Freelance Statistics 2026
By Axis Intelligence Research
Co-author: Mia Scarlett & David Park | Last updated: October 21, 2026 | License: CC BY 4.0
Fiverr closed June 2026 with 2.676 million annual active buyers, down 21.9% in a year, while each remaining buyer spent $368, up 15.6%. Upwork lost 4% of its active clients over the same twelve months. Freelancer.com’s marketplace volume fell 20.3% in the first half. MBO Partners, meanwhile, counts a record 74.9 million American independents.
Quick Answer
The freelance market in 2026 has more sellers and fewer buyers. All three listed freelance marketplaces reported shrinking client bases or volume in mid-2026, while survey counts of independent workers hit records. Axis Intelligence Research finds the platforms kept a larger share of each client dollar anyway: Upwork’s Freelance Platform Capture Rate (FPCR) rose to 19.83% and Fiverr’s to 28.0%. Official Census data shows the same pattern among solo tech freelancers: fewer businesses, higher average receipts.
Key Findings
- Fiverr reported 2.676 million annual active buyers on June 30, 2026, down 21.9% year over year, while annual spend per buyer rose 15.6% to $368 (Fiverr Form 6-K, July 29, 2026).
- Axis Intelligence Research calculates a Freelance Buyer Divergence Spread (FBDS) of 37.5 points for Fiverr and 9.1 points for Upwork as of June 30, 2026: both marketplaces are trading buyer count for buyer size, Fiverr four times faster.
- Axis Intelligence Research finds Upwork kept 19.83 cents of every client dollar in Q2 2026, up from 19.44 cents a year earlier, even as gross services volume fell 3.6% to $966.4 million (Upwork Form 8-K, August 10, 2026).
- Axis Intelligence Research finds that U.S. solo businesses in computer systems design fell 2.46% to 334,886 in 2024, the only decline among 14 freelance-heavy professions tracked, while their average receipts rose 4.89% to $62,902 (U.S. Census Bureau Nonemployer Statistics 2024, released September 29, 2026).
- Peer-reviewed research found a 21% drop in job posts for automation-prone writing and coding work within eight months of ChatGPT’s release, and a 17% drop for image-creation jobs (Demirci, Hannane and Zhu, Management Science, 2025).
How Many Freelancers Are There in 2026?
It depends on whether you count people or count transactions, and in 2026 the two counts are moving in opposite directions.
On the people side, MBO Partners’ 2026 State of Independence study puts the U.S. independent workforce at a record 74.9 million, with full-time independents up 114% since 2020 and 78% planning to stay independent or build a larger business (MBO Partners, 2026). MBO’s definition is broad: it includes consultants and freelancers but also temps and on-call workers, so it is a ceiling for “freelancers,” not a headcount of them. Inside that total, MBO identifies nearly 12 million Independent Service Professionals, the people who sell specialized expertise to businesses, and almost 6 million independents earning more than $100,000 a year.
Globally, the World Bank’s Working Without Borders report estimated 154 million to 435 million online gig workers, a category that combines online freelancing with microwork, and put the upper bound at up to 12% of the global labour force (World Bank, 2023). The report also found demand for online freelancers rising faster in developing countries than in industrialized ones.
For the federal count of U.S. independent contractors, which comes from a different instrument and answers a different question, see our gig economy statistics, where Axis Intelligence Research reconciles the BLS headline with its own second-job tables.
Freelancer headcount by source and definition
| Measure | Value | Scope | Source |
|---|---|---|---|
| U.S. independent workers (all types) | 74.9M | Consultants, freelancers, temps, on-call | MBO Partners, 2026 |
| U.S. Independent Service Professionals | Nearly 12M | Expertise sold to businesses | MBO Partners, 2026 |
| U.S. independents earning $100K+ | Almost 6M | All independents | MBO Partners, 2026 |
| U.S. businesses with no employees (2024) | 31,985,874 | Tax-filing solo and owner-only businesses, receipts of $1,000+ | U.S. Census Bureau, 2026 |
| EU-27 self-employed without employees (2025) | 17.68M (8.77% of employment) | Own-account workers, ages 15–64 | Eurostat, 2026 |
| Online gig workers, worldwide | 154M–435M | Online freelancing + microwork | World Bank, 2023 |
| Netherlands: primarily self-employed share | Around 12% of employed | Mid-2026 | De Nederlandsche Bank, 2026 |
David Park: The fear behind most “how many freelancers” searches is a simple one: am I about to compete with everyone? The supply numbers say yes, the field is bigger than it has ever been. But 74.9 million is a count of people who earned independently at all, and the median member of that group is not bidding against you for the same requisition. The number that matters for your pipeline is on the demand side, and that number went the other way this year.
