Technology research firm Gartner expects worldwide spending on artificial intelligence (AI) to reach $4.05 trillion in 2027, a 47.5% increase, as infrastructure takes 59% of the total.
The forecast spans infrastructure, software and services. Gartner projects infrastructure spending at $2.397 trillion, up 53% from $1.563 trillion in 2026. AI-optimized servers are the largest infrastructure category, with spending forecast to rise to $957.2 billion from $527.6 billion.
John-David Lovelock, Gartner’s Distinguished VP Analyst, Chief of Research, linked the infrastructure outlook to growing use of AI for inference and training tasks that need substantial computing capacity. Gartner also pointed to ongoing investment by hyperscalers and neocloud providers, as well as higher prices for AI-optimized servers.
Gartner forecasts spending on AI services will reach $746 billion in 2027, compared with $576 billion in 2026. Software spending is projected to rise to $649 billion from $458 billion over the same period. The figures show projected growth across both the computing systems used for AI and the services and software associated with it.
Spending on AI agents and assistants is forecast to increase to $67 billion in 2027 from $30 billion in 2026. Gartner lists that category separately in its outlook.
Reuters has reported that data-center spending is a growing source of revenue for industrial suppliers, including makers of generators, construction equipment, switchgear and chillers. The outlet also identified chipmakers and server developers among businesses benefiting from the buildout.
