Gartner predicts 16.8% growth in semi sales for next year

According to Gartner, semiconductor sales are projected to grow by 16.8% in 2024, reaching a total of $624 billion. This comes after a decline of 10.9% in the market in 2023, with sales totaling $534 billion.

"We are at the end of 2023 and strong demand for chips to support artificial intelligence (AI) workloads, such as graphics processing units (GPUs), is not going to be enough to save the semiconductor industry from double-digit decline in 2023," says Gartner VP Alan Priestley. "Reduced demand from smartphones and PC customers coupled with weakness in data center/hyperscaler spending are influencing the decline in revenue this year."

However, there is hope for a bounce-back in 2024. All chip types are expected to experience revenue growth, driven by double-digit growth in the memory market.

Figure 1. Semiconductors Revenue Forecast, Worldwide, 2021-2025 (Billions of U.S. Dollars)

Source: Gartner (December 2023)

The worldwide memory market is forecasted to record a 38.8% decline in 2023, but it is expected to rebound in 2024 with a growth of 66.3%. The decline in 2023 is attributed to anaemic demand and declining pricing due to massive oversupply. NAND flash revenue, for example, is projected to decline by 38.8% to $35.4 billion.

However, the situation is expected to improve for vendors in the next 3-6 months. Gartner analysts predict a robust recovery in 2024, with NAND flash revenue growing to $53 billion, a year-over-year increase of 49.6%.

Similarly, the DRAM market is facing oversupply and declining demand, leading vendors to reduce inventory by lowering market prices. The oversupply is expected to continue through the fourth quarter of 2023, triggering a pricing rebound. The full effect of pricing increases, however, will only be seen in 2024, when DRAM revenue is projected to increase by 88% to reach $87.4 billion.

The demand for high-performance GPU-based servers and accelerator cards in data centers is being driven by developments in generative AI (GenAI) and large language models. This, in turn, creates a need for workload accelerators to support both training and inference of AI workloads in data center servers.

Gartner analysts estimate that by 2027, more than 20% of new servers will include workload accelerators as AI techniques continue to be integrated into data center applications.