Every January, three reports drop and every freelance newsletter writes the same headline: *median freelance income is up X%*. The headline is not wrong. It is also not useful, because the median has stopped being a meaningful summary of how freelancers earn money.
The clearest read on what's actually happening comes from the MBO Partners 2025 State of Independence, the Staffing Industry Analysts coverage of the same dataset, and the Interview Guys' synthesis of multiple gig-economy surveys. Read together, the picture is bimodal: a fast-growing high-end and a fast-shrinking low-end, with very little middle.
The numbers, briefly
From MBO Partners' 2025 report on the US independent workforce:
- 72.9 million Americans identified as independent workers in 2025, ~45% of the labor force.
- 5.6 million US independents earned more than $100,000 in 2025 — up 19% year-over-year and 87% from 3 million in 2020.
- 27.6 million were full-time independents (essentially flat from 27.7M in 2024).
- 7.9 million were part-time independents — down from 11.1 million in 2022, a three-year decline.
- 37.4 million were "occasional" independents (irregular, periodic income).
- 74% of independents now use generative AI, up from 65% in 2024.
The headline most outlets will run on this is "freelance workforce hits record 72.9M." The story underneath is the part-time decline.
The middle has fallen out
Three years ago, "freelancing part-time" was a real economic category. Side gigs on Fiverr, evening Upwork shifts, weekend design work for friends-of-friends. That category is collapsing. The 11.1M → 7.9M decline over three years is roughly 29% of part-time freelancers leaving the segment.
There are three plausible explanations and they probably all matter:
- AI eating the bottom of the market. ChatGPT Agent and Sora 2 have eaten the kinds of small jobs that part-timers used to take — admin work, simple graphics, basic copywriting. We covered this dynamic in detail in our analysis of how AI agents crossed 40% of Upwork tasks.
- Labor market tightness. Through 2024–2025, real wage growth at W-2 jobs absorbed many people who used to side-gig because they had to. They no longer have to.
- Platform economics. Upwork, Fiverr, and Toptal have all hardened pricing, surfaced the more credentialed workers, and made it harder for the casual side-gigger to actually win bids. Per Snipework's Connects breakdown, a typical Upwork proposal costs $0.90 in Connects alone — the unit economics of casual freelancing on the platform have stopped working.
The 5.6M / $100K+ cohort is moving the opposite direction: up 19% in a year, 87% in five years. These are full-time freelancers with specialized skills, billing project work or retainers, often with one to three named clients each.
Why the average is misleading
If you compute mean income across 72.9 million people, you get a number that sounds reasonable — and tells you almost nothing about any specific freelancer's prospects.
Imagine the same arithmetic. Take the 5.6M earning >$100K (probably averaging $150–200K). Take the 7.9M part-time freelancers (probably averaging $15–25K). Take the 37.4M occasional freelancers (averaging $3–10K). The arithmetic mean lands somewhere in the $40–60K range. No actual freelancer earns that. They earn either ~$150K or ~$15K. The middle is empty.
What this means for where you position in 2026
The strategic implication is not "everyone should aim for $100K+." It is "the middle is being squeezed out from both ends — pick a side."
The high end is structural, not lucky. The 5.6M >$100K freelancers are not exceptional individuals. They are practitioners who specialized, productized, switched to retainers, and built repeat-client books. The path is reproducible. We've covered the components in detail across the freelance retainer playbook, retainer pricing replacing hourly, and productized services for freelancers.
The low end is becoming volunteerwork-adjacent. The 7.9M part-time number is going to keep falling. The economics no longer support generalist part-time freelancing on platforms. If you are in this category, the realistic options are: full-time freelance with a specialization, a W-2 job, or AI-augmented operator work.
The "occasional" tier is the largest and most fragile. 37.4M people earning some income occasionally is the bulk of the workforce — but it's mostly weekend work, hobby income, irregular gigs. It is real, but it is not a career.
What the 5.6M cohort actually looks like
From the same dataset, the demographic and behavioral pattern of high-income independents is consistent:
- Specialized. Most have one clearly-named role: "freelance brand strategist", "Series B SaaS landing-page designer", "fractional CMO for B2B SaaS", "fintech compliance writer". Not "freelance writer" or "freelance designer."
- Retainer-heavy. A typical $150K+ freelancer has 60–80% of their revenue on retainer or project deals lasting >3 months. They do not bill hourly.
- Three to seven named clients. Not 30. Not 200. The portfolio is small and durable.
- Direct-outreach pipeline. Most do not get clients from Upwork or Fiverr. They get clients from LinkedIn DMs, referrals, podcast appearances, and a personal newsletter.
- AI-augmented, not AI-resisted. Per the 74% AI adoption number, the high earners are using ChatGPT Agent, Sora 2, Cursor, Claude Code as accelerants — they ship more work in less time, then keep the margin.
If you map all of this onto the non-US freelancer working with US clients playbook the structural setup is the same regardless of where you live.
The 2026 game plan, distilled
Given the bimodal data, the highest-ROI moves for the next 12 months are:
- Pick a defensible specialization. "I do X for Y" is the single most important sentence in your business.
- Switch hourly clients to retainers or fixed-fee projects. The hourly model is what's collapsing.
- Build a direct-outreach pipeline. Stop relying on platforms whose unit economics no longer work for you.
- Use AI as a margin amplifier, not a competitor. Every hour AI saves you on execution is an hour you can spend on direction, sales, or ICP refinement.
- Track your income distribution, not just your total. Goal: 70%+ from your top 3 clients on retainer or multi-month deals.
FAQ
Q: Is the bimodal pattern unique to the US?
No. Similar patterns appear in Payoneer's freelancer income reports across global markets and in EU labor data. The shape is universal; the absolute numbers differ.
Q: I'm in the part-time / occasional bucket. Can I move up?
Yes, but the path is structural. Pick a specialization, build a repeatable offer, switch off platforms, and price by outcome. Most of the highest-paying freelance categories are accessible to specialists who commit to one lane.
Q: What about freelancers in their first or second year?
The MBO data doesn't separate them, but the cohort effect is real — top earners have on average 4–7 years of freelance experience. Your first two years are probably going to be in the lower buckets. The pivot to >$100K is what years 3–5 are for.
Q: How does this interact with the AI agent threat?
The cohort earning >$100K is also the most likely to use AI heavily. AI is not a threat to specialized, well-positioned freelancers. It's a threat to generalist execution work, which is what the part-time / occasional buckets do.
Q: What's the source data on the AI adoption number?
74% of independents using generative AI in 2025 vs 65% in 2024 is from the MBO Partners 2025 State of Independence press release. The trend is consistent across other surveys (Upwork Research Institute, Fiverr's Anywhere Worker Study).
Delivvo helps freelancers run client-by-client retainer and project work — contracts, invoices, deliverables, branded portal — so you can build the small, durable book of named clients the >$100K cohort runs. See how it works →
Written by The Delivvo team · May 7, 2026
More from the blog →
