'There is no reliable, economical one-size-fits-all model on the horizon': Experts claim AI costs will grow fivefold by 2028, as demand continues to soar

Gartner analysts are warning that the costs of implementing AI systems will increase fivefold by 2028, despite cheaper model prices. This phenomenon, known as the "inference paradox," occurs when improved token prices are offset by the costs of more complex AI agents. Sophisticated workflows, such as those used by AI agents, require more AI tokens than standard chatbots, leading to higher expenses.

While efficiencies in foundational model costs and AI efficiency are driving more sophisticated AI applications, the use of AI tokens in these workflows is outpacing the corresponding value. Agentic AI costs are currently outpacing results, and businesses are struggling to manage these rising expenses. Product leaders cannot count on more efficient token economics to offset AI costs.

Solutions may include partnerships among companies to share the burden, but the complexity and cost of these models will continue to rise. There is no reliable, economical one-size-fits-all model in sight, and developing competitive AI products will require creating and maintaining intricate multimodel ecosystems.