The Demand Inversion: 56.9% of Fab Sells One Thing, Buyers Are Asking About Another

We counted the whole marketplace instead of guessing

Ask ten Unreal developers what sells on Fab and you'll get ten confident, contradictory answers. The marketplace is large, opaque, and mostly discussed through anecdote and gut feel. So rather than add another opinion, we went and counted the whole thing: 80,082 live listings, 9,877 distinct sellers, 14 categories, every review and every public buyer question attached to them.

What came back was not a close call. It was an inversion. The category where sellers concentrate their effort is almost exactly the category where buyers concentrate the least. Sit with that for a moment, because most of the marketplace advice you've ever read points you in precisely the wrong direction.

One category is more than half the entire store

A single category — game-ready 3D models — accounts for 45,589 listings on its own. That is 56.9% of the entire catalogue. More than half of everything on Fab is a mesh.

If you ship another model, you are walking onto the single most crowded shelf in the building, competing on price and luck against tens of thousands of near-identical neighbours. And the surrounding numbers do not make that shelf any kinder. The median paid listing on Fab goes for $19.99. Only 3.1% of the catalogue is free, so price is doing real work as a signal — and on a shelf of 45,589 listings, the one lever you can reliably pull is to undercut.

This is the part of the picture sellers already feel in their gut. The crowding is obvious. What is not obvious is where everyone else went.

The buyers went somewhere else entirely

Here is the move that changes everything. Instead of asking what sells, we measured what buyers ask about — the public questions posted on listings before purchase. There are 183,238 of them across the catalogue, an average of 2.3 per listing. Divide a category's questions by its listings and you get demand intensity: how much pre-purchase attention the average listing in that category attracts.

Ranked by that metric, the marketplace turns upside down. The full table is below, but look at the two ends of it. Game Templates draws 22.8 questions per listing from a category of just 667 listings. 3D Models draws 1.3 from a category of 45,589. That is about 17 times more buyer attention per listing in the smallest serious category than in the largest one.

The big shelf is not where the buyers are. It is just where the sellers are.

Why we trust questions over reviews and over star ratings

You might reasonably ask why we lean on questions rather than the obvious signals — reviews and ratings. Two reasons, and both matter.

First, reviews are a biased sample. A review comes from the small, self-selected group who bought the asset and then chose to write something afterwards. A question comes from a far larger crowd weighing a purchase before paying. One measures a sliver of past buyers; the other measures live, unfiltered intent. If you want to know where demand is heading, you want the second crowd.

Second, the conventional signals are barely signals at all. 68.4% of live listings have zero reviews — for more than two-thirds of the store, there is no review data to read. And where ratings do exist, they're useless for discrimination: the average rating across reviewed listings is 4.46, with almost everything clustered just below five stars. A metric where nearly everything scores the same tells you nothing about what to build. Questions, by contrast, spread across two full orders of magnitude — from 0.2 to 22.8 — which is exactly what a useful signal looks like.

The gap is the opportunity

That gap — between where sellers keep building and where buyers actually look — is the single most useful finding in this entire report. It is, in plain terms, a map of where the marketplace is under-served.

The top of the demand table is not random. Game Templates, Tools & Plugins, and Game Systems are the three categories where buyers ask the most per listing, and all three share a trait: they save the buyer time and engineering. A template ships a working slice of a game. A plugin solves a problem the buyer would otherwise have to build. These are not decorations; they are shortcuts, and shortcuts are what a developer with a deadline pays for and asks careful questions about first.

Meanwhile the supply keeps pouring into models, where the average listing has drawn barely one question across its whole life — 1.3 of them — while competing against tens of thousands of others on price.

If you sell on Fab, the strategic read is uncomfortable but simple. The crowded shelf feels safe because it's where everyone is. The thin shelves feel risky because they're empty. The data says you have it backwards. An empty shelf with buyers standing in front of it is not a problem. That is an opening.

None of this means a great model can't sell. It means the average model is fighting the hardest fight on the floor, while the average template or plugin is being actively sought out. Build where demand is dense, not where supply is.

The fine print

A few honest caveats, because the conclusion is only as good as its coverage. The supply figures — listing counts, category sizes, prices — are current to May 2026. The buyer-question and review figures are complete through September 2024 and held for reference: the demand signal is frozen at that date, so treat the intensity ranking as a strong directional read rather than a live feed. Categories with very few listings (Environments sits at just 24) should be read with appropriate caution; their per-listing numbers move easily.

Every figure here was recomputed from our own full crawl of the live Unreal asset catalogue — no third-party estimates, no extrapolation, no rounding to flatter the story. The full category-by-category breakdown lives in our State of the Fab Marketplace report.

Demand intensity — buyer questions per listing

Buyer questions per listing by category — higher means more buyers asking per unit of supply. Demand figures complete through September 2024.

FAQ