Global spending on generative AI (GenAI) is expected to skyrocket to $644 billion in 2025, a 76.4% increase over 2024, according to a new forecast from Gartner. The surge will be led by hardware investments, with AI-capable devices and servers accounting for nearly 80% of total spending.

“The market’s growth trajectory is heavily influenced by the increasing prevalence of AI-enabled devices, which are expected to comprise almost the entire consumer device market by 2028,” says John-David Lovelock, Distinguished VP Analyst at Gartner. He notes that while consumer demand isn’t necessarily driving adoption, manufacturers are embedding AI features as standard.

According to Gartner’s breakdown, device spending will nearly double to $398.3 billion, while servers will climb to $180.6 billion. Software and services will follow at $37.2 billion and $27.8 billion, respectively.

Despite this explosive growth, Lovelock warns that enterprise enthusiasm for GenAI is being tempered by high failure rates in internal proof-of-concept (PoC) projects. “Expectations for GenAI’s capabilities are declining due to high failure rates and dissatisfaction with results,” he says. Issues such as poor data quality, user resistance, and unclear ROI have derailed many early efforts.

As a result, Gartner expects a strategic pivot in 2025. CIOs will move away from costly internal development and turn instead to commercial, off-the-shelf GenAI solutions embedded within existing software platforms. This shift could reduce standalone GenAI software spending, even as functionality becomes ubiquitous.

Notably, Lovelock emphasizes that hardware’s dominance will only grow: “The ratios shift more in hardware’s favor over time.” Future GenAI features will likely be bundled into broader software offerings, making it harder to isolate AI-specific spending.