What Should a Startup Research Before Entering an AI Market?
Eight checks to run before committing to an AI market, each tied to a number FalcoScan tracks for every market: competitors, crowding, room, wrappers, shut-downs, momentum, pricing and funding, with a current example of each.
25 September 2026 · 9 min read · Data updated 25 September 2026
What should you check before entering an AI market?
Based on FalcoScan data updated 25 September 2026, check eight things, each of which FalcoScan measures for all 29 AI markets it tracks: how many products already compete (from 60 to 435 active products per market), how crowded the market is (saturation), how much room is left (opportunity), how many incumbents are thin wrappers (wrapper share), how often products shut down (7.3% of everything tracked), momentum, how rivals price and who has raised money. No single check settles it: Gaming is one of the least crowded markets and loses the most products.
How do all 29 AI markets score on the checklist?
Every market FalcoScan tracks, against all eight checks. Momentum, pricing and funding are each reduced to one number here; the sections below explain each column, what counts as high, and what to do with it. Each market links to its page, where every product in it is listed with its own scores.
| Market | Active products | Saturation | Opportunity | Wrapper share | Shut-down rate | Hot | Paid only | Bootstrapped |
|---|---|---|---|---|---|---|---|---|
| 3D & AR/VR | 70 | 13 | 77 | 0% | 17.6% | 52% | 29% | 4% |
| Agents | 381 | 31 | 73 | 19% | 7.1% | 57% | 24% | 14% |
| AI Models | 402 | 30 | 74 | 1% | 8.6% | 53% | 45% | 7% |
| Automation | 324 | 23 | 73 | 0% | 3.9% | 21% | 78% | 3% |
| Avatars | 70 | 15 | 75 | 0% | 23.9% | 60% | 27% | 21% |
| Coding | 435 | 44 | 61 | 27% | 5.8% | 40% | 11% | 22% |
| Data | 377 | 33 | 69 | 3% | 3.8% | 39% | 36% | 8% |
| Design | 364 | 41 | 53 | 18% | 6.4% | 17% | 17% | 39% |
| Ecommerce | 315 | 36 | 65 | 12% | 5.1% | 22% | 47% | 6% |
| Finance | 379 | 35 | 68 | 2% | 5.0% | 24% | 69% | 1% |
| Gaming | 60 | 14 | 78 | 2% | 31.8% | 47% | 32% | 22% |
| Healthcare | 239 | 11 | 83 | 0% | 3.2% | 44% | 90% | 0% |
| Human Resources | 117 | 19 | 74 | 0% | 8.6% | 28% | 75% | 1% |
| Image | 396 | 52 | 49 | 10% | 8.8% | 17% | 19% | 35% |
| Learning | 271 | 42 | 63 | 30% | 2.9% | 23% | 31% | 17% |
| Legal | 122 | 11 | 81 | 0% | 9.0% | 56% | 77% | 2% |
| Marketing | 310 | 49 | 59 | 25% | 4.9% | 14% | 60% | 9% |
| Music | 63 | 19 | 73 | 1% | 23.2% | 44% | 10% | 17% |
| Productivity | 314 | 46 | 63 | 51% | 7.6% | 23% | 29% | 30% |
| Real Estate | 95 | 13 | 76 | 0% | 15.2% | 46% | 67% | 1% |
| Research | 126 | 14 | 76 | 1% | 10.0% | 37% | 57% | 9% |
| Sales | 301 | 44 | 65 | 18% | 5.9% | 31% | 55% | 10% |
| Search & SEO | 77 | 24 | 69 | 2% | 11.5% | 15% | 29% | 60% |
| Security | 357 | 33 | 70 | 0% | 1.9% | 42% | 72% | 6% |
| Social | 268 | 56 | 54 | 46% | 6.6% | 7% | 30% | 41% |
| Support | 253 | 39 | 68 | 39% | 6.3% | 34% | 65% | 7% |
| Video | 304 | 43 | 63 | 14% | 11.4% | 30% | 22% | 21% |
| Voice | 298 | 41 | 66 | 5% | 8.6% | 37% | 28% | 21% |
| Writing | 352 | 64 | 40 | 62% | 8.6% | 8% | 39% | 48% |
Active products have not shut down. Saturation and opportunity are market averages on a 0 to 100 scale. Wrapper share is the percentage of tracked products scoring 60 or more on wrapper risk. Shut-down rate is shut-down products as a percentage of everything tracked. Hot, paid only and bootstrapped are shares of active products with a known momentum label, pricing model and funding stage. Across the catalog: 7.3% shut down, 16.5% wrappers, 31% hot, 43% paid only and 17% bootstrapped. Source: FalcoScan catalog, data updated 25 September 2026.
