Leaner Startups, More Founders: How AI Is Reshaping the Economics of Starting a Business
Key Findings
Businesses started in 2024 in AI-enabled industries are operating with fewer employees in their first year than those founded in 2023. At 12 months post-founding, 2024 cohorts have approximately 6% lower headcount on average.
At the same time that new businesses are getting smaller, business formation is also surging, potentially leading to overall net job growth. In AI-enabled industries, new firms are operating with smaller teams in their first year as productivity gains reduce the need for early hires. Meanwhile, business applications climbed to near historic highs. Starting in late 2025 business applications started to accelerate, reaching over 500,000 applications per month by late 2025 pointing to a surge in new firm creation.
The shift is being driven by SMBs located in tech hubs. The decline in early-stage headcount is strongest in major tech hubs, especially the San Francisco Bay Area, where businesses started in 2024 have roughly 16% fewer employees at the 12-month mark compared to the 2023 cohort. Other large tech hubs (e.g., NYC, Seattle, Austin, Boston) also show meaningful declines in 2024, around 10% on average. However outside of major tech hubs average firm sizes have not yet meaningfully changed.
Introduction
For decades, the relationship between business growth and hiring has been well understood. If a business wanted to do more, it needed to hire more. But what if that relationship is beginning to shift?
Over the past two years, AI has moved rapidly from an experimental technology to a practical tool embedded in everyday business operations. Tasks that once required additional staff such as drafting marketing content, summarizing research, generating code, analyzing data, and managing customer communication, can now be completed efficiently and effectively with AI tools.
As AI capabilities raise productivity, they could alter early growth patterns of newly formed businesses – leading to leaner companies – while at the same time reducing the barriers to entry for new entrepreneurs, causing more new firms to be established.
To explore this possibility, we compared how new businesses in AI-enabled industries scaled headcount in their first year, focusing on firms founded in 2023 versus 2024. We examined two sectors where AI tools are especially relevant:
Information, which contains many tech, software, and media business;
Professional, Scientific, and Technical Services, which includes businesses like law firms and accountants.
These industries have jobs that are concentrated in coding, writing, research, and digital communication, areas where AI capabilities have advanced rapidly.
Using payroll records, we tracked how brand new firms scaled headcount over their first 12 months. We use 2024 as the comparison year because it represents the first full year of widespread adoption of generative and agentic AI tools in business settings. While ChatGPT captured public attention in late 2022 and GPT-4 raised the capability bar in 2023, small business adoption remained limited through 2023. Only 23% of small businesses reported using generative AI in 2023, a figure that nearly doubled to 40% by 2024.Firms founded in 2024 were, for the first time, building from day one in an environment where AI tools were a practical reality for a meaningful share of small businesses.
AI-enabled firms are 6% smaller when they reach 12 months.
At 12 months post founding, firms in the Information and Professional Services sector launched in 2024 have about 6% fewer employees on average than comparable firms founded in 2023. The difference is modest in absolute terms, but the trend is highly statistically significant. Notably, this is not a widespread shift across the economy, outside of these sectors, industries where AI tools are less universally applicable for many job tasks, there is no discernible difference in early stage headcount growth.
The pattern is also geographically concentrated. The decline in early-stage headcount is strongest in major technology hubs, where AI tools, talent, and capital are most densely clustered. In the San Francisco Bay Area, businesses founded in 2024 have about 16% fewer employees at the 12-month mark than those founded in 2023. Other major tech hubs, including New York City, Seattle, Austin, and Boston, show a similar but smaller effect, with teams roughly 10% smaller on average.
Taken together, the evidence suggests an early shift in how startups scale. The trend is emerging first in AI-intensive industries and tech-forward regions, however if AI adoption continues to diffuse across sectors and geographies at its current pace, similar patterns could become more widespread across the economy.
Smaller Teams, But More Small Businesses
At first glance, smaller startup teams might sound like a warning sign for the labor market. But the broader employment story is more nuanced. Economists often distinguish between changes on the intensive margin and the extensive margin. The intensive margin refers to how large firms are, for example, how many employees each startup hires. The extensive margin refers to how many firms are created in the first place. Our findings suggest AI may be affecting both.