How Much Do Freelancers Earn? Census Data on 31.99 Million Solo Businesses
The most complete official count of U.S. freelance businesses is not a survey. It is the Census Bureau’s Nonemployer Statistics, built from federal business income tax records: every business with no paid employees and at least $1,000 in annual receipts ($1 in construction). The 2024 edition, released September 29, 2026, counts 31,985,874 nonemployer businesses with $1.861 trillion in receipts, up 5.12% from 30,427,808 in 2023 (U.S. Census Bureau, Nonemployer Statistics 2024).
A nonemployer business is a tax return, not a person: one freelancer can file for two businesses, and many are side businesses run alongside a salaried job. Receipts are gross revenue before expenses, not take-home pay. With those two cautions, the file answers the question freelance surveys cannot: what solo businesses in each profession actually gross.
Axis Intelligence Research calculates that the average U.S. nonemployer business grossed $58,194 in 2024, but 60.35% grossed under $25,000 and only 13.31% (4,257,187 businesses) reached $100,000 or more. The average is pulled up by a small top tier; the typical solo business is a part-time income.
Freelance Micro-Receipts Share (FMRS)
The Freelance Micro-Receipts Share (FMRS) is an Axis Intelligence Research metric: the share of nonemployer businesses in an industry with annual receipts under $10,000. It measures how much of a freelance profession is occasional, side-income activity rather than a working business. A falling FMRS means the profession is professionalizing; a rising one means it is filling with casual entrants.
FMRS = nonemployer businesses with receipts under $10,000 ÷ all nonemployer businesses × 100
(Census receipts-size classes "under $5,000" + "$5,000 to $9,999")
All U.S. nonemployers, 2024: 37.29% (2023: 37.84%)
Computer systems design & related, 2024: 35.99% (2023: 37.11%)
Independent artists, writers & performers: 54.57%
Legal services: 27.36%
Solo freelance businesses by profession, United States, 2024
| Industry (NAICS) | Businesses, 2024 | Change vs 2023 | Average receipts | Change in average | FMRS | Share at $100K+ |
|---|---|---|---|---|---|---|
| Management, scientific & technical consulting (5416) | 1,187,781 | +9.78% | $60,339 | +0.84% | 37.66% | 14.15% |
| Advertising, public relations & related (5418) | 221,163 | +6.40% | $62,029 | +1.97% | 38.22% | 14.08% |
| Translation & interpretation (54193) | 56,914 | +5.00% | $28,040 | +5.39% | 44.19% | 4.54% |
| Photographic services (54192) | 243,502 | +2.89% | $29,379 | +1.90% | 46.90% | 6.25% |
| Document preparation (56141) | 98,949 | +2.83% | $19,289 | +4.76% | 49.21% | 2.36% |
| Motion picture & video (5121) | 132,060 | +2.63% | $45,318 | +2.80% | 43.36% | 9.93% |
| Independent artists, writers & performers (71151) | 1,115,783 | +2.55% | $32,015 | +3.29% | 54.57% | 5.13% |
| Specialized design services (5414) | 284,294 | +2.36% | $46,837 | +1.63% | 42.60% | 10.76% |
| Accounting, tax prep & bookkeeping (5412) | 405,112 | +2.09% | $35,869 | +3.87% | 41.07% | 8.14% |
| Architectural & engineering services (5413) | 239,366 | +1.53% | $62,435 | +1.49% | 32.58% | 16.77% |
| All other professional & technical services (54199) | 853,017 | +1.34% | $71,993 | +3.72% | 35.09% | 16.09% |
| Legal services (5411) | 275,815 | +0.66% | $88,270 | +4.41% | 27.36% | 24.61% |
| Educational services, incl. tutors (611) | 924,679 | +3.49% | $18,746 | +2.32% | 63.10% | 3.11% |
| Computer systems design & related (54151) | 334,886 | −2.46% | $62,902 | +4.89% | 35.99% | 16.24% |
| Professional, scientific & technical services (54), all | 4,233,838 | +3.88% | $58,928 | +2.52% | 37.68% | 14.03% |
| All U.S. nonemployer businesses | 31,985,874 | +5.12% | $58,194 | +1.01% | 37.29% | 13.31% |
| Contrast: taxi & limousine incl. ridesharing (4853) | 1,602,297 | +16.41% | $29,924 | −1.91% | 38.16% | 6.10% |
| Contrast: couriers & messengers (492) | 1,595,662 | +10.94% | $14,873 | +8.74% | 60.04% | 1.16% |
Source: Axis Intelligence Research calculations on U.S. Census Bureau Nonemployer Statistics 2023 and 2024, U.S. files. Averages are receipts ÷ businesses. Census applies noise infusion to receipts; every value used here carries the low-noise flag (changed by less than 2%).
