Two numbers from this year's surveys tell the whole story of AI and personal income, and they need to be read together or not at all.
The first is enormous. Menlo Ventures, in a nationally representative survey of 5,067 US adults fielded by Morning Consult in July, found that 48% of AI users have earned money with AI. That works out to roughly 30% of all US adults, about 79 million people. Not "could earn". Have earned, within the past six months.
The second is small. A Florida Tech survey of 831 full-time US workers aged 22 to 28 asked the people actually doing it how much they make. Among active AI side hustlers, 51% earn between $1 and $499 a month, and another 19% make $500 to $999.
So the AI side hustle is simultaneously the most widespread new source of personal income in years and, for most people, worth a few hundred dollars a month. Both facts matter, and the space between them is where every misleading promise on the internet lives.
What the national survey actually found
| Finding | Figure |
|---|---|
| AI users who have earned money with AI | 48% |
| Share of all US adults that represents | ~30%, about 79 million people |
| Most common route: content creation | 22% of AI users |
| Second: freelancing | 18% |
| Gen Z AI users who have earned money | 62% |
| Boomer AI users who have earned money | 15% |
| Sample | 5,067 US adults, Morning Consult, July 2026, weighted to be nationally representative |
The generational split is the sharpest line in the data. Nearly two thirds of Gen Z AI users have made money with these tools against one in seven boomers. That is not mainly a skills gap, it is an exposure gap: younger workers already sell things online, already freelance, already post content, so adding AI to an existing money-making behaviour is a small step rather than a new career.
For context, Bankrate found that 27% of Americans had a side hustle of any kind in 2025, and that share was falling. Set that against 30% of adults earning something from AI and you get the real shape of this: AI has not created a new army of side hustlers so much as it has quietly become the tool the existing ones use, plus a layer of people earning occasional money who would never call it a side hustle at all.
What they actually earn
This is the part that almost never appears next to the big participation numbers.
| Monthly earnings from AI-assisted work | Share of active AI side hustlers |
| $1 to $499 | 51% |
| $500 to $999 | 19% |
| Above $999 | not reported in the published findings |
| Time invested: 1 to 5 hours per week | 53% |
Put the earnings next to the hours and the arithmetic becomes honest. Somebody making $300 a month on four hours a week is earning roughly $17 an hour. That is real money and it is worth having. It is not a replacement income, and nobody should quit anything over it.
A separate survey of 1,007 side hustlers by Network Solutions found AI users earning a median of $500 a month against $250 for those not using AI, with averages of $555 and $333. The authors were careful, and so should we be: people who reach for AI tools are probably also the people who treat their side work more seriously. The correlation is clear. The causation is not proven.
The gap between the median and the dream
Look at the general side-hustle numbers and you see the pattern that explains why AI income content is so misleading. Bankrate's 2026 survey puts the median side hustler at $200 a month and the average at $885.
That spread is the whole phenomenon. A median of $200 and an average of $885 means a small number of people earn a great deal and pull the mean far above what a typical person experiences. Every "I make $8,000 a month with AI" video is technically from inside that distribution. It is just from the thin end of it, presented as the middle.
| What you are being shown | What it usually is |
| "People are making $5,000 a month with AI" | the top of the distribution, as a recruiting pitch |
| "Average AI side hustler earns $555" | the mean, inflated by a few large earners |
| "Median is $200 to $500" | what a typical person actually experiences |
| "79 million Americans have earned money with AI" | true, and most of them made pocket money |
None of this argues against trying. It argues against quitting your job on the strength of someone else's best month, and against the gap between the two I keep coming back to in the AI wealth divide.
The one finding that should change what you use
Buried in the Florida Tech data is a comparison I have not seen anywhere else, and it is worth sitting with.
Among AI side hustlers, 45% of Claude users earn $500 or more a month, against 30% of ChatGPT users. ChatGPT was far more widely used (87% against 34%), so this is not about popularity.
Be careful with what that means. The most likely explanation is selection rather than magic: people who go out of their way to use a less default tool tend to be the more technical, more deliberate users, and those are the people who build something sellable rather than asking a chatbot for ideas. The tool is a marker of the user, not a cause of the income.
That said, it is a useful marker. If the people clearing $500 are disproportionately the ones who chose their tools on purpose, the lesson is to be the kind of person who chooses on purpose. Our most-read guide happens to be making money with Claude, and the honest companion to it is the ChatGPT version; what separates the earners in both is not the logo but whether they sell a defined service to someone with a budget.
The spending side mirrors this almost exactly. The median AI payer spends about $25 a month, while the top 1% spend around $903, mostly on automation and production tools that do paid work. What people actually pay for AI shows that the heavy spenders and the heavy earners are largely the same people.
Why most people stop at a few hundred dollars
Three things show up repeatedly in the data, and each one is fixable.
Hours. Most people put in one to five hours a week. At any realistic rate, that caps the outcome before skill enters the picture. The people earning thousands are not working four hours a week with a better prompt.
