AI Agent Costs Set to Surge Fivefold by 2028 as Enterprises Race to Adopt Despite Hidden Expenses

AI agent costs are set to surge fivefold by 2028 as enterprises rush to adopt the technology despite hidden expenses and unpredictable returns, with Gartner warning of an 'inference paradox' where efficiency gains actually drive higher spending, while Nvidia's $105 billion financing of OpenAI's Ohio data center highlights just how scarce and critical AI compute infrastructure has become.

• Gartner predicts AI inference costs per agentic workflow will surge more than fivefold by 2028, driven by an 'inference paradox' where efficiency gains paradoxically fuel higher overall spending through increased token consumption.
• Enterprises are under intense top-down pressure to aggressively adopt AI agents despite unpredictable returns, with experts warning that companies often lack a full understanding of total cost of ownership before diving in.
• Demand for AI compute is so overwhelming that Nvidia is providing up to $105 billion in financing for OpenAI's Ohio data center, underscoring how compute infrastructure has become the defining — and increasingly scarce — resource of the AI era.