Global AI spending projected to reach $2.5 trillion amid ROI concerns and price competition

Global artificial intelligence spending is projected to reach $2.52 trillion by 2026, according to Gartner, with infrastructure costs such as servers and data centers accounting for over half of that total. However, a PwC survey of 4,454 CEOs across 95 countries revealed that 56% have yet to see clear financial benefits from their AI investments, and only 12% reported seeing both cost savings and revenue increases.

Simultaneously, the generative AI market is seeing intense price competition. Data from the OpenRouter platform shows a shift in market share: the combined token processing share of major US firms like OpenAI, Anthropic, and Google dropped from 54.6% in January to 33.4% in June. Conversely, the share of five major Chinese companies rose from 20.3% to 48.3% during the same period.

As enterprises prioritize cost efficiency, model developers are diversifying their product lines to include various price tiers. Experts suggest that reducing token costs and improving inference efficiency through low-power semiconductors will become critical competitive factors for developers facing shrinking profit margins per token.