Which AI markets are heating up, and which have gone quiet

Everything in AI is described as accelerating, which makes the word useless. The more answerable question is comparative: of the twenty-nine markets we track, which ones have a majority of products still gaining, and which ones have quietly stopped? The gap between the two ends is a factor of ten.

What a growth signal is

Each product carries one of four labels — hot, rising, stable or declining — derived from movement in the signals we track over time, then decayed so that a reading we have not refreshed recently counts for less. It is a statement about trajectory, not about quality: a stable product can be excellent and profitable, and frequently is. What the label says is whether the thing is still gathering pace.

328 live tools carry no signal yet, because that requires a history of checks we have not accumulated for recent additions. They are excluded from every rate below. The full definition and the decay we apply are on our methodology page.

Almost nothing is falling off a cliff

Only 46 products out of 6,113 are classified declining. That sounds like good news and mostly is not — it is a description of how decline actually looks in this category. Products do not crash; they stop moving. The 1,454 stable products are where the story is, and in a market growing as fast as this one, holding still is a relative loss.

The ten-fold spread

Legal at 76% and Healthcare at 72.1% lead, and they are the same two markets that top our opportunity table. That is not circular — opportunity measures unmet demand and momentum measures observed movement — but it is consistent, and it is the clearest signal in this data that those two markets are genuinely in motion rather than merely under-served.

At the other end, Social (7.2%) and Writing (8%) are the two markets where the fewest products are still gaining. Both are also the two most saturated markets in the catalog, and the two with the highest wrapper density. Three different measurements arriving at the same conclusion is about as strong as this kind of evidence gets.

Momentum runs opposite to size

Line the markets up by number of products and by share of hot products, and the two orders largely invert. The five biggest markets are Coding, Image, Design, Writing and Agents. Image ranks 24th of 29 for momentum, Design 25th and Writing 28th. Coding sits mid-table at 13th. Only Agents, at 4th, is both large and still moving — and it is the youngest of the five by some margin.

This is the most practically useful thing in the piece. Market size is a measure of where attention went eighteen months ago. Momentum is a measure of where it is going now. Choosing a market on size alone means arriving exactly where everybody else already is — and the size-to-momentum inversion is how you can see that before your own numbers tell you.

What the moving end looks like

Six of the highest-opportunity products currently carrying a hot signal, across different markets.

What cooling looks like from the inside

The products that lose momentum in this catalog tend to do so for one of three reasons, and none of them are that the product got worse. A model provider ships the capability natively, and a category that was a product becomes a feature. The company is acquired and the standalone product stops being invested in. Or a version is superseded by its own successor — several entries in the declining set are simply earlier generations of models that are still, in a sense, thriving under a different name.

All three are worth distinguishing from failure, and none of them are avoidable by building a better version of the same thing. The defence is the same one the rest of our data keeps pointing at: own something the model provider does not.

Every market, by momentum

Counts of each signal per market. Markets are ordered by their opportunity-to-saturation gap, the same order as our room-to-build analysis.

Photo: Mike Huang / Pexels. Colour-graded for Falcoscan.

← All articles