The finding. Computer systems design, the profession most exposed to AI coding tools, is the only one of the 14 freelance-heavy professions in the table where the number of solo businesses fell in 2024. The survivors earned more: average receipts rose 4.89%, one of the largest gains on the list, and the FMRS fell 1.12 points as casual sellers left. That is the same fewer-but-bigger pattern Upwork and Fiverr report, observed independently in IRS-based federal data a year earlier.
Management consulting moved the other way, adding 105,820 solo businesses (+9.78%) with nearly flat average receipts. That fits MBO’s finding that expertise sold to businesses is where independent work is growing, and Upwork’s AI Strategy & Consulting growth.
David Park: If you are deciding whether to go independent, the FMRS column is the reality check the averages hide. More than half of independent artists, writers and performers grossed under $10,000. In legal services, about one in four solo practices cleared $100,000. Same tax form, very different careers. Before quitting a salaried role, compare your realistic first-year bookings with the FMRS line for your profession, not with the average.
Are Freelance Clients Disappearing? What Upwork, Fiverr and Freelancer.com Reported
Yes, on the open marketplaces. Every listed freelance marketplace lost buyers or volume in the first half of 2026, and each management team attributed part of it to AI absorbing low-complexity work.
Fiverr: 21.9% fewer buyers, 15.6% more spend per buyer
Fiverr’s second-quarter filing shows 2.676 million annual active buyers on June 30, 2026, against 3.425 million a year earlier. Marketplace revenue fell 15.5% to $63.1 million. Spend per remaining buyer rose to $368 from $318 (Fiverr Form 6-K, July 29, 2026).
Fiverr’s own explanation is that AI is absorbing high-volume, transactional orders while larger projects hold up: clients completing $1,000-plus projects grew 13% on a trailing basis. The company cut its full-year 2026 revenue guidance to $356 million–$372 million and guided the third quarter to $80 million–$88 million, citing demand and traffic weakness in categories most exposed to automation.
Fiverr does not publish gross merchandise value (GMV) in this release, but it defines annual spend per buyer as trailing-twelve-month GMV divided by annual active buyers. Multiplying the two back together, Axis Intelligence Research estimates Fiverr’s trailing GMV at roughly $984.8 million to June 2026, down about 9.6% from roughly $1,089.2 million a year earlier. That is less than Upwork processed in a single quarter.
Upwork: fewer clients, bigger clients, and an AI work category growing fast
Upwork reported gross services volume (GSV) of $966.4 million in Q2 2026, down 3.6% from $1,002.7 million, and 763,000 active clients against 796,000 (Upwork Form 8-K, August 10, 2026). GSV is the total amount clients spend on the marketplace and managed services, including fees. GSV per active client hit a record $5,230, up 5%.
The growth is concentrated. GSV from AI-related work rose more than 22%, AI Strategy & Consulting more than 50%, and the Business Plus tier for small businesses 174%, with its client count up 219%. Upwork guided full-year 2026 revenue to $730 million–$750 million.
Freelancer.com: volume down 20.3%, and not only because of AI
Freelancer Limited’s marketplace GMV fell 20.3% to A$42.4 million in the half to June 30, 2026, from A$53.2 million; in constant currency the decline was 11.7%. Marketplace revenue fell 23.2% to A$14.9 million (Freelancer Limited, ASX, July 28, 2026).
Freelancer.com is the useful control case. Management named three causes, and two had nothing to do with demand: two-factor login friction that drove returning-customer churn, and a payment-gateway migration that cut acceptance rates. The third was AI-adjacent in an unexpected way: aggressive AI scraping forced rate limits that led to search de-indexation. Meanwhile AI development projects reached 7.4% of the marketplace’s Q2 volume, according to the same filing.
Listed freelance marketplaces, mid-2026
| Platform | Demand metric | Latest | Year earlier | Change | Source |
|---|---|---|---|---|---|
| Fiverr | Annual active buyers | 2.676M | 3.425M | −21.9% | Fiverr 6-K, Jul 29, 2026 |
| Fiverr | Annual spend per buyer | $368 | $318 | +15.6% | Fiverr 6-K, Jul 29, 2026 |
| Upwork | Active clients | 763K | 796K | −4.15% | Upwork 8-K, Aug 10, 2026 |
| Upwork | GSV per active client | $5,230 | — | +5% | Upwork 8-K, Aug 10, 2026 |
| Upwork | Gross services volume, Q2 | $966.4M | $1,002.7M | −3.6% | Upwork 8-K, Aug 10, 2026 |
| Freelancer.com | Marketplace GMV, H1 | A$42.4M | A$53.2M | −20.3% | Freelancer Ltd, ASX, Jul 28, 2026 |
Mia Scarlett: Read the three filings together and the shape is the same in each: the denominator shrinks, the per-unit number rises, and the press release leads with the per-unit number. That is a legitimate strategy, not a disguise. But it changes what these companies are. A marketplace built on millions of $50 orders and a marketplace built on fewer $5,000 engagements carry different acquisition costs and different churn risk, and the guidance cuts at Fiverr say the transition is costing more revenue than the per-buyer gains replace this year.
