Where the catalog actually sits
Twenty-nine markets, 6,441 live products, and a distribution far from even. Ten markets hold 52.6% of everything built — and they are, almost exactly, the markets with the least room left.
16 September 2026 · 8 min read · Data frozen 16 September 2026
Twenty-nine markets, 6,441 live products, and a distribution that is far from even. Ten markets hold 52.6% of everything built. This report sets out where the catalog actually sits, and then does the thing that makes the distribution worth reading: puts it next to how much room each of those markets has left.
The short version
- The five largest markets hold 29% of live products; the ten largest hold 52.6%.
- Coding is the largest at 422 products — 6.6% of the live catalog on its own.
- The ten largest markets average 18.7 points of room. The ten smallest average 58.4. Building effort is concentrated almost exactly where it is least needed.
- Market size and failure rate are unrelated. The deadliest markets in the catalog are among the smallest.
What is counted
Every figure here counts live products — 6,441 of 7,030 tracked, with shut-down and acquired products excluded except where mortality is being measured explicitly. A product sits in exactly one market, so shares sum to 100%.
Finding one: a long, flat head and a short tail
6,441 live products across 29 markets. Falcoscan catalog, 16 September 2026.
The shape is less extreme than the geography data. There is no single dominant market: Coding leads at 422 products, but eight markets sit within a hundred products of it. The concentration is in the aggregate rather than in any one category — ten of twenty-nine markets account for 52.6% of everything.
| Rank | Market | Products | Cumulative share |
|---|---|---|---|
| 1 | Coding | 422 | 6.6% |
| 2 | Image | 392 | 12.6% |
| 3 | Design | 357 | 18.2% |
| 4 | Writing | 351 | 23.6% |
| 5 | Agents | 344 | 29% |
| 6 | AI Models | 313 | 33.8% |
| 7 | Data | 309 | 38.6% |
| 8 | Ecommerce | 304 | 43.3% |
| 9 | Productivity | 303 | 48.1% |
| 10 | Video | 296 | 52.6% |
| 11 | Finance | 281 | 57% |
| 12 | Marketing | 280 | 61.4% |
| 13 | Sales | 276 | 65.6% |
| 14 | Voice | 272 | 69.9% |
| 15 | Social | 266 | 74% |
| 16 | Learning | 258 | 78% |
| 17 | Security | 257 | 82% |
| 18 | Support | 230 | 85.6% |
| 19 | Healthcare | 137 | 87.7% |
| 20 | Automation | 125 | 89.6% |
| 21 | Human Resources | 96 | 91.1% |
| 22 | Research | 87 | 92.5% |
| 23 | Legal | 83 | 93.8% |
| 24 | Real Estate | 82 | 95% |
| 25 | 3D & AR/VR | 68 | 96.1% |
| 26 | Search & SEO | 68 | 97.1% |
| 27 | Avatars | 66 | 98.2% |
| 28 | Music | 61 | 99.1% |
| 29 | Gaming | 57 | 100% |
Markets ordered by size, showing the running percentage of live products covered. Falcoscan catalog, 16 September 2026.
Finding two: effort is concentrated where room is not
This is the finding that makes the rest of the table matter. Take the ten largest markets and average their opportunity-to-saturation gap: 18.7 points. Do the same for the ten smallest: 58.4.
| Market | Live products | Share | Opportunity | Saturation | Room |
|---|---|---|---|---|---|
| Coding | 422 | 6.6% | 60 | 44 | +16 |
| Image | 392 | 6.1% | 49 | 52 | -3 |
| Design | 357 | 5.5% | 53 | 41 | +12 |
| Writing | 351 | 5.4% | 40 | 64 | -24 |
| Agents | 344 | 5.3% | 73 | 31 | +42 |
| AI Models | 313 | 4.9% | 73 | 30 | +43 |
| Data | 309 | 4.8% | 68 | 33 | +35 |
| Ecommerce | 304 | 4.7% | 65 | 36 | +29 |
| Productivity | 303 | 4.7% | 63 | 46 | +17 |
| Video | 296 | 4.6% | 63 | 43 | +20 |
| Finance | 281 | 4.4% | 67 | 35 | +32 |
| Marketing | 280 | 4.3% | 58 | 49 | +9 |
| Sales | 276 | 4.3% | 64 | 44 | +20 |
| Voice | 272 | 4.2% | 66 | 41 | +25 |
| Social | 266 | 4.1% | 54 | 56 | -2 |
| Learning | 258 | 4% | 63 | 42 | +21 |
| Security | 257 | 4% | 68 | 33 | +35 |
| Support | 230 | 3.6% | 67 | 39 | +28 |
| Healthcare | 137 | 2.1% | 82 | 11 | +71 |
| Automation | 125 | 1.9% | 72 | 23 | +49 |
| Human Resources | 96 | 1.5% | 73 | 19 | +54 |
| Research | 87 | 1.4% | 76 | 14 | +62 |
| Legal | 83 | 1.3% | 80 | 11 | +69 |
| Real Estate | 82 | 1.3% | 76 | 13 | +63 |
| 3D & AR/VR | 68 | 1.1% | 77 | 13 | +64 |
| Search & SEO | 68 | 1.1% | 69 | 24 | +45 |
| Avatars | 66 | 1% | 75 | 15 | +60 |
| Music | 61 | 0.9% | 73 | 19 | +54 |
| Gaming | 57 | 0.9% | 78 | 14 | +64 |
Every market, ordered by size, with its room (opportunity minus saturation) alongside. Falcoscan catalog, 16 September 2026.
Read the two right-hand columns together and the pattern is unmistakable. The markets where thousands of people are building have almost no gap left. The markets with the widest gaps — Healthcare at 137 products, Legal at 83, Gaming at 57 — are among the smallest in the catalog.
The mechanism is the one our wrapper-risk report describes: the large markets are large because they were cheap to enter, and the small ones are small because they were not. Concentration here is not a signal of where value is. It is a map of where the barriers were lowest.
Finding three: size says nothing about survival
Share of each market's tracked products that no longer resolve, ordered by failure rate. Falcoscan liveness sweep, 16 September 2026.
The worst failure rates belong to Gaming (85 tracked products), Avatars (88) and Music (80) — three of the smallest markets we track. The largest markets sit in the safe half of the table. Whatever drives products to die, market size is not it; per our mortality report, it is who the buyer is.
That combination — small markets being both the roomiest and the deadliest — is the single most important caveat on the concentration story. An open market is not a safe market. It is an untested one.
What to do with it
For founders. If you are choosing a market by how much activity you can see in it, you are choosing by a lagging indicator of where entry was easy. The cumulative-share table is a list of places where you will be the four-hundredth entrant.
For investors. Deal flow volume will track this distribution almost exactly, because supply follows the same low barriers. If your pipeline mirrors the top of this table, it is describing the shape of the market rather than your selection.
For anyone reading a market map. Count-based charts of the AI landscape — including ours — measure past entry, not present opportunity. The gap column is the one that answers the question people usually think the size column is answering.
Photo: Arndt-Peter Bergfeld / Pexels. Colour-graded for Falcoscan.
Citing these numbers
Every figure here is from the Falcoscan catalog as it stood on 16 September 2026, and is frozen at that reading. Later changes to the catalog will not alter this page, so a number you quote today will still say the same thing when someone checks it.
Falcoscan, “Where the catalog actually sits”, 16 September 2026. https://falcoscan.com/reports/where-the-catalog-actually-sits