Stanford doesn’t speculate. Its AI Index Report is one of the few annual publications that tracks AI’s economic footprint across employment, investment, and skills data, and the 2026 edition contains a finding that the markets audience has been waiting for someone authoritative to confirm.
According to the 2026 Stanford AI Index Report, employment for software developers aged 22 to 25 has fallen by approximately 20% since 2024. The report also found that AI skills demand in the information sector grew to 13.2% in 2025, up from 7.8%, according to Stanford HAI. These two data points aren’t separate stories. They’re the same story from opposite ends of the labor market: routine technical work is being automated while demand for workers who can build and manage AI systems continues to rise.
The workforce reduction finding adds a second signal. According to the 2026 AI Index Report, as reported by CIO Dive, roughly one-third of organizations expect AI to cause workforce reductions in the coming year. Note carefully: this is a different finding from a separate figure in the same report, the HAI report also found that 33% of Americans expect AI to improve their own jobs, compared to a 40% global average. Both figures come from the Stanford HAI 2026 AI Index. They’re measuring different things. The organizational expectation of reductions is the labor market signal. The worker sentiment figure is a confidence measure. Don’t conflate them.
The developer employment finding is particularly significant because of the age cohort it targets. Workers aged 22 to 25 are entry-level. They’re doing the work that AI code generation tools are most capable of handling today: writing boilerplate, drafting tests, implementing well-specified features. The 20% decline isn’t a surprise to anyone who has watched GitHub Copilot and its successors become standard tools in engineering organizations. What Stanford has done is put a verified number on it.
This is sector-level econometric data, not a named-company layoff event. Stanford is inferring displacement from aggregate workforce statistics, not from a press release. That makes the attribution classification `ai-adjacent` rather than `ai-direct`, the causal link to AI automation is analytically strong, but it isn’t an explicit company statement. That distinction matters for how you interpret the finding. It also means the 20% figure reflects the aggregate outcome of thousands of individual hiring and non-hiring decisions made across the industry, not one company’s restructuring announcement.
The skills demand data is the forward signal. A jump from 7.8% to 13.2% AI skills demand in the information sector in a single year is steep. Organizations cutting entry-level developer headcount while simultaneously increasing demand for AI-skilled workers aren’t downsizing their technical workforce overall. They’re replacing one kind of technical worker with another. That’s a different problem than a cyclical tech layoff, and it requires a different response from employers, educators, and the workers themselves.
The Stanford HAI 2026 AI Index is a primary source event. Its findings carry institutional weight that secondary reporting and analyst estimates don’t. For anyone tracking AI’s workforce impact, this report is the benchmark to measure other data against.