How Fast Are Freelance Marketplaces Losing Small Buyers? The Freelance Buyer Divergence Spread (FBDS)
The Freelance Buyer Divergence Spread (FBDS) is an Axis Intelligence Research metric that measures how fast a marketplace is trading buyer count for buyer size. It subtracts the year-over-year change in active buyers from the year-over-year change in spend per buyer, using the company’s own filed figures.
FBDS = (YoY % change in spend per active buyer) − (YoY % change in active buyers)
Fiverr, June 30, 2026: 15.6 − (−21.9) = 37.5 points
Upwork, June 30, 2026: 5.0 − (−4.15) = 9.1 points
A reading near zero means the buyer base is stable in both size and mix. A high reading means the platform is shedding many small buyers while the survivors spend more, which is what upmarket repositioning looks like in the numbers.
FBDS reading, June 30, 2026 (baseline reading)
| Platform | Spend per buyer, YoY | Active buyers, YoY | FBDS | Source |
|---|---|---|---|---|
| Fiverr | +15.6% | −21.9% | 37.5 pts | Axis calculation on Fiverr 6-K |
| Upwork | +5% | −4.15% | 9.1 pts | Axis calculation on Upwork 8-K |
What the reading says. Fiverr’s spread is four times Upwork’s. Fiverr’s marketplace is built around “Gigs” ordered from a catalog, the format best suited to the small, repeatable tasks both companies say automation is absorbing; Upwork’s growth is coming from small-business and enterprise tiers with larger scopes. The divergence is the market sorting itself by task complexity.
What would reverse it. If Fiverr’s active-buyer decline stabilizes while spend per buyer keeps rising, the spread falls and the repositioning has worked. If buyers keep falling and spend per buyer flattens, the spread falls for the wrong reason, and that is the reading to watch in Fiverr’s Q3 2026 filing.
David Park: For a freelancer, FBDS translates into a pricing decision. A falling buyer count with rising spend per buyer means the client who stays is buying something larger. If your offer is still priced and packaged as a single small task, you are selling into the shrinking half of the market. The comp-band move here is the same one employees make: stop quoting the entry rate for the entry task.
How Much Do Upwork, Fiverr and Freelancer.com Keep? The Freelance Platform Capture Rate (FPCR)
The Freelance Platform Capture Rate (FPCR) is an Axis Intelligence Research metric: platform revenue divided by the gross volume clients spent over the same period. It answers a question freelancers and clients both ask, which is how much of each dollar the platform keeps, using filed numbers rather than published fee schedules, which differ by plan, contract type and promotion.
FPCR = platform revenue ÷ gross volume (same period) × 100
Upwork, Q2 2026: $191.660M ÷ $966.439M = 19.83%
Upwork, Q2 2025: $194.939M ÷ $1,002.650M = 19.44%
Fiverr, TTM June 2026: marketplace take rate, company-reported = 28.0%
Fiverr, TTM June 2025: company-reported = 27.6%
Freelancer.com, H1 2026: A$14.9M ÷ A$42.4M = 35.14%
Freelancer.com, H1 2025: A$19.4M ÷ A$53.2M = 36.47%
FPCR by platform, 2025 vs 2026
| Platform | Period | FPCR, prior year | FPCR, latest | Client dollar remaining after platform revenue | Source |
|---|---|---|---|---|---|
| Upwork | Q2 | 19.44% | 19.83% | 80.17¢ | Axis calculation on Upwork 8-K |
| Fiverr | TTM to June | 27.6% | 28.0% | 72.0¢ | Fiverr 6-K (take rate) |
| Freelancer.com | H1 | 36.47% | 35.14% | 64.86¢ | Axis calculation on Freelancer Ltd ASX filing |
The “client dollar remaining” column is an upper bound on what reaches freelancers, because payment processing, currency conversion and taxes come out of what is left.
The finding. Two of the three marketplaces raised their capture rate while their volume shrank. Upwork’s FPCR rose 0.39 points on falling GSV; Fiverr’s take rate rose 0.4 points while buyers fell by more than a fifth. Freelancer.com is the exception, and its exception is instructive: its capture fell because payment acceptance problems and search de-indexation cut revenue faster than volume.
Mia Scarlett: A rising take rate on falling volume is the classic late-cycle marketplace move: when you cannot grow the pie, you price the slice. It works until it does not, because the take rate is also the price of using the platform rather than invoicing the client directly. The most useful number on this page for a CFO deciding whether to route contractors through a marketplace is the FPCR, compared against what an employer-of-record or direct contract costs.
