Gartner: AI Agent Inference Costs to Rise Fivefold by 2028
Gartner forecasts that AI agent inference costs per workflow will rise more than fivefold through 2028, despite falling prices per token. The firm says multi-step agent workflows use far more tokens, often rely on pricier models and require more reasoning and tool calls. A complex workflow could cost five to 30 times as much as a basic chatbot exchange, while training more capable models may cost 2.5 times more. Gartner calls this the “Inference Paradox”: cheaper AI tokens can encourage wider use, increasing overall spending. It recommends routing simpler tasks to lower-cost models, coordinating different models and monitoring inference costs closely. The forecast concerns AI infrastructure spending and has no direct cryptocurrency market impact.
Neutral
The forecast is about AI operating costs, not cryptocurrency markets, so it has no clear direct effect on crypto prices, trading volumes or market stability. In the short term, traders may take it as a signal that demand for computing capacity and AI infrastructure could continue to grow, potentially influencing sentiment around AI-linked technology or crypto projects. However, any such response would likely depend on broader market conditions and company-specific developments rather than this forecast alone. Over the longer term, higher inference costs could encourage investment in more efficient computing, decentralized infrastructure or AI-related blockchain projects, while also putting pressure on firms whose business models rely on expensive AI workloads. Similar AI and data-centre narratives have at times affected technology-sector sentiment, but they do not reliably translate into sustained moves across crypto assets. Traders should therefore treat this as an indirect thematic development, not a standalone directional catalyst, and watch broader risk appetite, token-specific news and liquidity conditions.