The State of Unreal Asset Selling 2026

The almanac, and how to read it

Welcome to the first annual *State of Unreal Asset Selling*. This is the almanac we wish we'd had when we started selling on Fab: not a hype reel and not a leaderboard, but a sober map of the marketplace you are about to spend a year of your life building for. Every figure here comes from the MythicLemon Marketplace Index, our own compiled dataset, and every figure is stamped with the date it is honest as of. I am Phil; I run MythicLemon, and we sell on this same marketplace, so read everything that follows as a working seller talking to other working sellers rather than an analyst with no skin in the game.

Two stamps govern the whole document, and they matter enough to put up front. Supply, pricing, engine-version and category-mix figures are honest as of May 2026. Buyer-demand figures — questions, reviews and ratings — are frozen at September 2024 and are shown unchanged for reference. We are deliberately careful never to dress the frozen numbers up as current, because the single biggest mistake you can make reading a marketplace report is to confuse *we stopped measuring* with *the market went quiet*. More on that when we reach the asset graveyard.

If you only take one thing from the next ten sections, make it this: the marketplace's supply and its demand point in opposite directions, and the gap between them is the most valuable thing on the board. The rest of this report is an attempt to draw that gap precisely enough that you can aim a build at it.

Catalogue size and growth

Fab is large and still getting larger. As of May 2026 the catalogue holds 79,792 listings, and 1,754 of those were added in May alone. A marketplace adding the better part of two thousand listings a month is not a market that is winding down — it is a market where the scarce resource is not assets, it is *attention*. Buyers are not short of things to buy. You are short of ways to be seen.

That reframes the whole problem of what to build. In a thin marketplace, almost anything competent finds a buyer. In one this size, the question is never just "is my asset good" but "is my asset good *and* aimed somewhere the crowd is not already three deep." Growth this steady is also why we refuse to extrapolate the frozen demand numbers: a catalogue that adds ~1,754 listings a month has changed shape enormously since our buyer-activity snapshot closed in September 2024, and pretending otherwise would be the most dangerous kind of false precision.

The one-category problem

Here is the defining structural fact of the marketplace: 3D models make up 57.1% of the entire catalogue as of May 2026 — more than every other category combined. There is no second category that comes close. When sellers ask "what should I make," the revealed answer of the crowd has been, overwhelmingly, *another 3D model*.

This is the one-category problem. It is not that 3D models are bad to sell — it is that more than half the entire marketplace is competing in the same lane, which means that lane is where being merely good is least likely to be enough. The interesting question is whether buyer demand follows supply into that lane. It does not, and the divergence is so stark that we have given it its own section. For now, simply hold the number: 57.1%. Almost everything else in this report is a reaction to it.

The $19.99 anchor

Pricing on Fab is remarkably concentrated. The median price is $19.99, and so is the modal price — the single most common price point — with heavy clustering on .99 endings throughout the catalogue. A small slice of the marketplace is free: 3.1% of listings carry no price. The picture is of a market that has quietly agreed on a psychological anchor and crowds around it.

For a seller, the anchor cuts two ways. It makes pricing easy — list near $19.99 with a .99 ending and you are speaking the marketplace's native dialect — but it also means price is a weak lever for standing out, because nearly everyone is pulling it to the same spot. A blunt warning while we are on the subject of money: price is not revenue, and this report contains no unit-sales or revenue figures. We do not know what anything earns, and neither does anyone quoting list prices at you. A $19.99 median tells you what sellers *ask*, not what buyers *pay across volume*. Anyone converting that median into an income projection is inventing a number, and we will not be the ones to do it.

The UE5 takeover

If the catalogue has one clear direction of travel, it is the engine version. As of May 2026, UE5 accounts for 93.2% of new monthly releases — new work is overwhelmingly being made for UE5, and that share is the clearest forward-looking signal in the whole dataset. New content is a UE5 story now, full stop.