Our 1099 vs W-2 calculator runs the other half of that comparison for U.S. contractors.
Is AI Replacing Freelancers? What the Research Measured
The peer-reviewed evidence says AI reduced demand for exposed freelance work quickly, and did not spare experienced freelancers.
Demirci, Hannane and Zhu, studying a leading global freelancing platform in Management Science (volume 71, issue 10, 2025), found a 21% decrease in job posts for automation-prone writing and coding jobs relative to manual-intensive jobs within eight months of ChatGPT’s release, and a 17% decrease in image-creation job posts after image generators arrived. They also found the remaining automation-prone posts were more complex and paid more (Imperial College London record; DOI 10.1287/mnsc.2024.05420).
Hui, Reshef and Zhou, using Upwork data in Organization Science (2024), found freelancers in highly exposed occupations saw about a 2% decline in contracts and about a 5% drop in earnings, and that the declines were largest for experienced, higher-rated, higher-priced freelancers (the authors’ summary at Brookings, July 8, 2025; DOI 10.1287/orsc.2023.18441).
Measured effect of generative AI on freelance demand
| Study | Outcome | Effect | Segment | Source |
|---|---|---|---|---|
| Demirci, Hannane & Zhu (2025) | Job posts | −21% | Writing and coding vs. manual-intensive | Management Science 71(10) |
| Demirci, Hannane & Zhu (2025) | Job posts | −17% | Image creation | Management Science 71(10) |
| Hui, Reshef & Zhou (2024) | Contracts | About −2% | AI-exposed occupations, Upwork | Organization Science |
| Hui, Reshef & Zhou (2024) | Earnings | About −5% | AI-exposed occupations, Upwork | Organization Science |
Where the studies and the 2026 filings disagree. Both studies measured the first months after ChatGPT’s late-2022 release. The 2026 filings show the effect did not fade: Fiverr explicitly cited “persistent weakness” in automation-exposed categories in July 2026. But the filings also show something the studies could only hint at, a second market forming in the same platforms. Upwork’s AI Strategy & Consulting category grew more than 50% and Freelancer.com’s AI development work reached 7.4% of volume. AI is removing one kind of freelance demand and creating another, and they are not the same freelancers.
David Park: The finding that should change behavior is the one about experienced freelancers. The usual 11pm reassurance is “AI replaces juniors, seniors are safe.” Hui and Reshef’s Upwork data says the opposite held for writing and image work: the higher-rated, higher-priced freelancers lost more, likely because AI substitutes best for the polished deliverable clients were paying a premium for. Seniority protects you when it means judgment and scope, not when it means a better version of the same deliverable.
Are Freelancers Leaving Self-Employment in Europe? Country Ranking 2025
Yes, slowly, and while total employment keeps rising. Eurostat’s Labour Force Survey counts 17.68 million self-employed people without employees in the EU-27 in 2025, the closest official equivalent to solo freelancers, or 8.77% of employment among people aged 15 to 64 (Eurostat, lfsa_egaps). Eurostat calls them own-account workers.
That number peaked at 17.91 million in 2023, the highest level since Eurostat’s 2021 series break, and has fallen for two consecutive years: −0.59% in 2024 and −0.72% in 2025. Over the same 2025, total EU employment rose 0.47%. Axis Intelligence Research calculates that the EU lost 232,700 solo self-employed workers between 2023 and 2025 (−1.3%), and that their share of employment slipped from 8.96% in 2021 to 8.77% in 2025.
The mix is changing too. Women made up 37.36% of EU own-account workers in 2025, up from 35.3% in 2021.
