How many Roblox games actually make money?

Roblox does not publish per-game revenue. Nobody outside a studio can tell you what a specific game earns — not us, and not the calculators that imply otherwise.

What can be done is estimate every game the same way, state the assumptions, and look at the shape of the result. That is what this is: 3,959 Roblox games with a measurable daily visit rate, each run through the same model, measured in the week to 19 August 2026 according to BloxBunny's hourly sampling of the Roblox public API.

How much does the average Roblox game make?

The median game is estimated at $948 a month.

| Estimated developer earnings | Games | Share | |---|---|---| | $0 – $99 | 1,025 | 25.9% | | $100 – $999 | 978 | 24.7% | | $1,000 – $9,999 | 1,161 | 29.3% | | $10,000 – $99,999 | 689 | 17.4% | | $100,000 or more | 106 | 2.7% |

Half of these games are estimated below $948 a month. A quarter are below $100 — effectively nothing, after tax and any time counted as cost.

That median is higher than most people expect, and the reason matters: this is not a sample of all Roblox games. It is a sample of games Roblox actually surfaces in its own discovery sorts. Roblox hosts millions of experiences; the overwhelming majority never appear anywhere a player would find them, and they are not in this dataset. If they were, the median would be approximately zero. Read every number here as "among games that already have an audience".

Who actually takes the money?

Concentration is the real story, and it is severe:

• The top 1% of games take an estimated 38.1% of all earnings in this dataset.
• The top 10% take 80.5%.

The gap between the middle and the top is not incremental. The median game is at $948 a month. The 90th percentile is at $26,674. The 99th is at $219,342. Moving from typical to top-decile is not a 2× improvement — it is roughly 28×.

How many Roblox developers make a living?

Depends where you set the bar, and it is worth being conservative. Our model produces a range per game rather than a single figure, because the two quantities that decide the answer — what share of visitors spend anything, and how much — cannot be observed from outside. Counting only games that clear each threshold at the pessimistic end of their range:

| At least | Games clearing it (conservative) | Using the midpoint | |---|---|---| | $100 / month | 2,644 | 2,934 | | $1,000 / month | 1,518 | 1,956 | | $10,000 / month | 453 | 795 | | $100,000 / month | 39 | 106 |

So: roughly 1,500 games on Roblox plausibly earn their developer $1,000 a month or more. Around 450 clear $10,000. Fewer than 50 clear $100,000.

For anyone sizing up Roblox as a career rather than a hobby, that middle row is the honest target, and there are a few hundred seats in it.

Why is this an estimate and not a measurement?

Because the inputs Roblox makes public stop short of revenue. The model works forward from what is observable:

1. Visits per day, measured — not the lifetime total. This is the part most calculators get wrong: they ask you to type in a game's all-time visit count, which cannot distinguish five years of accumulation from one good month.
2. Robux spent per visit, assumed. The base rate is a community heuristic, adjusted by genre, approval rating, favourites per thousand visits and session length.
3. Roblox's 30% marketplace cut, then conversion at the DevEx rate of $0.0038 per earned Robux.

Step 2 is judgement, not measurement, and it is where the uncertainty lives. Typical conversion is somewhere around 1–3% of visitors spending anything at all, and what they spend varies by an order of magnitude between games. No dataset available from outside a studio settles it.

Which is why every figure above is a distribution rather than a claim about any one game — the errors are large per game and largely cancel across four thousand of them.

How do I check a specific game?

Paste its Roblox link into the Roblox revenue calculator. It fills in the visit rate we actually measured for that game over the past week and shows every multiplier it applies, so you can disagree with the assumptions and change them.