Quick Overview

  • AI data center spending will account for more than half of total semiconductor revenue over the next four years, up from 36.5% in 2026, according to a forecast released by Gartner on Monday.
  • The research firm said growth in data center spending will push global semiconductor revenue past $1.6 trillion this year, nearly doubling year over year.
  • Although memory revenue remains the main contributor to overall semiconductor industry growth this year—expected to account for 54% of semiconductor revenue in 2026—the increased share of AI data center revenue highlights a "shift in how semiconductor value is created and how industry demand is evolving," said Ben Lee, a director analyst at Gartner.

Deep Insights

AI demand is driving more data center construction to support computing. However, enterprise executives are entering a more cautious deployment era as they face rising AI costs.

Providers are rapidly shifting from fixed-rate subscription pricing to usage- or outcome-based models. This pricing shift is not something organizations are well equipped to handle, Justin St-Maurice, a technology advisor at Info-Tech Research Group, said in an email to CIO Dive.

"In the early days, most vendors' strategy was to get enterprises 'hooked' on technology usage," he said. "During the adoption phase, the real costs were actually hidden."

But now, bills are piling up quickly. Continuous agent loops are the biggest contributor to rising AI costs, St-Maurice said. The technology second-guesses itself and reprocesses data, driving up final costs in unpredictable ways, he said.

Amid growing concerns, vendors such as Snowflake, AWS, and Oracle are implementing features to help enterprises control AI costs.

However, CIOs and other executives looking to better manage AI spending may need to look to China, which has created open-source models that can compete with U.S. frontier models, St-Maurice said. Depending on the use case, CIOs should consider using these models.

"The future strategy will involve multi-tier computing, with desktop devices handling most of the basic work and complex tasks escalating when needed," St-Maurice said.

Travelers Insurance built its own large language model to help mitigate rising AI costs. TravelersLLM handles insurance-related queries and costs less for the company to run than frontier models, said Mojgan Lefebvre, executive vice president and chief technology and operations officer at Travelers.

The company still uses frontier models for broader reasoning or coding queries. However, cheaper internal models help offset higher frontier model costs, Lefebvre said in a July interview.