The One-Person Business: What AI Actually Replaced (And What It Didn’t)

The one-person business stopped being a fallback somewhere around 2023, and most people haven’t updated their mental model since.

The story you’ve probably absorbed goes like this: AI arrived, a few founders started making millions alone, and the traditional company became optional. It’s a good story. It’s also missing the part where the numbers live.

This article looks at what actually changed, using government data rather than success stories — including the figure that almost nobody puts in a headline.

The Scale Is Larger Than You Think

Start with the size of the category, because it reframes everything that follows.

According to the U.S. Census Bureau, there were 30,427,808 nonemployer establishments in 2023 — businesses with no paid employees at all. That’s 78.4% of all U.S. establishments, bringing in nearly $1.8 trillion in revenue, or 6.4% of GDP.

Read that percentage again. Roughly four out of five American businesses have zero employees.

That figure tends to surprise people because it contradicts the mental image of what a business is. We picture offices and payroll and org charts. The statistical reality is one person, a laptop or a van, and a client list.

This was true before AI existed. Working alone isn’t a new invention — it’s the overwhelming default, and always has been. What changed is the ceiling.

What AI Actually Replaced

Be specific here, because vague claims about “AI changing everything” are why so much advice in this space is useless.

What genuinely got absorbed is a particular category of work: the first draft. Not the finished thing — the rough version that used to consume hours and required a person.

That covers a lot of ground. Initial copy. Basic design assets. Research summaries. Code scaffolding. Routine customer responses. Data cleaning. Transcription and translation. Meeting notes.

Each of these used to be either a hire, a contractor, or an evening you weren’t getting back.

The change is easy to underestimate because it happened gradually and each individual case seems small. But stack them together and a single person now covers work that would have required a small team in 2021 — not at the same quality in every case, but at a quality that clears the bar for most commercial purposes.

Adoption reflects this. Roughly 21% of U.S. workers now use AI in some part of their job, up from 16% the year before — a fast shift for something that requires changing how people work.

What got absorbed	What didn’t
First drafts of almost anything	Deciding what should exist
Basic design and layout	Taste and judgment about quality
Research summarisation	Knowing which question to ask
Routine customer replies	Difficult conversations
Code scaffolding	Architecture decisions
Transcription and translation	Cultural and contextual nuance

What It Didn’t Replace

The column on the right of that table is where a solo operation actually lives or dies.

Every item in it shares a property: it requires someone to decide, and to be accountable for having decided. A tool can generate twenty options. It cannot tell you which one matches what you’re actually trying to build, and it cannot answer for the choice afterward.

This isn’t speculation. The measured employment data tells the same story — we covered that in detail in what AI still can’t do, where the gap between theoretical capability and actual delegation turns out to be much wider than the predictions suggested.

For a solo operator, the practical translation is uncomfortable: AI removed the reason you needed help with execution. It did nothing about the part where you have to know what you’re doing.

This is why the “anyone can do it now” framing misleads. The barrier moved rather than disappeared. It used to sit at capability — could you produce the thing at all. Now it sits at discernment — can you tell whether what you produced is any good, and whether it was worth producing.

The Number Nobody Puts in the Headline

Now the part that the million-dollar-solo-founder genre leaves out.

Take the Census figures and do the division. Nearly $1.8 trillion in revenue, spread across 30.4 million nonemployer businesses. That works out to roughly $59,000 per business, per year — in gross revenue, before any expenses.

That’s the average. The median is lower, because a small number of very large solo operations pull the average up.

So the honest picture is this: working alone is the most common form of business in America, and the typical operation generates something closer to a modest salary than to wealth.

None of this means the model doesn’t work. It means the distribution is wildly skewed, and almost all the content you read comes from the thin end of it.

Why the Success Stories Mislead

There’s nothing dishonest about a founder writing up how they built a seven-figure solo company. The distortion happens in aggregate.

You read twenty such stories and form an impression that this is what solo work looks like. You never read the thirty million that grossed $59,000, because nobody writes that post and nobody would share it.

This matters practically, not just philosophically. If your expectation is calibrated to the outliers, you will interpret normal early results as failure and quit something that was working.

Calibrate to the distribution instead, and $2,000 a month in year one reads as genuinely above average rather than as evidence you should have picked a different path.

Survivorship bias has a specific cost here, and it isn’t just disappointment. It pushes people to abandon working approaches early, and to chase the tactics that appear in outlier stories — which are frequently the least transferable part of those stories.

Where Solo Operators Actually Cluster

The Census sector data is more useful than most advice about “which business to start.”

In 2022, the three largest categories by number of establishments were Professional, Scientific and Technical Services with 4.01 million, Transportation and Warehousing with 3.85 million, and Real Estate with 3.15 million.

The first of those is where knowledge work sits — consulting, design, writing, development, analysis. It’s also the category AI affects most directly, in both directions: the tools compress the work, and the compressed work makes the category more accessible to people entering it.