The *installed* catalogue tells a more nuanced tale. Of the version-tagged listings, UE5 is 53.3% and UE4 is still 46.7% — almost half the tagged catalogue is UE4. That 46.7% is not a sign of a thriving UE4 market; it is a modernisation backlog, a large body of older listings that have not been brought forward. For a seller, the two numbers together are a strategy: build new work on UE5 to ride the 93.2% tailwind, and treat the UE4 backlog as a clue that there is room for well-made, modernised versions of things buyers still want. (We are describing the marketplace's UE5 majority, not asserting maturity dates for individual engine subsystems — for that, see our companion Engine Feature Radar, which is anchored to Epic's own release notes.)

The seller economy

Who are you competing with? As of May 2026 there are 9,869 active sellers, with 142 new ones arriving in May. The marketplace is concentrated at the top and enormously fragmented at the bottom. The top ten sellers alone account for 10.0% of all listings — ten accounts holding a tenth of the entire catalogue — while at the other end, between 42.8% and 43% of sellers have published exactly one listing.

That shape is the most encouraging chart in the report for a newcomer, and the most sobering for anyone hoping to coast. Encouraging, because a marketplace where four in ten sellers are one-and-done is not one where a small number of incumbents have locked everything up — there is constant churn and constant room at the bottom of the table to climb. Sobering, because the one-listing-and-stop pattern is what *failure looks like here*: it is the seller who shipped once, didn't see the traction they hoped for, and walked away. The sellers who climb toward that top-10 band are, almost without exception, the ones who treat selling as a catalogue to be built over years rather than a single lottery ticket.

The demand inversion

This is the centre of the report, and the finding everything else orbits. We measure demand with the cleanest proxy we have: buyer questions per listing — how often, on average, a listing in a category provokes a buyer to actually speak up. (These figures are frozen at September 2024; treat them as the shape of demand, not its live level.) When you rank categories that way, the marketplace nearly turns upside down.

Game Templates are the most asked-about at roughly 43 buyer questions per listing. Audio is the least at about 0.3. 3D models — 57.1% of all supply — draw only about 2.5. That means buyers interrogate a game template around 17 times more than they do a 3D model. The category sellers have crowded into hardest is among the quietest in buyer demand, and the thin, hard-to-build categories — templates and tools — are the loudest. We call this the demand inversion: supply and demand point in opposite directions, and the gap between them is where an under-served buyer is waiting.

The honest reading of this is not "abandon 3D models" — it is half the marketplace for a reason, and there is real money in being excellent there. It is: if you are choosing where to point new effort, weight toward where buyers are dense rather than where sellers are. A competent template or tool lands in a far less crowded room with a far more engaged audience than a competent prop dropped into the 57.1% scrum.

The asset graveyard — and why freeze is not decay

Two more frozen-demand findings round out the picture, and they come with the single most important caveat in this report. First: 83% of all reviews are 5-star. Ratings barely discriminate — when nearly everything is five stars, a five-star rating tells a buyer almost nothing, which is part of why the questions-per-listing proxy is a better demand signal than star ratings. Second, the graveyard: 70% of the 2023 launch cohort have zero reviews, up from 43% for the 2015 cohort. The longer the marketplace has run, the larger the share of listings that arrive and are never reviewed at all.

Now the caveat, and please read it twice. Our buyer-activity data is frozen at September 2024. Roughly 45% of the catalogue was released *after* that date, and it has zero questions and zero reviews not because it failed, but because the data froze before buyers could weigh in. This is the freeze-is-not-decay trap, and it is the easiest way to misread this entire marketplace. A category that is young, or quiet, or full of recent releases is not therefore "saturated," "declining" or "dead" — it may simply sit beyond the edge of our demand snapshot. We never rate a young or quiet category as crowded on the strength of frozen numbers; the most we will say is that it is *uncertain* and *one to watch*. The graveyard is real for old cohorts; for recent ones, absence of reviews is absence of *data*, not absence of *life*.

The cross-marketplace premium

Fab does not exist in isolation, and one cross-marketplace pattern is worth a seller's attention. Across 2,260 assets that are cross-listed on two stores, the Unreal listing is typically about 150% more expensive than the very same asset elsewhere. Whatever the reasons — engine-specific integration work, a buyer base with deeper budgets, or simply what the market will bear — the Unreal-side listing commands a meaningful premium for identical content.