1. How many products already compete?
Why it matters. The count is how many alternatives a buyer already has before they hear of you, and how many products you will be compared with in every sale.
How to read it. Use active products: everything in the market that has not shut down. It counts products, not companies, so a company with several products in a market counts once for each. A large count means buyers have choices; it does not on its own make a market crowded, which is what the next check measures.
Example. Coding has 435 active products, the most of any market. Gaming has 60, the fewest. A product entering Coding will be compared with about seven times as many rivals as one entering Gaming.
2. How crowded is the market?
Why it matters. Saturation is the most direct answer to how hard it will be to be noticed. Higher means more competitors and less room for a new entrant.
How to read it. FalcoScan scores saturation from 0 to 100 and a market's figure is the average across its products. The median market scores 33.1; the further above that, the harder the fight for each buyer.
Example. Writing scores 64, the highest of the 29. Healthcare scores 11, the lowest. The full rankings are in the most competitive AI markets and the AI markets with the lowest saturation.
3. How much room is left?
Why it matters. Saturation says how hard the fight is; opportunity says whether the prize is worth it. A crowded market with strong demand can still be worth entering, and a quiet one with weak demand may be quiet for a reason.
How to read it. Opportunity is FalcoScan's 0 to 100 estimate from its opportunity model, which weighs addressable demand, competitive saturation and differentiation. Because saturation is one of its inputs, the two tend to move in opposite directions. Opportunity minus saturation is what we call room. The median market scores 68.5 on opportunity.
Example. Voice and Design are about equally crowded, both at 41 on saturation, but Voice scores 66 on opportunity and Design 53. On crowding alone they look the same; on room they are 14 points apart. For which markets have the most room and what it is worth, read where there is still room to build in AI, from data frozen on 16 September 2026.
4. How many incumbents are thin wrappers?
Why it matters. A wrapper is a product that is mostly a layer over someone else's model. Where many incumbents are wrappers, a new thin product is as easy to copy as they are, and the company whose model they use can ship the same feature itself.
How to read it. Wrapper share is the percentage of a market's tracked products that score 60 or more on FalcoScan's 0 to 100 wrapper risk score; across the catalog it is 16.5%. A high share is a warning if you plan to build another thin product. A low share means the incumbents own something of their own, such as a model, data or integrations, so the bar for a new entrant is higher.
Example. In Writing, 62% of tracked products are likely wrappers, the highest share of any market. Security, with 357 active products against 352 in Writing, has no tracked product scoring 60 or more. The wrapper risk index has the measure for every market.
5. How often do products in the market shut down?
Why it matters. This is the one check that looks at outcomes rather than at the field as it stands today. It tells you whether the products that entered before you found buyers and stayed.
How to read it. FalcoScan checks whether each product is still running. The shut-down rate is shut-down products as a percentage of every product tracked in the market; across the catalog it is 7.3%, or 590 of 8,031. Acquired products are not counted as shut down.
Example. Gaming has lost 28 of the 88 products tracked there, 31.8%, the highest rate of any market, while scoring only 14 on saturation. Healthcare, about as uncrowded at 11, has lost 3.2%. Of the 14 markets below the median on saturation, ten lose products faster than the catalog as a whole. In those, part of the reason the field is thin is products that did not last.
Each point is one market. Across: average saturation, 0 to 100. Up: shut-down products as a percentage of all products tracked in the market. Catalog-wide shut-down rate: 7.3%. Named: the markets used as examples above. Source: FalcoScan catalog, data updated 25 September 2026.
For which markets bury their products and at what age, see what died, and where it went.
6. Which way is momentum pointing?
Why it matters. Momentum tells you whether buyers and builders are paying attention to a market now. Entering while a market heats up is a different job from entering one that has gone flat.
How to read it. Momentum is FalcoScan's editorial label for each product: hot, rising, stable or declining. It is a judgement, not a measured growth rate. The useful number is the share labelled hot, 31% of active products across the catalog. Declining tells you little on its own: only 46 of the 7,441 products that have not shut down carry it, because the label mostly marks products that have already gone (589 of the 590 shut-down products).
