error: ±1.68 percentage points.
Key movements
Observed facts: among tool signals tracked this month, the largest counts are 6,049 repos with the top signal, 4,582 with the next, and 467 for a third signal. The tracked corpus this site reports includes 36,012 AI-related repositories and 11,616 repos with explicit AI-tool signals. Continuity of development over a three-month window is 0.1771 (17.71%). There are 5,025 candidate solo-developer repositories and 5,501 web-launch candidate repositories in the tracked corpus.
Interpretation: tool-signal concentrations remain concentrated in a few signals, suggesting a small number of toolchains or frameworks drive much of the detectable tooling footprint. The continuity rate (≈17.7%) implies a minority of repositories created three months ago sustain active pushes, which affects expectations for longer-term project maturation.
- Top tool signals (repo counts): 6,049; 4,582; 467; 312; 39; 30; 9 repos.
- Tracked-corpus counts (this site): 36,012 AI-related repos; 11,616 with tool signals.
- Continuity (3-month): 0.1771 (17.71%).
- Solo-dev candidates: 5,025 repos; web-launch candidates: 5,501 repos.
Twelve-month trend
Observed facts: between 2026-04 and 2026-09 the share identified as having AI built moved through these values: 7.51% (Apr), 8.02% (May), 8.86% (Jun), 10.95% (Jul), 13.04% (Aug), and 12.74% (Sep). The share identified as AI-using in the same months was: 4.58% (Apr), 4.60% (May), 4.40% (Jun), 5.12% (Jul), 4.35% (Aug), and 4.42% (Sep). Sample sizes for those months ranged from 1,680 to 1,811, with 1,515 in the current sample.
Interpretation: built-share rose notably from April through August and is near the recent peak in September, while the using-share has remained relatively flat in the 4.3–5.1% range. The divergence between the growing built-share and the steadier using-share suggests increased incorporation of AI components in projects rather than an equivalent rise in projects that merely call out third-party AI usage.
- Built-share by month: Apr 7.51%, May 8.02%, Jun 8.86%, Jul 10.95%, Aug 13.04%, Sep 12.74%.
- Using-share by month: Apr 4.58%, May 4.60%, Jun 4.40%, Jul 5.12%, Aug 4.35%, Sep 4.42%.
- Monthly sample sizes: 1,811; 1,784; 1,773; 1,698; 1,680; 1,515.
Outlook
Observed facts: notable high-profile repositories this month include projects with star counts up to 98,290 and AI scores ranging in the dataset (examples include scores of 100, 70, 55, 50). The largest tool-signal counts are concentrated in a couple of signals totaling thousands of repositories. The tracked corpus contains tens of thousands of AI-related repositories and several thousand candidate solo and web-launch projects.
Interpretation: near-term monitoring should prioritize the dominant tool signals and projects with sustained pushes because they appear to drive visible changes in the built-share. Given the modest continuity rate (≈17.7%), many new AI-related repos may not convert to long-lived projects; focusing on signal concentration and continuity offers actionable insight into which initiatives are likely to persist.
- High-impact indicators to monitor: dominant tool-signal counts; continuity rate (0.1771 / 17.71%); counts of solo-dev and web-launch candidates (5,025 and 5,501 repos).
- Representative repository signals to watch: top repositories by stars and AI score (examples with up to 98,290 stars and AI score up to 100).