Gartner forecasts global spending on AI infrastructure will reach $2.4 trillion in 2027, up 53% on 2026. The projected surge puts the infrastructure build-out, not just the models running on it, at the centre of AI’s next spending cycle.
Watch Desk analysis
What happened
ITPro reports that Gartner expects total AI spending to reach $4 trillion next year. Its forecast also puts AI infrastructure at almost 60% of that total, growing faster than overall AI spending, which is projected to rise 47.5%.
Within the infrastructure total, spending on AI-optimised servers is forecast to rise from $527.6 billion in 2026 to $957.2 billion in 2027. Gartner also projects spending on AI agents and assistants to grow 123%, cybersecurity 69%, and generative AI models 82%. These are Gartner forecasts, not final spending figures.
Why it matters
The numbers suggest the AI boom is becoming a vast infrastructure and operations bill: more servers, memory, power and capacity to run systems after they are trained. Gartner analyst John-David Lovelock attributes the rising forecast to faster-than-expected growth in inference demand, alongside compute-intensive training workloads.
For organisations, the headline is not simply that AI budgets are rising. The mix matters: infrastructure is forecast to claim most of the spend, while agents and assistants are among the fastest-growing categories. That makes the coming year a test of whether growing investment produces useful services and returns, rather than an impressive invoice and a rack of expensive kit.
Our read
A forecast this large is a signal of expectations, not proof that every planned build will happen or pay off. Still, the breakdown is useful: Gartner expects the machinery behind AI to absorb more spending than any other part of the market, with server costs set to climb sharply. The sensible question for buyers is what they need the capacity for, and whether a forecast demand curve is a good reason to commit now.
What to watch
- Whether Gartner’s 2027 spending forecasts translate into actual investment.
- How much of the projected infrastructure growth goes to servers, memory and other constrained components.
- Whether rising spending on agents and assistants brings measurable results for organisations.
Discussion spark: If AI infrastructure is forecast to take nearly 60% of AI spending, should organisations build capacity ahead of demand or wait for clearer evidence that the investment pays off?
Sources and evidence
- Source update (16 November 2026, 09:33 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.