AI Startups Hire 15% Fewer Entry-Level Workers, Study Finds

Smaller Teams, Fewer Junior Roles

AI-native startups are building fundamentally different workforces than their traditional counterparts, according to new research from Harvard Business School and INSEAD. These companies employ 25% smaller teams overall, with entry-level positions and management roles each reduced by roughly 15% compared to non-AI startups.

The working paper analyzed Y Combinator startups and venture-backed US companies that received their first financing between 2020 and 2024. Researchers defined AI-native firms as those leveraging artificial intelligence through two distinct channels: internally to boost employee productivity in tasks like coding and design, and externally by embedding AI into products that replace work previously done by human teams.

The Expert Advantage

While junior positions shrink, demand for senior expertise is surging. AI-native startups employ 20% more senior workers than their peers, and the engineering share of the workforce runs about 13% higher. But the concentration goes beyond seniority alone.

These companies disproportionately recruit from a narrow talent pool: graduates of elite institutions, workers based in Silicon Valley, and male employees. The pattern suggests AI tools may be amplifying existing advantages rather than creating new pathways into technology careers.

Why It Matters

The findings challenge a popular narrative about AI democratizing the workplace. While tools like AI coding assistants theoretically lower barriers to technical work, the data shows AI-native companies are actually raising the bar for entry. This has immediate implications for workforce development and diversity in tech. If AI adoption accelerates learning primarily for those already using it, performance gaps could widen between workers with access to these tools and those without—creating a feedback loop that reinforces existing inequalities in both hiring and entrepreneurship.

Rethinking the Career Ladder

The research arrives as "vibecoding"—using AI to generate code from natural language descriptions—has made it easier for non-engineers to build prototypes. Yet this hasn't translated into more entry-level opportunities at AI-focused companies. Instead, the technology appears to be collapsing the traditional career ladder, where junior employees learn through routine tasks before advancing.

The authors express particular concern about demographic implications. Differential adoption rates of AI tools could translate into widening performance gaps, both for individual workers within companies and for the entrepreneurs who launch them. This dynamic risks entrenching advantages among groups already overrepresented in technology.

The findings were first reported by Business Insider, based on the Harvard Business School and INSEAD working paper titled "AI-Native Firms."