- The wave of consolidation that has already hit GenAI is about to sweep the Agentic AI market
- Gartner's Will Sommer argued this doesn't necessarily mean there's an AI bubble
- Those who have the resources to wait for customer demand to mature will capture a piece of the $3.3 trillion market
There’s a saying that always comes up just before a showdown in old Western movies: “This town ain’t big enough for the both of us.” The same could be said about the current agentic AI market, and that means some players – like their movie counterparts – will soon find themselves fighting to survive.
Agentic AI is in a weird spot at the moment. It’s poised to be a massive market in the near future – worth $3.3 trillion by 2029, according to Gartner. But in skating to where the puck will be in a few years, vendors have ended up too far out in front of where their customers are in the present day. The result is something that never bodes well for anyone selling something: too much supply and not enough demand.
“We’ve produced extensive research indicating that the adoption of agentic AI has been shallow and patchy, that implementation failure rates are high, that customers are dissatisfied with how AI agents are priced, and that there’s still a lot of concern out there about hallucinations, liability, security, what AI agents are making decisions,” Gartner Senior Director Analyst Will Sommer told Fierce.
That means there’s a disconnect between the number of providers out there selling agentic AI tools and the current size of the market. The natural result of this imbalance, of course, is consolidation.
We’ve already seen this play out in the generative AI space, and agentic AI deals are starting to pick up. Sommer pointed to deals between Salesforce and Informatica, SAP and SmartRecruiters, and ServiceNow and Moveworks this year as evidence of this trend.
What bubble?
Notably, Sommer said this does NOT mean the oft-cited AI bubble has burst. In fact, he argued the AI bubble doesn’t even exist – yet.
That might be hard to believe given $110.1 billion in global funding was poured into AI in 2024 and another $149.1 billion came in the first three quarters of 2025, according to Crunchbase data. The absurdly high values ascribed to recent AI compute deals involving OpenAI, Nvidia, Oracle and others doesn’t help either, nor do the wild valuations assigned to AI startups that haven’t even released their first product.
So Fierce pressed the issue with Sommer, who said in this case definitions really do matter.
Sommer, who has a master’s degree in economic policy and used to work as an economic modeler for the U.S. Energy Information Administration, said the definition of a bubble is when market valuations are disconnected from intrinsic value.
Remember, the market in question is slated to grow to $3.3 trillion. In that context, even $12 billion (the valuation given to Thinking Machines Lab in the link above) is a drop in the bucket (or 0.36% if you want to be precise).
While Sommer acknowledged that AI valuations are historically high and there are risks that could lead to a future bubble, he said he doesn’t believe we’re in Dot Com territory at the moment.
“We can actually say with absolute certainty that many agentic AI companies will fail or be acquired as the market concentrates towards a few dominant providers to capture what will be a multi-trillion dollar market for solutions,” he said. But, “this process of consolidation, it’s normal. It’s an expected part of the product lifecycle and it’s fundamentally distinct from a bubble.”
Winners and losers
Ok. So the not-a-bubble isn’t bursting but there IS a wave of consolidation coming. Who wins?
Sommer said there are three different groups of players in the AI race. The Inside Track cohort includes application service providers and large language model companies; the Field group includes hyperscalers and application service providers; and the Outside Track is basically everyone else – all the players that provide AI applications and services but which aren’t in favored positions.
In short, he said the winners will be the companies with the resources to sustain and scale agentic AI until the market reaches maturity (aka, until customers catch up). But that’s only one form of winning. Sommer said for startups, being acquired can also count as a victory.
“If you’re a large application software provider, one method by which you can continue to improve your likelihood of being in the dominant position when the dust settles is to acquire technologies that you don’t have,” he concluded. “If you’re an application software provider and you aren’t going to be able to dominate the market that you exist in, your best strategy may be acquisition.”