AI startups want engineers, not entry-level hires
NEW YORK, UNITED STATES — A working paper analyzing thousands of United States startups over four years confirms what hiring managers have been anecdotally reporting: AI-native companies are building leaner, more senior workforces, and entry-level workers are largely absent from their headcount plans, Business Insider reports.
The data: leaner teams, senior skew
Harvard Business School researcher Rembrand Koning and INSEAD researcher Hyunjin Kim examined Y Combinator cohorts from 2020 to 2024, plus a broader group of US venture-backed companies, to map how deep AI adoption changes workforce structure at each stage of a startup’s growth.
The researchers found that AI-native startups — defined as companies that use AI to enhance internal productivity and embed it in their products — run about 25% smaller than comparable non-AI firms, yet reach similar valuations, meaning they generate significantly more value per employee.
“These workers are especially likely to be graduates from elite institutions, concentrated in Silicon Valley, and male,” Koning and Kim wrote, describing the senior hires that define AI-native payrolls.
AI is concentrating startup opportunity at the top of the talent credential pyramid, not broadening it.
Entry-level roles are disappearing from the pipeline
IMF Managing Director Kristalina Georgieva has urged policymakers to ‘ensure AI’s economic benefits are broadly shared’ — a benchmark the data from Koning and Kim suggest the startup sector is actively moving away from.
AI-native firms carry roughly 13% more engineers than their non-AI counterparts, pushing technical complexity into the core of even the smallest teams.
They employ about 15% fewer entry-level workers and managers — hollowing out the rung that has traditionally served as the entry point to tech careers.
Senior staff make up around 20% more of total headcount at these companies than at non-AI peers of comparable size. The pattern is consistent: AI-native startups skip the junior cohort and build from the top down.
Business process outsourcing (BPO) and outsourcing firms are not passive observers of this shift — they are watching their core delivery model get compressed from two directions.
Entry-level roles in data processing, quality assurance, and tier-one support are the first categories AI automation absorbs, and they are also the workforce layer that has historically trained people for mid-level BPO delivery.
As AI-native clients scale and begin sourcing offshore services, they arrive expecting leaner teams with higher baseline skill profiles, compressing the entry-level market that outsourcing providers have long relied on to build capacity.