EU-27 solo self-employed workers by country, 2025
| Rank | Country | Own-account workers, 2025 | Share of employment, 2025 | Share, 2024 | Change in count, 2024–2025 | Change in count, 2021–2025 |
|---|---|---|---|---|---|---|
| 1 | Greece | 721.1K | 17.16% | 19.05% | −8.68% | −6.70% |
| 2 | Poland | 2,345.3K | 14.00% | 14.07% | −0.63% | −0.40% |
| 3 | Czechia | 649.3K | 12.82% | 12.95% | +0.06% | +2.33% |
| 4 | Italy | 2,921.4K | 12.56% | 12.76% | −1.14% | −5.28% |
| 5 | Slovakia | 311.7K | 12.24% | 11.63% | +4.60% | +3.80% |
| 6 | Netherlands | 1,068.6K | 11.34% | 12.01% | −5.51% | +9.81% |
| 7 | Belgium | 514.0K | 10.20% | 9.91% | +4.18% | +11.91% |
| 8 | Lithuania | 136.8K | 9.85% | 9.84% | +0.37% | +14.38% |
| 9 | Spain | 2,138.5K | 9.81% | 9.82% | +2.29% | +4.40% |
| 10 | Malta | 31.2K | 9.67% | 9.58% | +4.70% | +11.83% |
| 11 | Romania | 698.8K | 9.18% | 10.23% | −12.13% | −10.86% |
| 12 | Portugal | 450.1K | 8.95% | 8.80% | +4.75% | +13.18% |
| 13 | Latvia | 73.8K | 8.92% | 7.80% | +14.77% | +19.42% |
| 14 | Slovenia | 85.7K | 8.77% | 8.97% | −2.50% | +6.46% |
| 15 | France | 2,329.2K | 8.09% | 7.87% | +4.17% | n/a (series break) |
| 16 | Hungary | 353.7K | 7.79% | 7.32% | +5.65% | +4.06% |
| 17 | Finland | 190.6K | 7.64% | 7.55% | +0.79% | −9.67% |
| 18 | Ireland | 202.6K | 7.54% | 7.46% | +3.05% | +9.28% |
| 19 | Estonia | 44.7K | 6.89% | 6.84% | 0.00% | +19.20% |
| 20 | Cyprus | 33.3K | 6.87% | 7.08% | −0.89% | −6.46% |
| 21 | Bulgaria | 192.5K | 6.85% | 6.38% | +7.00% | n/a (series break) |
| 22 | Austria | 279.6K | 6.36% | 6.43% | −1.06% | +11.53% |
| 23 | Croatia | 101.2K | 6.06% | 5.77% | +6.19% | +21.63% |
| 24 | Luxembourg | 16.3K | 5.04% | 4.87% | +4.49% | +8.67% |
| 25 | Sweden | 250.0K | 4.97% | 5.07% | −1.81% | −1.46% |
| 26 | Denmark | 122.6K | 4.19% | 4.45% | −5.69% | n/a (series break) |
| 27 | Germany | 1,413.9K | 3.48% | 3.68% | −5.79% | −8.11% |
| EU-27 | 17,676.3K | 8.77% | 8.87% | −0.72% | +2.43% |
Source: Axis Intelligence Research calculations on Eurostat lfsa_egaps (EU Labour Force Survey, ages 15–64, data updated September 10, 2026). Shares are own-account workers ÷ employed persons. 2021–2025 changes are not shown where Eurostat flags a series break after 2021 (France, Bulgaria, Denmark).
Germany has the lowest share in the EU at 3.48%, and it is still falling: German own-account workers dropped 5.79% in 2025 and 8.11% since 2021. At the other end, Greece leads at 17.16% but lost 8.68% of its solo self-employed in a single year, and Romania fell 12.13%. Eurostat’s own-account category includes farmers and small shopkeepers as well as freelancers, so the countries with the highest shares are not necessarily the largest freelance markets.
The growth is in smaller markets, the Baltic states, Croatia, Belgium and Portugal: Latvia’s solo self-employed rose 14.77% in 2025, Croatia’s 21.63% since 2021.
The Netherlands: two official sources, one direction
The Netherlands shows how definitions matter and why direction is the safer signal. Eurostat counts 1,068.6 thousand Dutch own-account workers aged 15 to 64 in 2025, down 5.51% in a year. De Nederlandsche Bank, using a broader national measure, reported on October 15, 2026 that the number of self-employed fell by around 131,000 between the end-2024 peak of almost 1.3 million and mid-2026, with only 7% of leavers becoming unemployed (De Nederlandsche Bank, October 15, 2026). Axis Intelligence Research calculates the DNB decline at roughly 10.1% of the peak (a floor, since the peak is “almost” 1.3 million). The levels differ; both sources show Dutch solo self-employment shrinking from a 2024 high.
Mia Scarlett: The European numbers rule out the easy story. If AI were simply pushing knowledge workers into freelancing, the EU count would be rising with employment; it is falling against it. If AI were simply destroying freelance work, the decline would be uniform; instead Germany and the Netherlands fall while Belgium, Portugal and the Baltics grow. The DNB finding that only 7% of Dutch leavers became unemployed points to the same reading as the U.S. data: workers are moving between arrangements, and the open-marketplace end of freelancing is where demand is thinning.
What Freelancers and Hiring Managers Disagree About
MBO’s 2026 enterprise survey adds a demand-side view that rarely makes headlines: 54% of enterprises expect AI to increase their use of independent talent over the next two years (MBO Partners, 2026). That sits awkwardly beside the marketplace contraction until you read it as a statement about which talent: enterprises want applied AI skills they cannot hire full-time, the same category growing inside Upwork.
