Research from Harvard Business School and INSEAD has revealed intriguing insights about the evolving landscape of startups that leverage artificial intelligence. The findings, encapsulated in their working paper titled "AI-Native Firms," highlight a noteworthy shift in team dynamics among AI-native startups compared to traditional ones.
Analyzing Y Combinator startups from the years 2020 to 2024, alongside a wider array of U.S. venture-backed startups that secured initial funding during this timeframe, the researchers defined a new breed of companies: "AI-native startups." These firms exhibit two significant productivity transformations. The first involves internal optimization, where AI tools enhance employee efficiency in various tasks such as coding, selling, designing, or coordinating. The second entails the integration of AI into products, enabling customers to carry out tasks that would typically require human intervention.
What stands out is that AI-native startups tend to operate with approximately 25% fewer employees than their non-AI counterparts. Interestingly, they maintain a workforce with about 13% more engineers while seeing a 15% reduction in both entry-level employees and managerial positions.
These findings challenge the widespread belief that the rise of AI is democratizing job opportunities at the entry level. Instead, it appears that these advancements allow entry-level workers to take on greater responsibilities more quickly, while simultaneously automating mundane tasks. The emergence of tools like vibecoding has blurred the lines around technical qualifications needed to prototype ideas, further complicating the landscape.
However, the study also indicates a rising demand for highly skilled professionals within AI-native startups. These firms have a 20% higher ratio of senior employees, predominantly attracting individuals from elite educational backgrounds, particularly in Silicon Valley, and skewed towards male candidates.
This trend raises important questions about equity in access to career opportunities. The authors express concern that, while AI may enhance learning for its users, varying adoption rates might lead to increasing disparities in performance—both among individual employees and the entrepreneurs leading these firms. The research suggests that rather than broadening opportunity, AI-native startups might be reinforcing privilege among an already select group of highly qualified workers.