On the intensive margin, startups founded in 2024 are operating with smaller teams. In AI-enabled industries, they have about 6% fewer employees at the 12-month mark compared to similar firms founded in 2023
But on the extensive margin, business formation is accelerating. U.S. Census Bureau data show that new business applications rose in late 2024 and continued to rise through 2025. New business formation is now hovering near historically high levels, approaching 500,000 filings per month. And at the same time, Gusto survey data indicate that 30% of entrepreneurs say AI has made it easier to start a business.
Taken together, this points to a shift in the mechanics of entrepreneurship. AI may reduce the average size of new firms (the intensive margin) while increasing the total number of firms being created (the extensive margin). By lowering the cost and complexity of starting a business, AI could enable more entrepreneurship overall, meaning smaller teams per firm may be offset by more firms in total, with a neutral or even positive effect on job growth.
A Structural Shift, Not a Cyclical One
These findings are notable given the broader macroeconomic backdrop. Interest rates peaked in August 2023 and began to decline in late 2024, meaning firms founded in 2024 actually entered a more favorable financing environment. In theory, that should have supported stronger early-stage hiring relative to the 2023 cohort. Instead, we observe slower headcount growth.
If cyclical or financial conditions were driving the results, we would expect the opposite pattern. For that reason, the observed decline in early headcount is best interpreted as a lower bound estimate. The fact that it appears despite improving macro conditions is consistent with a structural shift in how new firms are scaling, rather than a temporary economic constraint.
Additionally, two pieces of evidence support the interpretation that AI adoption, rather than broader economic forces, is driving this shift.
First, the effect is concentrated in industries where AI tools are most directly applicable to core job tasks — coding, writing, research, and digital communication. If the decline were driven by general macroeconomic conditions or shifts in business strategy unrelated to AI, we would expect to see similar patterns across a wider range of industries. In fact, the difference in early-stage headcount between the 2023 and 2024 cohorts is not statistically significant for firms outside of these AI-enabled sectors.
Businesses founded in 2024 in other industries are scaling at roughly the same pace as those founded in 2023. This industry-specific pattern, combined with the geographic concentration in tech hubs where AI adoption is earliest, points to AI productivity gains as the most likely explanation.
Conclusion
For decades, entrepreneurship and early hiring moved together. Now, startups may be able to grow further before they need to expand their teams.
The early evidence suggests that AI is beginning to reshape how new businesses start and scale. We see this shift emerge first in AI-enabled industries and in major tech hubs. Firms founded in 2024 are operating with smaller teams in their first year, even as overall business formation remains elevated. At the same time, small businesses report meaningful productivity gains from AI tools, and many entrepreneurs say those tools are making it easier to get started.
It is still early, and adoption is uneven. But if AI continues to lower the cost and complexity of starting a company, more people may choose to launch businesses of their own. Small businesses are poised to play an even more central role in an AI-driven economy.
Methodology
This analysis uses anonymized and aggregated payroll records from Gusto's platform, which serves over 500,000 small and medium-sized businesses across the United States.
We compared firms founded in 2023 and 2024, tracking their headcount trajectories over the first 12 months of operation. To isolate the potential effects of AI on early-stage hiring, we focused on two NAICS sectors where AI tools are most directly applicable to core job functions: Information (NAICS 51), which includes technology, software, and media businesses; and Professional, Scientific, and Technical Services (NAICS 54), which includes consulting, legal, accounting, and research firms. We refer to these collectively as "AI-enabled industries." As a check, we also examined industries outside these sectors and found no significant difference in headcount between the 2023 and 2024 cohorts.
We further segmented firms by geography, classifying them into major technology hubs (San Francisco Bay Area, New York City, Seattle, Austin, Los Angeles, Washington D.C., and Boston) and all other regions.
In addition to Gusto payroll data, this analysis draws on U.S. Census Bureau Business Formation Statistics to measure trends in new firm creation, and Gusto survey data on small business AI adoption and its perceived impact on entrepreneurship.
This analysis is observational and does not establish a causal link between AI adoption and startup headcount changes. However, several features of the data support an AI-related interpretation: the effect is concentrated in AI-exposed industries and absent elsewhere; it is geographically concentrated in tech hubs; and it emerges despite improving macroeconomic conditions in 2024, which would be expected to support rather than suppress hiring. For these reasons, we interpret the observed decline as a conservative, lower-bound estimate of AI's effect on early-stage firm scaling.
