--- title: AI Startups Are Staying Small by Hiring Fewer Junior Staff description: AI-native startups are 25% smaller than peers, with fewer entry-level staff and managers, a Harvard and INSEAD study finds, as solo founding hits a record. author: Darie Nani (Editor-in-Chief) date: 2026-09-24T23:51:42.727Z updated: 2026-09-24T23:51:42.737Z canonical: https://www.sovereignmagazine.com/article/ai-startups-hiring-fewer-junior-staff image: https://cdn.nanimediahouse.com/pexels-two-young-adults-focused-on-work-in-a-modern-office-using-la-7212931.jpg categories: Startups content_type: Analysis region: United States publication: Sovereign Magazine schema_type: Article --- Floqer, a Halifax startup that sells AI software for organizing customer data, reached seven figures in annual recurring revenue in less than two years with a team of nine. Its co-founder and chief executive, Shivam Mahajan, has said he wants to build a billion-dollar company with fewer than 20 employees. Other early-stage founders are working to the same plan. Startups built around AI are 25% smaller than comparable companies, and the shares of entry-level workers and managers in their workforces are each about 15% lower, according to a [working paper](https://www.hbs.edu/ris/download.aspx?name=26-090.pdf) published in June by Hyunjin Kim of INSEAD and Rembrand Koning of Harvard Business School. The Wall Street Journal reported on September 21 that some founders now ask whether AI can do a job before they hire anyone for it, and that some have laid off staff or dropped contractors and agencies they no longer think they need. ## Small Teams With Large Revenues Lindy, which sells an AI assistant that works inside Slack, has about two dozen employees and has raised $50 million. It laid off a handful of marketers in January during a change of product direction and did not refill the roles. Two marketers working with AI agents now look after more than half a million users, and Lindy no longer pays the five creative agencies it once used for content and video, the Journal reported. Butternut AI, which makes an AI website builder for small businesses, has about 10,000 paying customers and four employees, down from nine at its peak, co-founder Simar Singh told the Journal. It laid off five engineers at the end of last year after deciding AI agents could do their work. Gamma, which makes AI software for presentations, said in November that it had passed $100 million in annual recurring revenue with about 50 employees and had been profitable for more than two years. It raised $68 million at a $2.1 billion valuation at the same time. ## The Harvard and INSEAD Study of AI-Native Startups Kim and Koning studied about 2,900 Y Combinator companies from batches between 2020 and 2024, along with tens of thousands of US venture-backed startups tracked by PitchBook, and matched them to workforce records from Revelio Labs. They compared each AI-native startup, meaning one built around AI capabilities, with other startups in the same industry that were founded in the same period. The AI-native companies have a 13% higher share of engineers and a 20% higher share of senior staff, and their hierarchies are half a seniority level flatter. Their valuations are comparable to those of other startups, which the authors say suggests each employee is creating more value. The authors found that building AI into the product itself, rather than only giving staff AI tools, is a primary way these startups handle knowledge work without large teams of knowledge workers. AI-native startups are also 20 percentage points more likely to be based in San Francisco, and their staff are more likely to be male, to hold graduate degrees and to have studied at highly ranked universities. Studies of individual tasks have tended to find that AI helps less experienced workers the most. At the level of the firm, Kim and Koning found "a shift broadly in the opposite direction." ## AI and Entry-Level Jobs Researchers at Stanford's Digital Economy Lab, using payroll data from ADP covering millions of US workers through June 2026, found that employment of 22- to 25-year-olds in the [occupations most exposed to AI](https://www.sovereignmagazine.com/article/world-bank-ai-jobs-risk-developing-economies) is 19% below where it would be had it kept pace with less exposed peers. Experienced workers show no comparable gap. The [Stanford authors](https://digitaleconomy.stanford.edu/publication/canaries-in-the-coal-mine-six-facts-about-the-recent-employment-effects-of-artificial-intelligence/), Erik Brynjolfsson, Bharat Chandar and Ruyu Chen, found that the gap comes mainly from employers hiring fewer young workers rather than letting existing ones go, and that it has widened steadily since they first measured it in August 2025. Their examples include [young software developers](https://www.sovereignmagazine.com/article/anthropic-study-finds-ai-coding-assistants-speed-up-work-but-reduce-skills) and customer service representatives. They found no evidence of economy-wide job losses, and they describe their figures as early indicators rather than proof that AI is the cause. The authors suggest AI may affect entry-level workers more because it is better at replacing codified knowledge, the kind taught in formal education, than the experience senior employees build up on the job. ## Solo Founders at a Record High Solo founders accounted for 63% of the US corporations formed through Stripe Atlas in the second quarter of 2026 as of late May, the highest share Stripe has recorded. Stripe found that the most successful solo founders, measured by revenue in their first two years, were about twice as likely as the median founder to be building a product whose core function depends on AI. Maor Shlomo started Base44, [a tool for building apps without code](https://www.sovereignmagazine.com/article/vibe-coding-killing-saas-tools), alone as a side project. He took no outside funding and had eight employees when Wix bought the company for $80 million in cash in June 2025, six months after launch. One solo founder the Journal spoke to, Advait Paliwal, pays for about 20 Claude and Codex subscriptions instead of hiring staff. He interviewed four candidates over the past year, one of them a 14-year-old in Belgium, and hired none of them. ## Where Staying Small Stops Working Liz Wessel of First Round Capital told the Journal that whether a company can stay small depends on its product and how fast its revenue grows, and that companies growing very quickly still compete for engineers. Aileen Lee of Cowboy Ventures said mid-market and enterprise customers paying $60,000 to $75,000 a year want to buy from a company with people in it. She also said a startup raising a Series A is now expected to show $4 million to $5 million in revenue in its first year. [Lovable](https://www.sovereignmagazine.com/article/sweden-s-ai-darling-lovable-is-actually-a-us-company-and-that-s-europe-s-real-problem), whose software builds apps from text prompts, raised $400 million in August at a $13.3 billion valuation and plans to increase its headcount by half this year, to about 450 people. ## AI and the 2022 Funding Slump Carta, whose software manages share ownership records for tens of thousands of startups, recorded net hiring at those companies peaking in January 2022, ten months before OpenAI released ChatGPT. Hiring then fell as interest rates rose and venture funding dropped, so the decline began before ChatGPT and the AI tools that came after it. The Harvard and INSEAD comparison is made between startups in the same industry founded in the same period, so a funding climate shared by all of them does not account for the 25% gap. > "A significantly smaller team, using the tools we're building, can do more and do it better." > — Jack Dorsey, Block Block, the payments company run by Jack Dorsey, cut more than 4,000 jobs in February, nearly 40% of its staff. Dorsey wrote in a letter to shareholders that a smaller team using Block's own tools "can do more and do it better," and that he expects most companies to make similar changes within the next year. Kim and Koning say the move toward smaller, more senior teams may affect career opportunities, but they caution against reading their results as a forecast for the whole job market. ## FAQ **Q: What are AI-native startups?** In the Harvard and INSEAD study, an AI-native startup is one built around AI capabilities, both in how its staff work and in the product it sells. Floqer, Gamma and Lindy all sell products whose main function relies on AI, which lets them serve large numbers of customers with small teams. **Q: Are companies hiring less because of AI?** The evidence so far points to fewer hires of young workers in the jobs most exposed to AI, not broad job losses. The Stanford study finds no economy-wide displacement, but employment of 22- to 25-year-olds in the most exposed occupations is 19% below trend, mainly because fewer of them are being hired. Startup hiring also fell after January 2022 for reasons that came before AI tools were widespread, including higher interest rates and less venture funding. **Q: What entry-level jobs will AI take over?** The Stanford researchers found the largest declines for young workers in occupations where AI is mostly used to do tasks in place of people, with software development and customer service among their examples. Where AI mainly assists workers, employment is flat or rising. At AI-native startups, the Harvard and INSEAD study found the shortfall in entry-level roles and in management, while the share of engineers and senior staff is higher.