Dive Brief:
- Generative AI spending will increase 76% year over year to $644 billion in 2025, driven primarily by vendor investments in the technology, according to Gartner data published Monday. The segment exploded last year, growing 337% to $365 billion.
- “We are now in the midst of a technology provider build-out phase,” said John-David Lovelock, Gartner distinguished VP analyst, pointing to massive hyperscaler capital expenditures on GPU infrastructure and the software vendor rush to deploy generative AI tools.
- In contrast, enterprises are paring back on in-house AI projects in favor of off-the-shelf tools. “CIOs are no longer building generative AI tools, they’re being sold technology,” said Lovelock. “The vendors are coming in saying, ‘I’ve got what you need.’”
Dive Insight:
Enterprises hit a wall with generative AI last year. The learning curve was sharp and painful, as pilot programs stalled and proofs of concept racked up eye-watering bills.
“The capital cost to create your own model was much higher than organizations expected; the inference costs were higher than expected, and the value was less than they expected.” Lovelock said.
AI skills gaps and GPU scarcity coupled with budget constraints and lagging infrastructure caused project delays, as many companies opted to abandon pilots before they reached the production stage.
“We overestimated how much money would be spent, but what we did get right was the failure rate,” Lovelock said. “Enterprise data wasn't up to it, the enterprise itself wasn't up for changing, and the ROI wasn’t there.”
Companies that started 2024 with hundreds of generative AI pilots pared the list back to just a handful by the end of the year, according to Lovelock.
Despite enterprise frustration with the technology, cloud and software providers soldiered on, driving massive spikes throughout the AI market. Spending on servers and software more than doubled year over year and the device category saw unprecedented growth of 846%, according to Gartner. Hardware accounted for 80% of the total AI spend in 2024.
Software emerged as the nexus for enterprise AI spend, Lovelock said. The category more than tripled in size last year and is expected to nearly double in 2025 to $37 billion, driven by AI software services consumption.
“This is the year that we go from a little bit of money being spent on software with Gen AI to a tidal wave of money being spent on it,” Lovelock said. “If you're using a cloud software or platform service, you may just end up one day coming in and seeing a new button that is the Gen AI feature.”
CIOs are bracing for the impact, setting 9% of the IT budget aside for price increases on existing services, according to Gartner research.
“Most companies aren't getting 9% budget increases,” Lovelock said. “So, they're taking away from their opportunistic budget in order to pay for price increases.”