Agentic AI costs set to balloon fivefold by 2028 The cost of agentic AI workflows is forecast to increase more than fivefold by the end of 2028 as users adopt more complex applications of the technology. As Nvidia and other tech giants push inference and agentic AI as the next stage of the AI wave, Gartner warns that the cost of implementing these systems will rise even as …
I don't know the term, but for some products, there's a perceived maximum price and I think we're going to see this with AI because the VCs are not going to be able to buy up the competition. Currently, there is mainly focus on the capability of the new models, but that's mainly because they're being reported upon by people who aren't into the economics. Most of the time, most of the users won't need most of the features and won't want to pay for them and if you look at Chinese markets, you'll see that providers are already reflecting this. China has by far the greatest uptake of AI – especially beyond coding and presentation automation – and is likely to become the benchmark for use.
"Where a simple chatbot must read and interpret a query and quickly respond with a probabilistic reasonable answer, an AI agent must constantly reason, negotiate, and question itself," he [Gartner senior director analyst Will Sommer] explains.Those processes add up: routing a task to an agentic reasoning model increases inference costs at least fivefold, and potentially by much more as the task becomes more complex.
That may or may not be true, but only until AI agents constantly reasoning, negotiating, and questioning themselves have the answer, or even just enough of the answer to the ultimate question of Life, The Universe & Everything.
Then does Life, The Universe & Everything have more than just otherworldly problems to negotiate and fail to vanquish without AIgentic Assistance for and from ITs CyberIntelAIgencies.
Companies that were previously telling their employees to use all the AI they could are now telling them to be more sparing in its use as they see their token costs exploding. When something like this hits a lot of CEOs don't want to be seen by the board/shareholders as behind the curve, so they adopt it even when they don't know what it is good for. We saw some of this in dotcom days too - a company would put up an internet presence but it was basically marketing materials and a contact us link. It took years for non tech native companies like plumbing suppliers or trucking logistics firms to make proper use of what the internet offered them.
Gartner always wants to extrapolate the latest hype ad infinitum. They are the ones who are responsible for the trend of bosses telling employees they MUST use a certain amount of AI resources every month, as if that guarantees smart adoption of AI where it is useful. It is like the "marketing materials and contact us link" websites from old line companies thing all over again.
Makes you just want to turn everything over to Artificial Ignorance decision makers like Altman & co. would like, eh?
Or not, if you have any functioning brain cells about how reliable those decision makers are in reality, and how vulnerable they are to being "engineered" into making offers you wouldn't like too much.