The same study found 61% of independent professionals rate schedule flexibility as very important, but only 36% of enterprises recognize that. For a client competing for scarce specialists, that 25-point gap is a free lever.
| Survey finding (MBO Partners, 2026) | Independents | Enterprises |
|---|---|---|
| Schedule flexibility rated very important | 61% | 36% recognize it |
| Expect AI to increase use of independent talent | — | 54% |
| Plan to stay independent or grow a larger business | 78% | — |
Related reading: our AI job market statistics track the same applied-AI skills in full-time postings, and our side hustle statistics cover second-job holders, many of whom freelance part-time.
Methodology
Collection. Every figure on this page was retrieved from its producer on October 21, 2026: Upwork’s Q2 2026 results filed with the SEC on Form 8-K (August 10, 2026); Fiverr’s Q2 2026 results filed on Form 6-K (July 29, 2026); Freelancer Limited’s H1 FY26 results lodged with the ASX (July 28, 2026); MBO Partners’ 2026 State of Independence landing page; the World Bank’s Working Without Borders record; De Nederlandsche Bank’s October 15, 2026 analysis; Eurostat’s lfsa_egaps table (EU Labour Force Survey, employed persons by professional status, ages 15–64, 2016–2025, extracted October 21, 2026); the U.S. Census Bureau’s Nonemployer Statistics U.S. files for 2023 and 2024 (record layout 2017–2024, receipts in thousands of dollars, all legal forms, U.S. total rows); and the two peer-reviewed AI studies via their institutional records and the authors’ own published summary. Company figures are company-reported. Survey figures are labeled with their survey sponsor.
Freelance Platform Capture Rate (FPCR). Platform revenue divided by gross volume for the same period. For Upwork, total revenue (marketplace plus enterprise) over total GSV, both as filed. For Fiverr, the company’s own marketplace take rate, which Fiverr defines identically as marketplace revenue over GMV, on a trailing-twelve-month basis. For Freelancer.com, marketplace segment revenue over marketplace GMV for the half-year. Freelancer.com’s revenue includes membership and upgrade fees that are not per-transaction commission, so its FPCR overstates commission alone; it still measures what the platform earns relative to the volume it intermediates. The three are reported on different periods (quarter, trailing year, half-year) and are compared on direction of change, not ranked as a league table.
Freelance Buyer Divergence Spread (FBDS). Year-over-year percentage change in spend per active buyer minus year-over-year percentage change in active buyers, in percentage points. Fiverr inputs are the company-reported changes (+15.6%, −21.9%). Upwork’s spend input is the company-rounded +5%; its client change (−4.15%) is calculated from the filed counts of 763,000 and 796,000. Freelancer Limited does not disclose a buyer count, so no FBDS is computed for it. June 30, 2026 is the baseline reading.
Freelance Micro-Receipts Share (FMRS). Nonemployer establishments in receipts-size classes “under $5,000” and “$5,000 to $9,999”, divided by all nonemployer establishments in the same industry and year. Average receipts are total receipts (thousands of dollars × 1,000) divided by establishments. The share at $100,000 or more is all establishments minus those in the five classes below $100,000. Size-class counts sum to each industry total within two establishments, a gap caused by Census disclosure protection. Industries were chosen for their concentration of knowledge-work freelancers; ridesharing and couriers are shown only as contrast, and app-based work is covered on our gig economy statistics page.
EU solo self-employment. Own-account workers (self-employed without employees) ÷ employed persons, both ages 15–64, by country and year. Eurostat flags a series break in 2021 for most countries, so trend comparisons start in 2021. France (2024), Bulgaria (2024) and Denmark (2023) carry later breaks and are excluded from 2021–2025 comparisons; year-on-year 2024–2025 changes remain comparable. Spain and France carry Eurostat’s “definition differs” flag throughout. Own-account workers include farmers and small traders as well as knowledge freelancers; Eurostat does not split them by occupation in this table.
Fiverr implied GMV. Annual active buyers × annual spend per buyer, which reproduces trailing GMV under Fiverr’s own definition. Because both inputs are rounded in the release, the result is an estimate.
What we did not do. We do not add Upwork clients and Fiverr buyers into a single “marketplace buyers” total: each company defines an active buyer differently and the populations overlap. We do not merge MBO’s survey count with platform data, because one counts people who earned independently and the other counts paying clients. We do not publish a global “freelance market size” in dollars: the figures in circulation come from commercial estimates without auditable methods.
About This Dataset
Contents. 438 observations: Eurostat own-account worker counts, shares of employment and changes for the EU-27 and each member state, 2021–2025; U.S. Census Bureau nonemployer counts, receipts, averages, FMRS and $100K+ shares for 18 industry rows in 2023 and 2024; Upwork, Fiverr and Freelancer.com operating metrics for mid-2025 and mid-2026; FPCR and FBDS readings with their inputs; Fiverr implied GMV; MBO Partners 2026 independent-workforce and enterprise survey figures; World Bank online gig workforce estimates; De Nederlandsche Bank self-employment figures; and effect sizes from two peer-reviewed studies of generative AI and freelance demand.