No named buyer. Content creation is the most common route at 22% of AI users, and it is also the one with the longest and least reliable path to money. Freelancing, at 18%, pays sooner because somebody specific agrees to pay for something specific.
Motivation set to "experiment". In the Gen Z data, 40% are doing it for a short-term income boost and 35% to experiment and learn, while only 11% are building a long-term career path. Those are reasonable goals, and they produce exactly the income you would expect from them.
The step from a few hundred to a few thousand is not a tool upgrade. It is the switch from making things and hoping, to selling a defined outcome to a named client, which is the organising idea behind the nine business models that actually pay and the reason the hardest and most valuable step is landing the first client.
One of the clearest examples of this shape is also one of the newest: people paid by the hour to film their own chores so robots can learn from them. It needs no skill, pays $13 to $25 an hour in the US, and works out at almost exactly the median once rejected footage is counted. The numbers are in what filming chores for robot training pays.
What this means for you
If you have not started, the participation numbers should be encouraging rather than intimidating. Thirty percent of US adults have already earned something, mostly in ordinary ways: writing, design, small freelance jobs. The entry bar is low and the realistic first result is a few hundred dollars a month, which is a perfectly good reason to begin. Side hustle ideas for beginners is the practical starting list.
If you are already earning a few hundred a month, you are the median, and the median is not a ceiling. The data says the constraint is usually hours and the absence of a paying client rather than ability. Pick one service, name a price, and find three people who need it. That single change moves more income than any amount of further tool learning.
If you are considering going full-time, read the distribution again before you resign. Half of 22 to 28 year old AI side hustlers earn under $500 a month, and 52% say they would leave their job if the income became significant, which means a lot of people are waiting for a number most of them have not reached. Build to a stable, boring, repeatable income first. The honest income ladder is in how much you can really make.
If you run a small business, the owner-side numbers are separate and more encouraging: a median of $2,200 a year spent and five hours a week of your own time back, with two-thirds reporting more revenue. The small business owner's AI math covers who actually turned that time into money.
If you employ people, note that a third of adults are now earning money outside work with the same tools you are deciding whether to license. Some of that is harmless. Some of it is your staff discovering that their AI skills have a market price that your salary structure does not reflect, which is the pay gap I covered in the 62% premium most employers never pay.
What to watch
Whether the median moves. Participation is clearly rising. The number that matters for ordinary people is whether the typical earner goes from $300 a month to $800, or whether more people simply join at the bottom. Those are very different economies.
Whether content creation keeps leading. It is the most common route and one of the least reliable. If freelancing and services overtake it, the aggregate income numbers will improve, because those have buyers attached.
Better data. Almost everything here is self-reported, and the participation surveys and the earnings surveys cover different populations. Nobody is running an audited study of AI side income. Until somebody does, treat every figure in this field, including the ones in this article, as directionally right and precisely wrong.
The honest take
What strikes me about these numbers is how undramatic the truth is. Seventy-nine million people have earned money with AI, and the typical experience is a few hundred dollars a month for a few hours a week. That is neither the revolution the hype sells nor the nothing the sceptics claim. It is a large number of people making modest, real money, which is how most technologies actually arrive in ordinary lives.
The useful conclusion is that the barrier is no longer access. Everyone has the tools, 79 million people have already made something with them, and the thing separating $300 from $3,000 is not a secret prompt. It is the ordinary, unglamorous business question of whether somebody with a budget has agreed to pay you for a specific thing, on a schedule.
So here is the question worth sitting with if you are one of the 79 million: do you have a customer, or do you have a tool?
Frequently asked questions
How many people actually make money with AI?
About 48% of US AI users have earned money with AI, roughly 30% of all US adults or around 79 million people, according to a nationally representative survey of 5,067 adults fielded by Morning Consult for Menlo Ventures in July 2026.
How much does the average person earn from an AI side hustle?
Among active AI side hustlers aged 22 to 28, 51% earn $1 to $499 a month and 19% earn $500 to $999. A separate survey of side hustlers put the median for AI users at $500 a month against $250 for non-users. General side hustle data shows a median of about $200 and an average of $885, a gap caused by a small number of high earners.
What is the most common way people earn money with AI?
Content creation, at 22% of AI users, followed by freelancing at 18%. Freelancing tends to pay sooner because a specific client agrees to pay for a specific deliverable.
Do people using Claude earn more than people using ChatGPT?
In the Florida Tech survey, 45% of Claude users cleared $500 a month against 30% of ChatGPT users. The likely explanation is selection rather than the tool itself: people who choose a non-default tool tend to be more deliberate users.
Can you replace a full-time salary with AI income?
Some people do, but they are the thin end of the distribution. Half of young AI side hustlers earn under $500 a month, and most spend one to five hours a week on it. Treat full replacement as a target to build toward with a real client base, not a realistic first-year outcome.
Sources: Menlo Ventures: The State of Consumer AI 2026; Florida Tech: How Gen Z makes money with AI, 2026 survey data; Black Enterprise on the Network Solutions side-hustle survey.