Read carefully, this is a pricing-power signal, not a revenue figure. It says the Unreal buyer tolerates a higher *asking price* for the same thing; it does not say anything about how many copies move, and we will not pretend it does. The same dataset also shows the broader 3D-asset publishing world cooling — past its peak — which makes the Unreal premium more notable, not less: the surrounding adjacent market is softening while the Unreal listings hold their higher price. For a seller weighing where to put a cross-listable asset, that premium is a genuine argument for treating the Unreal listing as first-class rather than an afterthought.

A note on the categories we sell in

Honesty obliges a disclosure. MythicLemon sells on this marketplace, and several of the categories this report discusses are ones we publish into — VFX and Niagara packs, audio packs, content packs, and developer tools. So when we point you toward thinly-supplied, highly-asked-about categories, understand that we have, in places, taken our own advice and have a commercial interest in those lanes being healthy.

We have tried to keep the analysis honest in spite of that, and the way we do it is by never naming or rating a single competing product anywhere in this report — the figures are about *categories and the marketplace as a whole*, never about an individual rival listing. One category-level data point we can share without conflict, because it is about the shape of VFX supply rather than any product: VFX packs are a small slice of the catalogue (roughly 4.6%), they cluster around the same ~$19.98 anchor as everything else, and they draw mid-table buyer demand at about 4.9 questions per listing on the frozen September-2024 numbers — denser than 3D models, far quieter than templates. Use that the way you'd use any other line on the demand map: as one more reading of where the room is less crowded.

Materials, audio, and a shifting mix

Underneath the headline 57.1% one-category number, the *composition* of new releases moves year to year, and one recent shift is worth flagging because it shows the marketplace is not frozen in shape even where it looks dominated. As of May 2026, materials have recently overtaken audio in annual new-release volume — sellers are minting new material packs faster than new audio packs, a reordering of the mid-table categories that was not true a couple of years ago.

On its own this is a small fact, but combined with the demand inversion it is a useful corrective: the relative effort sellers pour into a category drifts over time, which means today's crowded lane is not guaranteed to be tomorrow's, and a quiet lane can fill up. It is one more argument for reading the marketplace as a moving map rather than a fixed leaderboard — and for re-checking the mix before you commit a year to any single category.

What this means if you build

Pull the threads together and a strategy falls out — stated at the level of categories and practices, never of products, because that is the only level at which honest advice is possible. First, do not aim a year of work into the 57.1% scrum by default. 3D models are half the marketplace and the hardest place to be merely good; enter that lane only if you have a genuine angle, not because it is the obvious thing to make.

Second, follow the demand inversion. Buyers ask roughly 17 times more about game templates than 3D models, and tools and templates sit thin in supply and loud in demand. If you can build in those lanes, you land in a less crowded room in front of a more engaged audience. Third, build on UE5 — it is 93.2% of new releases — and treat the ~46.7% UE4 backlog as a hint about modernised versions of things buyers still want. Fourth, price into the $19.99/.99 anchor rather than fighting it, remembering always that price is an asking-figure, not revenue.

Fifth, and most important, play the long game. The single clearest dividing line in the seller economy is between the 43% who ship once and stop and the handful who build a catalogue over years toward that top-10 band. And sixth, mistrust quiet. Roughly 45% of the catalogue is too new to appear in our frozen September-2024 demand data; a young or silent category is uncertain, not dead, and should never be written off as saturated on numbers that simply stopped being collected. Build where buyers are dense, on the engine they are moving to, at the price they expect, with the patience to keep going — and re-read this almanac next year, because the map will have moved.

• ✓Don't default into the 57.1% 3D-model scrum — enter only with a real angle.
• ✓Follow the demand inversion: weight toward templates/tools (~43 q/listing) over 3D models (~2.5).
• ✓Build new work on UE5 (93.2% of new releases); read the ~46.7% UE4 backlog as a modernisation opening.
• ✓Price into the $19.99/.99 anchor; never mistake list price for revenue.
• ✓Build a catalogue over years — avoid the 43% one-and-done failure pattern.
• ✓Treat young/quiet categories as uncertain, not saturated — remember the demand data froze in Sep 2024.

The marketplace at a glance

Supply, pricing, engine and seller figures as of 2026-05. Demand figures (questions/listing) frozen at 2024-09 and shown unchanged for reference — not current, never extrapolated.

FAQ