Provenance. Each row carries source_org, source_document, source_url, retrieved_date, is_primary and axis_calculated; every calculated row has a method_note with its formula and inputs.
License and citation. CC BY 4.0. Free to reuse with attribution: Axis Intelligence Research, Freelance Statistics 2026, 2026.
Cite This Research
APA: Axis Intelligence Research, Scarlett, M., & Park, D. (2026, October 21). Freelance statistics 2026: More freelancers, fewer buyers, and what the platforms keep. Axis Intelligence. https://axis-intelligence.com/freelance-statistics/
MLA: Axis Intelligence Research, Mia Scarlett, and David Park. “Freelance Statistics 2026: More Freelancers, Fewer Buyers, and What the Platforms Keep.” Axis Intelligence, 21 Oct. 2026, axis-intelligence.com/freelance-statistics/.
Chicago: Axis Intelligence Research, Mia Scarlett, and David Park. “Freelance Statistics 2026: More Freelancers, Fewer Buyers, and What the Platforms Keep.” Axis Intelligence, October 21, 2026. https://axis-intelligence.com/freelance-statistics/.
Freelance Market Questions, Answered With 2026 Data
Is freelancing dying because of AI?
No, but the cheapest tier of it is shrinking. Fiverr lost 21.9% of its annual active buyers in the year to June 30, 2026 while spend per remaining buyer rose 15.6%, and Upwork’s AI Strategy & Consulting work grew more than 50% in Q2 2026. Demand is moving from small, repeatable tasks to larger, AI-adjacent projects.
How much of a client’s payment does Fiverr keep compared with Upwork?
Fiverr’s marketplace take rate was 28.0% for the twelve months to June 30, 2026, per its SEC filing. Axis Intelligence Research calculates Upwork’s Freelance Platform Capture Rate at 19.83% for Q2 2026. Both include fees charged to clients and freelancers combined, so neither is the commission a single freelancer sees on one contract.
Why does Upwork have fewer clients if it is making more per client?
Upwork reported 763,000 active clients on June 30, 2026, down from 796,000, while GSV per active client reached a record $5,230. The company says lower-complexity work is shifting to automation while larger engagements and its small-business Business Plus tier grow. Fewer, larger clients is the result.
Which freelance skills are growing in 2026?
Applied AI work. Upwork’s AI-related GSV grew more than 22% and its AI Strategy & Consulting category more than 50% in Q2 2026, and AI development reached 7.4% of Freelancer.com’s Q2 volume. MBO Partners found 54% of enterprises expect AI to increase their use of independent talent.
How much does the average freelancer make in the U.S.?
The Census Bureau’s 2024 tax-record data shows the average U.S. business with no employees grossed $58,194, but 60.35% grossed under $25,000 and only 13.31% reached $100,000. By profession, average receipts ranged from $19,289 in document preparation to $88,270 in legal services. Receipts are revenue before expenses, not take-home pay.
Do experienced freelancers lose less work to AI?
Not in the first evidence. Hui, Reshef and Zhou found that on Upwork, AI-exposed freelancers lost about 2% of contracts and 5% of earnings, with the largest declines among experienced, higher-rated freelancers. Demirci, Hannane and Zhu found the surviving automation-prone jobs were more complex and paid more.
How many people freelance worldwide?
The World Bank estimates 154 million to 435 million online gig workers globally, including both online freelancers and microworkers, or up to 12% of the global labour force. The wide range reflects different counting methods and whether marginal participants are included.
Is the number of freelancers falling in Europe?
Yes, for two years running. Eurostat counted 17.68 million self-employed people without employees in the EU-27 in 2025, down 0.72% from 2024 and 1.3% from the 2023 high, while total employment rose 0.47%. In the Netherlands, De Nederlandsche Bank reported about 131,000 fewer self-employed in mid-2026 than at the end-2024 peak.
Which European country has the most freelancers?
By share of employment, Greece leads with 17.16% of workers aged 15 to 64 self-employed without employees in 2025, followed by Poland (14.00%) and Czechia (12.82%), per Eurostat. Germany has the lowest share in the EU at 3.48%. By headcount, Italy has the most, with 2.92 million.
Why did Freelancer.com’s volume fall 20%?
Freelancer Limited named three causes for its 20.3% first-half 2026 marketplace GMV decline: two-factor login friction that drove returning-customer churn, a payment-gateway migration that reduced acceptance rates, and search de-indexation after rate-limiting aggressive AI scrapers. In constant currency, the decline was 11.7%.