Notice what isn’t on that list. No “AI startup” category. The solo operations that exist in volume are doing ordinary work for ordinary clients, with better tools than they had three years ago.

What Actually Constrains a Solo Business

If AI removed the execution bottleneck, something else became the binding constraint. Three candidates, in rough order of how often they bite.

Distribution. You can produce more than ever and still have nobody see it. Getting in front of the right people did not get easier — arguably it got harder, since everyone else’s output also multiplied.

Judgment. More options generated per hour means more decisions per hour. Deciding well under uncertainty is now the scarce skill, and it’s one of the few that genuinely improves with experience rather than tooling. The research on how we make those calls is genuinely counterintuitive — including why you may decide better in a second language than in your first.

Your own consistency. No colleagues, no manager, no structure. The failure mode of working alone is rarely a dramatic collapse. It’s a slow stop, usually somewhere around month three.

That third constraint is the one most people underestimate, because it doesn’t feel like a business problem while it’s happening. It feels like a bad week, then a busy month, then a gap you tell yourself you’ll close.

What This Means If You’re Starting

Four things follow from the data rather than from encouragement.

Expect the median, plan for the median. If $59,000 gross is the average, treat anything approaching it as a real outcome rather than a disappointment.

Compete on judgment, not output volume. Everyone has the same tools now. Producing more is no longer a moat; producing the right thing is.

Solve distribution before you scale production. The most common failure is building more and better content, product, or service for an audience that doesn’t exist yet. Ten people who reliably see your work are worth more than a back catalogue nobody has encountered.

Pick one capability and get genuinely good at it. The durable skills are not the ones the tools replaced — our breakdown of the skills that increase your income goes through what the labour data actually rewards, and it is not what most career advice suggests.

Frequently Asked Questions

How many solo businesses are there really? The U.S. Census Bureau counted 30,427,808 nonemployer establishments in 2023 — businesses with no paid employees — representing 78.4% of all U.S. establishments.

How much does a typical solo business make? Dividing total nonemployer revenue by the number of businesses gives roughly $59,000 in annual gross revenue. The median is lower, since a small number of large solo operations raise the average.

Did AI actually create this trend? No. Nonemployer businesses were already the overwhelming majority of U.S. businesses before generative AI existed. What AI changed is how much one person can produce, not whether people work alone.

Which work did AI genuinely absorb? Primarily first-draft work: initial copy, basic design, research summaries, code scaffolding, routine customer replies, transcription. Finished work still requires someone to judge and decide.

Is it realistic to reach seven figures alone? It happens, but it sits far out on the tail of the distribution. Treating it as the expected outcome rather than the exception is the most common calibration error in this space.

What is the biggest constraint today? Distribution, most often. Production capacity increased dramatically for everyone simultaneously, which made attention scarcer rather than easier to obtain.

Should I use AI tools or avoid them? Use them, but for the layer they actually handle. They compress execution. They do not supply judgment, and judgment is what the remaining value concentrates in.

Final Thoughts

The one-person business is not a new phenomenon that AI invented. It is the default structure of American business, and it has been for decades.

What changed is narrower and more interesting than the headlines suggest. A single person can now execute work that previously required several, which raises the ceiling considerably. It does nothing about knowing what to build, who to build it for, or how to get anyone to notice.

Put differently: the tools solved the problem of not having enough hands. Almost nobody’s business was actually failing for that reason.

The $59,000 average isn’t a discouraging number. It’s a calibrating one. It tells you what normal looks like, which is exactly what you need in order to recognise progress when it’s actually happening rather than measuring yourself against the thin tail of a very skewed distribution.

We’re also preparing a more detailed course walking through the exact systems used to build and monetize this site, step by step. Readers of these articles will get early access when it’s ready.

A Note from Lucas

I went into this expecting to write about how AI is making solo businesses viable, and the Census numbers quietly rearranged the argument for me. Four out of five American businesses already had zero employees before any of this happened — the model was never waiting on the technology. What genuinely stuck with me was the $59,000 average, because it’s the number that never appears in anything written about this topic, and it changes how you read your own early results. My honest opinion: the most useful thing AI gave a solo operator isn’t leverage, it’s the removal of an excuse. You can no longer say you’d build it if you had a team.

About the Author 

Lucas Palma is the founder and editor of Info Is The Key. He writes about building an online business with AI, working through primary sources and testing what he finds in practice.

Last Updated: September 2026 | Reading Time: 10 minutes

Sources

• U.S. Census Bureau, Census Bureau Data Tell the Small Business Story (2026)
• U.S. Census Bureau, Statistics Shed Light on Self-Employment by Sector and State
• U.S. Census Bureau, Business Owner Characteristics Press Release (November 2025)
• Pew Research Center, About 1 in 5 U.S. Workers Now Use AI in Their Job