The semiconductor industry is the foundation of electronic information technology, which is widely used in computers, communications, consumer electronics, automotive electronics, industrial control and other fields. In the future, with the rapid development of 5G, Internet of Things, artificial intelligence, cloud computing and other technologies, the semiconductor industry will usher in round after round of growth opportunities.
Based on research from Deloitte, this report analyzes the global semiconductor industry trends, key drivers, potential challenges, and strategies to address them in 2024. Particular attention is paid to the role of generative artificial intelligence (Gen AI) in driving chip demand, the progress of smart manufacturing, the geography of global supply chains, cybersecurity threats, and the impact of geopolitical factors on the industry.
According to Deloitte's forecast, chip sales will rebound in 2024 driven by the generative AI boom, but geopolitical influences could complicate the situation. 2023 was the 7th downturn experienced by the semiconductor industry and a challenging year, with sales falling to $520 billion, but global sales are expected to rebound to $588 billion in 2024, up 13% from 2023 and 2.5% higher than the record $574 billion in 2022.
As always, the memory chip market is the most volatile. In 2022, the memory market sales were nearly $130 billion, accounting for less than 23% of the entire chip market, but in 2023, it fell by 31% (about $40 billion), while logic chip sales fell by 1%. The market is expected to return to a stable level in 2024.
Chart: Global semiconductor industry outlook for 2024 (Source: Deloitte).
Generative AI and the new wave of chip supply
As generative AI technology continues to evolve, so does the need for high-performance computing power. Gen AI technology enables the creation of new data patterns and content, opening up new possibilities for chip design and applications. Gen AI model training and inference require a lot of computing resources, which directly increases the demand for chips such as high-end GPUs, ASICs, and FGPAs. Gen AI chips are predicted to reach $50 billion in sales by 2024, accounting for 8.5% of total sales, with a significant impact on the semiconductor industry. The DRAM requirement for Gen AI applications can range from 5 to 13 million wafers, which is equivalent to 4 to 12 dedicated fabs. The total demand for NAND memory is expected to be between 2 and 8 million wafers, corresponding to 1 to 5 fabs.
The development of generative AI technology has driven a new wave of chip supply, which in turn has led to the further development of generative AI technology. Specifically, as chip performance improves and supply increases, generative AI algorithms can operate more efficiently, creating higher quality and more diverse content. At the same time, with the continuous popularization and application of generative AI technology, its demand for high-performance computing power will further increase, which will further promote chip suppliers to increase R&D and production, and promote the further development of chip supply.
However, there are some challenges to the mutually reinforcing relationship between generative AI and chip supply. First, as generative AI technology continues to evolve, so does the need for high-performance computing power, which could lead to a tight chip supply. Secondly, due to the complexity and diversity of generative AI technology, different application scenarios may require different types of chip products, which is also a huge challenge for chip suppliers. However, these challenges also present opportunities. For example, by developing and producing chip products that can meet the needs of different application scenarios, chip suppliers can further expand their market share and gain more business opportunities.
