Gartner highlights that while initial AI infrastructure investments are often subsidized by hyperscalers and software vendors, a majority of these projects struggle to achieve full return on investment. This suggests that vendor incentives can mask underlying cost inefficiencies and project underperformance. Only a few AI initiatives, particularly in IT Service Management, demonstrate significant payoff, indicating a need for enterprises to look beyond initial subsidies for long-term financial sustainability.
The firm emphasizes a critical need for organizations to reassess AI investment strategies, as direct operational expenditures for AI integration are expected to increase. Enterprises must prepare for a transition from vendor-supported growth to managing potentially higher, direct costs. This strategic shift requires proactive financial planning to ensure sustained AI utilization and value realization beyond the initial adoption phase.
Gartner also points to evolving risks in specialized AI projects, with a significant number, such as mainframe exits, potentially failing. However, recent analysis suggests that migrating to mainframes could be a more cost-effective alternative for some organizations compared to other solutions, despite potential vendor lock-in concerns. This presents a nuanced view on infrastructure choices, balancing cost-efficiency against strategic risks.
Last updated May 10, 2026