Online Tutoring in 2026: The Side Hustle AI Was Supposed to Kill (It Didn't)

In 2023, everyone told me tutoring was dead. ChatGPT could explain calculus for free, so why would anyone pay a human $50 an hour?

Three years later, tutoring marketplaces are busier than ever and rates went up. The doomsayers misunderstood what parents were actually buying.

Here's the thing AI can't do: make a 16-year-old actually sit down on Tuesday at 5pm and work through practice problems. ChatGPT gives infinite explanations. It gives zero accountability. Parents figured this out fast, and they're paying for the accountability.

The whole business fits on one desk. That's kind of the point.

What tutors actually charge, by subject

Rates aren't random. They track two things: how scarce the skill is, and how much the outcome is worth to the person paying. A parent staring down college admissions will pay more for SAT prep than for third-grade reading help, even if the reading help is arguably more important.

Those are marketplace rates. Independent tutors in wealthy metro areas charge more, sometimes a lot more. But start on a platform, because a platform's job is to hand you students you didn't have to find.

Picking a platform (and what they take)

Wyzant is the biggest US marketplace and takes roughly a quarter of what you earn. You set your own rate, so build the commission into it. Preply starts new tutors at a steep 33% commission that drops as you log hours, which stings early but the student flow is real. Cambly pays around $10 to $12 an hour for casual English conversation, which I'd only recommend as a zero-prep way to test whether you enjoy teaching at all. Tutor.com hires you like a regular part-time job with set-ish rates, less upside, more predictability.

My honest take: treat platforms as a customer acquisition channel, not a career. Once you've got 5 to 10 regular students and a stack of five-star reviews, new students start coming from referrals, and referrals don't pay commission. Just read each platform's rules first, since some prohibit taking students off-platform for a period.

How to actually start this month

Week one: pick the one subject where you're strongest, write a profile that names specific outcomes ("raised my last student's SAT math score 120 points" beats "passionate about education"), and record a 90-second intro video. Profiles with video get dramatically more contact requests on every platform that supports them.

Week two: price yourself 10 to 20 percent below the going rate for your subject, just until you have three reviews. Then raise it. Underpricing forever is the most common tutor mistake, and it signals low quality to exactly the parents who'd pay the most. If you want to sanity-check whether the hours are worth it against other options, run your numbers through our side hustle calculator.

And use AI as your prep assistant. I generate practice problem sets, lesson outlines, and parent progress emails with it, which cuts my unpaid prep time roughly in half. The teaching hour stays human. The busywork doesn't have to.

The honest downsides nobody puts in their course

Tutoring income is lumpy and seasonal. Test prep booms from January through May, then summer arrives and half your students vanish to camp. Evenings and weekends are your prime hours forever, because that's when students are free. Cancellations will drive you insane until you adopt a written 24-hour cancellation policy, so do that in week one, not month six.

Also, this doesn't scale by itself. You're selling hours. Getting past that means group sessions, raising rates, or productizing your material, and honestly, selling study guides or Notion templates built from your lesson plans is the most natural upgrade path I've seen tutors pull off.

Frequently asked questions

Tutoring survived the AI wave because it was never really about information delivery. It's about a human showing up for another human, week after week. If you know something well enough to explain it patiently, someone will pay you for that in 2026. Probably more than you think.