Gartner: LLM token costs fall 95% by 2030 but agentic workflow inference costs rise 5x in two years

According to Gartner research reported in CIO, while LLM token prices will decline 95% by 2030, overall agentic AI workload costs are increasing more than fivefold over the next two years—a phenomenon Gartner terms the inference paradox. The cost increase occurs because AI app builders use more and more expensive tokens as models grow more complex, meaning innovation velocity outpaces cost reduction. This creates a divergence between unit economics and total cost of ownership for agent-based systems.

Topics #

Sources #

• Press

Read article

Go deeper #

This intelligence is sourced automatically from public sources across the web and synthesised by the Prefactor AI pipeline. Stories are reviewed before publication.