Artificial intelligence (AI)-related applications are expected to drive global semiconductor revenue to 14% growth in 2025, according to new research from Gartner Inc.
This growth will result in a total of $717 billion, up from $630 billion this year.
“The growth is driven by a continued surge in AI-related semiconductor demand and recovery in electronic production, while demand from the automotive and industrial sectors continues to be weak,” said Rajeev Rajput, Senior Principal Analyst at Gartner.
The recovery in the semiconductor sector will be led by the memory market and rebound in GPUs. Specifically, memory is forecast to record 20.5% growth in 2025 to reach $196.3 billion fueled by prices falling for NAND flash.
Additionally, DRAM supply and demand will recover in 2025 with improved prices and production, Gartner said.
AI impact
GPUs, which have dominated the training and development of AI models, are forecast to rise to $51 billion in global revenue in 2025, or 27% growth.
“However, the market is now shifting to a return on investment (ROI) phase where inference revenues need to grow to multiples of training investments,” said George Brocklehurst, VP analyst at Gartner.
Specifically, this means an increase in high bandwidth memory to enable AI server performance. Vendors are significantly investing in HBM production and packaging to match next-generation GPUs for AI memory requirements, Gartner said.
HBM revenue is forecast to increase a whopping 284% in 2024 and then another 70% in 2025 to reach $21 billion. By 2026, about 40% of HBM chips will be used in AI inference workloads, compared to 30% today.