Gartner's $1.5T 2025 AI forecast turns spend into capacity risk

Gartner's 2025 AI spending forecast is a evidence-led market-signal event. It matters because the forecast turns AI demand into a hardware and infrastructure capacity question, while the later 2026 update shows the benchmark continued to move upward.

Gartner's September 2025 AI-spending forecast is a market-signal event, not proof that every AI initiative is producing returns. The evidence-led claim is narrower and more useful: Gartner forecast worldwide AI spending at nearly $1.5 trillion in 2025, then pointed the mechanism toward infrastructure expansion, AI-optimized hardware, GPUs, servers, AI-enabled devices, software and services. The phrase "this year" refers to 2025, because the public forecast was published on 17 September 2025.

The human actors are straightforward: Gartner published the forecast, and John-David Lovelock is the named Gartner analyst whose public profile ties him to technology-market economics and global IT spending trends. The spending figure is not the story by itself. The story is the machinery underneath it: buyers and suppliers are being pushed into a cycle where servers, chips, devices, data centers, software packaging and cloud capacity all move together.

The intelligence value is the shift from AI narrative to capital allocation. Gartner's numbers frame AI as a hardware-and-infrastructure cycle: data centers must be financed, accelerated servers must be procured, semiconductors and memory must be allocated, and device makers must absorb AI features into refresh cycles. The pressure point is not whether AI is popular; it is whether buyers can turn a swelling spending base into capacity, product differentiation and measurable enterprise value.

A May 2026 Gartner update matters because it moves the benchmark forward: worldwide AI spending was then forecast at $2.59 trillion in 2026. That makes the September 2025 forecast a dated but still material baseline. The watchpoints are category mix, server and semiconductor pricing, hyperscaler capacity discipline, AI-enabled device adoption, enterprise software bundling, and whether spending growth outruns provable productivity gains.