Example. In Avatars, 60% of active products are labelled hot, the highest share of any market. In Social it is 7%, the lowest. For a monthly view of the products with the most momentum, see the Rising 10.
7. How do the others price?
Why it matters. Your rivals set the buyer's expectation before you name a price. If most are sold on paid plans only, buyers expect to pay and a free plan stands out. If most are freemium, you will be compared with a free version of something.
How to read it. Every product carries a pricing model: free, freemium or paid. The table shows the share that is paid only, with no free plan; across the catalog it is 43%.
Example. In Healthcare, 90% of active products are paid only, the highest share of any market. In Music it is 10%. The pricing report covers how pricing varies across markets.
8. Who has raised money?
Why it matters. Funding tells you what you are up against. A field of bootstrapped products competes on product and price; a field of Series A and later companies can also outspend you on sales and marketing.
How to read it. FalcoScan records the funding stage of the company behind each product, from bootstrapped to public. The table shows the bootstrapped share; across the catalog it is 17%. A high share means you can start without raising, and so can everyone else.
Example. In Search & SEO, 60% of active products are bootstrapped, the highest share of any market. In Healthcare it is 0%, and 73% come from companies at Series A or later. More in who funds AI, and at what stage.
How do you put the eight checks together?
Read them as a screen, not a score. Start with the four that describe risk: saturation below the median market's 33.1, opportunity above the median's 68.5, wrapper share below the catalog's 16.5%, and a shut-down rate below the catalog's 7.3%. Three of the 29 markets clear all four today: Healthcare (saturation 11.0, opportunity 82.9), Automation (saturation 22.8, opportunity 73.4) and Security (saturation 32.7, opportunity 70.2).
Then read the other checks for what entering will take. The paid-only and bootstrapped shares are 90% and 0% in Healthcare, 78% and 3% in Automation and 72% and 6% in Security, against 43% and 17% across the catalog.
All of them sell mostly to buyers who pay from the start, and fewer of their products were built without outside money than in the catalog as a whole. The opening is real, but it favours a funded team that can sell to businesses.
Finally, open the market page and look at the products themselves. A market average hides the shape of the field: which corner of it is crowded, who leads, and who has already left. How to analyze an AI competitive landscape works through that for one market, step by step.
How does FalcoScan measure these?
Every product in the catalog carries a saturation score, an opportunity score and a wrapper risk score, each from 0 to 100, and a momentum label. Pricing model and funding stage are recorded per product. Market figures are recomputed from these every day, and FalcoScan checks whether each product is still running, which is what separates active products from shut-down ones. The Market Heat score combines several of these checks into one number per market. This page is rebuilt from the latest refresh, and the date at the top is the date of the data it used.
Where can you run these checks yourself?
The Market Terminal shows every market on these measures at once. To start with the markets that pass the screen, open Healthcare, Automation and Security. For what moved this week, read the Pulse, and for how every score is built, the methodology.
Questions people ask
What should a startup research before entering an AI market?
Eight things, each measured by FalcoScan for all 29 AI markets it tracks: how many products already compete, how crowded the market is (saturation), how much room is left (opportunity), how many incumbents are thin model wrappers, how often products shut down, momentum, how rivals price and who has raised money. As of 25 September 2026, active products per market range from 60 in Gaming to 435 in Coding.
Is an AI market with few competitors easier to enter?
Not on its own. Of the 14 AI markets below the median on saturation, ten lose products faster than the catalog-wide shut-down rate of 7.3%, based on FalcoScan data updated 25 September 2026. Gaming scores 14 out of 100 on saturation but has lost 31.8% of the products tracked there.
Which AI markets pass every check?
Based on FalcoScan data updated 25 September 2026, Healthcare, Automation and Security are the only markets below the median on saturation, above it on opportunity, and below the catalog on both wrapper share and shut-down rate. That makes them places to start looking, not a verdict: pricing and funding show what it takes to compete there.
Related articles
Photo: Nataliya Vaitkevich / Pexels. Colour-graded for Falcoscan.
Citing these numbers
Every figure here is read from the FalcoScan catalog, and this page refreshes daily. The numbers above are from the data updated on 25 September 2026. When you quote one, quote that date with it.
FalcoScan, “What Should a Startup Research Before Entering an AI Market?”, data updated 25 September 2026. https://falcoscan.com/articles/what-to-research-before-entering-an-ai-market