Apex Institute: AI Spending Targets Infrastructure Over Models
A new report reveals that the majority of artificial intelligence (AI) expenditures are directed towards the underlying infrastructure, including compute, storage, and networking, rather than the AI models themselves.
The Apex Institute's latest analysis, "Where AI Spending Is Actually Going in 2026," clarifies a common misconception about the primary destinations of artificial intelligence (AI) funding. The report indicates that the largest portion of AI expenditure is not allocated to the AI models directly, but rather to the essential infrastructure required to operate them.
The focus of these investments lies heavily on compute power, storage solutions, and networking technologies, which are the foundational elements enabling complex AI models to function. This infrastructure layer constitutes the most significant part of AI budgets, contrary to a popular perception that often centers on the models themselves.
The analysis underscores the necessity of a robust and scalable technological foundation to support the development and maintenance of AI systems. This suggests that organizations must make substantial investments in hardware and network infrastructure to ensure the efficiency and reliability of their AI solutions moving forward.
The report provides valuable insights for businesses and investors planning AI initiatives, aiding in the strategic allocation of resources. It also highlights the growing importance of professionals working within the infrastructure domain for the advancement of the AI sector.