The most important number in American work isn't the unemployment rate — it's the one that says 72.9 million people no longer wait for a paycheck to arrive from someone else. That's MBO Partners' 2025 count of US independent workers, the 15th consecutive year they've measured it, and it's up again from 72.7 million in 2024. Within that total, 27.6 million work independently full-time, while 37.4 million earn independent income at least monthly. However you slice it, tens of millions of Americans are running a business of one.

But the headline everyone repeats — "half of America will freelance by 2027" — is a zombie stat, a 2017 projection that never came true and should be retired. The real story in 2026 is quieter and far more interesting: the independent workforce is growing slowly in headcount but rapidly in value. A record 5.6 million independents now earn $100,000 or more, up 19% in a single year and nearly double the 3 million of 2020. Skilled freelancers collectively earned $1.5 trillion. This is the two-track economy Michael Dermer keeps naming, showing up as a genuine, growing path to ownership — not a gig-work consolation prize.

The question stopped being "can you survive on your own." For millions of people, working for yourself now pays better than the job you left.
— Michael Dermer

Read one way, that's just a bigger gig economy. Read another way, it's the largest voluntary migration into entrepreneurship in modern history — millions of people learning to find clients, price work, and deliver on their own terms. And a new force is deciding who thrives in it.

Why "more clients" is the wrong lever — and what the data proves

Here's the trap the six-figure curve hides. The default freelance playbook is linear: need more money, take more clients, work more hours. But an hour billed is still an hour spent — freelancers hit the exact same ceiling employees do, just without the salary floor. The independents pulling away from the pack aren't the ones cramming in more billable hours. They're the ones who changed what an hour produces. And in 2026, the single biggest multiplier on an independent's hour is AI.

The numbers are stark. Demand for AI-tagged freelance skills grew 109% year over year on Upwork's marketplace, comparing full-year 2025 earnings against 2024. That's not a headcount number — it's an earnings-weighted signal that clients are pouring money into exactly these skills. Independents who can wire AI into a client's systems, generate and edit AI video, or build chatbots aren't competing on hourly rate anymore; they're pricing scarce, surging capability. This is the leverage principle rendered in freelance economics: a generic hour is worth one generic hour, but an hour of scarce, in-demand skill keeps commanding a premium that rises faster than inflation.

Taking more clients is not a strategy — it's a faster treadmill with no salary underneath it. The move is to make one hour worth more, not to find a twenty-fifth hour to sell.
— Michael Dermer

So which skills are actually repricing fastest — and by how much? The marketplace data sorts them cleanly.

Why independents are winning the AI race — and employees aren't

Look at who's actually adopting these tools and a decisive gap appears. MBO Partners found 74% of independents used generative AI in 2025, up from 65% a year earlier — and ahead of the 69% rate among traditionally employed workers. Upwork found 54% of skilled freelancers rate themselves advanced or expert with AI, and 62% use it several times a week versus 53% of full-time employees. Fiverr's survey found 76% of its freelancers use AI, and 98% of those plan to keep. Every source that measures both populations finds the same thing: independents are out-adopting employees.

The reason isn't that freelancers are smarter or younger. It's that they capture the value directly. When an employee saves nine hours a week with AI, that time mostly evaporates into more meetings or someone else's backlog — the salary is the same either way. When an independent saves those same nine hours (and MBO says AI-using independents save an average of nine hours weekly), it becomes their gain: more billable capacity, faster turnaround, higher margins, or simply a shorter work week. The incentive is pure and immediate, and it shows up in behavior. This is Dermer's "playground where nobody else is playing" — the freelancer who leans into AI isn't fighting a crowd, because the incentive to do so is strongest exactly where the org chart is thinnest.

An employee who saves nine hours hands them back to the company. An independent who saves nine hours keeps every one — that's the whole game.
— Michael Dermer

The stakes compound brutally. A freelancer who ignores AI for three years competes on the same hourly rate as everyone else, in a race to the bottom. A freelancer who spent those years building an AI-augmented workflow can deliver in a fraction of the time, take on more work, and charge for outcomes instead of hours. Same profession, same starting point — a canyon of difference in trajectory. Here's exactly how the two populations compare, head to head.

How to end up on the right side of the divide — starting this week

The comforting thing about the freelance economy in 2026 is that the barrier isn't credentials or capital — it's a decision about where you point your hours. Nearly 80% of independents plan to stay independent or grow their business, and only 10% want to return to full-time work, while 36% of skilled employees are considering going independent. The flow is almost entirely one direction. The people already on the independent side aren't leaving; the question for everyone else is how to build a foothold that compounds instead of a hustle that just rents your time.

The first move is to pick one AI-adjacent skill that's repricing and go deep, not wide. The data hands you the shortlist: AI integration (+178%), AI data work (+154%), AI video and image (+329% and +95%), chatbot development (+71%). You don't need all of them — you need one that a client will pay a premium for, built on top of a craft you already have. The second move is to capture the hours AI saves you as an asset, not just free time: turn a repeatable client deliverable into a productized service, a template, or a small system you can sell again. The third move is to protect the boring infrastructure that keeps an independent business alive — a tax buffer, a client pipeline, and rest — because the freelancers who burn out are almost always the ones who mistook a busy month for a stable business.

It's not OK to just tell someone to "learn AI" — that's handing them a homework assignment. You give them one skill that's repricing and one system to sell twice, and the freelancing becomes a business.
— Michael Dermer

The failure mode here isn't laziness — independents are, by definition, self-starters. It's staying generic: competing on hourly rate in a crowded lane while the premium quietly migrates to the AI-augmented specialists next door. Effort spent selling undifferentiated hours burns you out. Effort spent building a scarce, in-demand capability compounds into pricing power — and pricing power is what turns a freelancer into a founder.

What the freelance boom is really measuring

Zoom out from the survey percentages and the freelance boom is measuring a structural shift in how work itself gets packaged. For most of the last century, the unit of work was the job: a bundle of tasks assigned to one person, one employer, one salary. What the 2026 data shows is that bundle coming apart. Upwork found 77% of business leaders say AI is increasing their need for specialized, fractional talent rather than full-time roles — work is being re-scoped into projects before it's re-posted as jobs. The independent workforce isn't growing in spite of that; it's growing because of it.

And here's the part that should reframe how you see it: the leadership pipeline already runs through independent work. Upwork found 63% of C-level leaders have freelanced at some point, and 42% of CEOs have done skilled freelance work in their own field. Independent work isn't the runner-up to a "real" career — for a growing share of people it is the career, and increasingly the on-ramp to running something bigger. A country where 72.9 million people already know how to earn outside a paycheck, adopt new tools faster than institutions can, and keep every hour of leverage they create, is a country one deliberate step away from a wave of real, AI-augmented businesses. The pressure isn't going away. The question is whether you use the tools it handed you to build a treadmill or a foundation.