30 October 2024, Wed |
8:19 AM
According to the latest forecast from Gartner, Inc., global semiconductor revenue is expected to surge by 14% in 2025, totaling $717 billion. This growth follows a remarkable 19% increase projected for 2024, with the market set to reach $630 billion.
After a decline in 2023, the semiconductor sector is rebounding, anticipating double-digit growth for both 2024 and 2025. Rajeev Rajput, Senior Principal Analyst at Gartner, stated, “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.”
In the near term, the memory market and graphics processing units (GPUs) are expected to significantly boost global semiconductor revenue. The worldwide memory revenue market is forecasted to grow by 20.5% in 2025, reaching $196.3 billion. Sustained undersupply in 2024 will drive NAND prices up by 60%, though a 3% decline is anticipated in 2025. NAND flash revenue is expected to hit $75.5 billion in 2025, marking a 12% increase from 2024.
DRAM revenue is also set to recover, with total revenue projected to reach $115.6 billion in 2025, up from $90.1 billion in 2024. This rebound is attributed to improved supply, high-bandwidth memory (HBM) production, and rising prices of double data rate 5 (DDR5) memory.
AI’s Growing Influence on the Semiconductor Market
Since 2023, GPUs have played a crucial role in the training and development of AI models. Their revenue is expected to total $51 billion, reflecting a 27% increase in 2025. However, George Brocklehurst, VP Analyst at Gartner, notes that the market is shifting towards a return on investment (ROI) phase, where inference revenues must grow significantly to match training investments.
The demand for HBM, a high-performance AI server memory solution, is set to rise sharply. HBM revenue is projected to increase by more than 284% in 2024 and by 70% in 2025, reaching $12.3 billion and $21 billion, respectively. Gartner analysts predict that by 2026, over 40% of HBM chips will support AI inference workloads, compared to less than 30% today, largely due to increased inference deployments and limited repurposing of training GPUs.
As the semiconductor industry continues to evolve, these trends highlight the critical role of AI in shaping market dynamics and driving revenue growth in the coming years.
