You searched “print on demand with AI designs” and got 30 posts telling you to generate 100 shirt designs this weekend and wait for the money. Here’s the part they leave out. Listing is cheap. Selling is not. This post gives you the real per-listing cost, the real per-sale payout, and the hit rate you’d need to clear anything at all.

Honest baseline first. Published 2026 print-on-demand breakdowns put sellers with 25 to 100 active designs at roughly $100 to $600 a month, and sellers with 100 to 500 designs at roughly $400 to $2,500 a month. Those are industry estimates, not audited platform data, so treat them as a shape rather than a promise. The shape is the useful part: this is a volume business, and small catalogs make small money.

A listing costs about $1 a year, not 20 cents

Etsy charges a flat $0.20 to list an item. That listing stays live for 4 months, then you renew it for another $0.20. Three cycles gets you through a year, so a design that never sells costs you about $0.60 in listing fees. Add the auto-renew that fires each time an item does sell and you land at roughly $1 per listing per year in platform costs.

That’s the number people quote to prove print-on-demand is free money. It isn’t wrong. It’s just the smallest cost in the stack. The real cost is the hours you spend on design, mockups, titles, and tags, and those never show up on an invoice.

What one sale actually pays

On Etsy, a $24.99 t-shirt nets roughly $7 to $10 after production, shipping, and Etsy’s cut. On Amazon Merch on Demand, a $19.99 shirt pays roughly $5 to $7 in royalty. Marketplace sellers generally land in the 20% to 35% margin range once fees and production come out. Your own storefront moves that to 50% to 70%, because you keep the marketplace’s cut and pay for traffic instead.

Run that against the $1 listing cost and the picture is clear. One sale pays for 7 to 10 listings for a whole year. Your problem is never the fee. Your problem is the hit rate.

The hit rate is the whole business

Here’s the arithmetic nobody puts in the thumbnail. Say you want $500 a month on Etsy at $8 net per shirt. That’s 63 sales a month, or about 2 a day. If 1 in 20 of your listings sells once in a given month, you need roughly 1,250 live listings to get there.

You can move that number 2 ways. Raise the hit rate with tighter niches, or raise the price with better products. Most beginners do neither and just upload more. That’s why 500 designs and $0 is a common outcome, not a rare one.

Etsy’s 2026 rules changed what “AI designs” means

Etsy still allows AI-assisted work, but its 2026 Creativity Standards require real creative input from you, plus disclosure of AI use where it applies. Using a print-on-demand partner to produce the shirt is fine. Uploading raw generator output and calling it a shop is the part that gets pulled.

Practically, that means AI is your drafting tool, not your product. Generate 20 directions, pick 1, rework the composition and typography yourself, and keep a note of what you changed. That note is the difference between a catalog and a takedown.

Amazon Merch makes you earn the upload slots

Merch on Demand starts everyone at Tier 10, which means 10 live designs. You climb by selling, and each tier raises your upload and publish limits. Amazon also moved to a performance-based royalty structure in 2026, where what you earn per sale depends partly on where the traffic came from.

Translated: you can’t brute-force Amazon on day 1, and your first 10 designs matter more than your next 100. Pick 10 you’d actually bet on.

The 90-day version that isn’t a scam

1. Weeks 1 to 2: pick 1 niche you already know. Not “funny shirts”. Something like “pickleball for players over 60”.
2. Weeks 3 to 6: build 40 listings in that niche, AI-drafted and hand-finished. Etsy first, because you control the listings from day 1.
3. Weeks 7 to 10: find the 5 designs that got views, and make 20 more in the direction those 5 point.
4. Weeks 11 to 13: only now add Amazon Merch, using your 10 best-proven designs to fill the starting tier.

That’s 60 designs in 90 days, not 600. You’re buying information about what sells, and information is cheaper in small batches.

Which means for you

Print-on-demand isn’t passive and it isn’t free. It’s a $1-per-listing, $8-per-sale business where the only variable that really moves is how many of your listings people actually want. So don’t open with 200 designs and hope. Open with 40 in 1 niche you understand, and let the view counts tell you what to make next.

This week’s action: write down 1 niche you know better than most people do, and list 5 designs in it. Not 50. Five. You’re buying data, not inventory.

---

Want the starting point? Grab the free AI Side Hustle Starter Kit to pick your lane in an evening. When you want the full build, The AI Side Hustle Playbook ($27) walks the whole thing step by step. And if you want the no-code tooling behind it, The NoCode Claude Playbook ($19) covers the exact button clicks.

Income figures are published third-party ranges, not guarantees. Your results depend on hours and follow-through.