After a period of slow growth, signs are emerging of a rebound in the global semiconductor market, fueled by a surge in demand for AI technologies.
The Semiconductor Industry Association (SIA) reported this week that global sales in the semiconductor industry reached $48 billion in November 2023. This figure represents a 5.3% rise compared to the $45.6 billion recorded in November 2022 and shows a 2.9% increase from the $46.6 billion total in October 2023.
“Global semiconductor sales increased on a year-to-year basis in November for the first time since August 2022, an indication that the global chip market is continuing to gain strength as we enter the new year,” said John Neuffer, SIA president and CEO. “Looking ahead, the global semiconductor market is projected to experience double-digit growth in 2024.”
This is alongside Taiwan Semiconductor Manufacturing Company (TSMC) surpassing fourth-quarter revenue estimates and after a SEMI report that growth could resume in 2024, with sales forecast to reach a new high of $124 billion in 2025.
Other research organizations, including IDC, have revised their semiconductor market forecasts, signaling a turnaround and predicting accelerated growth in the coming year.
Why the current demand drive is different from before
As with many other things shaping the technology landscape at the moment, AI is set to play a vital role in the chip market recovery. The AI chip market is gaining momentum, driven by significant engagement from key players.
Nvidia and AMD have both announced new AI chips for desktop computers. Intel is making a strategic move into this space, launching AI-enabled chips for the automotive sector. TSMC’s improved performance was also primarily driven by strong demand from artificial intelligence sectors.
This means there is a distinction in the current market compared to earlier growth drivers. Analysts point out that each decade, the semiconductor industry has been driven by specific products. Once, it was laptops, followed by smartphones, and in between, gaming and graphics chips. What renders the current revolution significant is that the driving factor is a concept, not a tangible product.
“The current trend is AI, but what’s interesting, in my opinion, is the difference from previous trends,” said Arun Mampazhy, an independent market analyst. “Earlier, the drivers were product-based – smartphones, gaming chips, etc. – whereas AI is concept-based. Now, we’re seeing a situation where a concept is pushing chip demand, and this could permeate everything, from AI-driven automobiles, like self-driving cars, to the next level of AI in smartphones. It’s going to be interesting to see how this develops.”
The flip side, Mampazhy pointed out, is whether it can live up to the hype surrounding it.
“AI is a significant driving factor,” Mampazhy added. “It differs from previous trends as it’s concept-based rather than product-based. Whether it will stabilize or continue to grow is something that time will tell.”
Manish Rawat, semiconductor analyst at TechInsights, pointed out that chipmakers must concentrate on developing optimal solutions for AI workloads and chip optimization to gain market dominance. The development of power-efficient CPUs for edge devices will result from the transition towards edge computing, pushed by the demands of real-time processing.
“As AI applications need customized solutions, customization becomes essential and requires working with AI software developers,” Rawat said. “Maintaining a competitive edge requires constant R&D investment, and collaborations between startups and academic institutions could encourage innovation. It is essential to implement security measures, such as hardware-based encryption, to safeguard AI systems.”
For long-term development and innovation in the AI-driven future, effective global supply chain management and regulatory compliance that takes ethical considerations into account are essential. For semiconductor companies to properly position themselves in the changing landscape, they must negotiate these possibilities and difficulties.
Challenges ahead, but opportunities, too
AI applications require specialized hardware, such as GPUs and TPUs. This means semiconductor manufacturers must invest significantly in R&D. Designing and manufacturing these customized chips necessitates a deep understanding of AI algorithms. The semiconductor industry faces challenges in keeping up with the rapid expansion of the AI sector and the evolution of AI models and algorithms.
“It is essential to keep financing R&D and maintaining a flexible manufacturing environment,” Rawat said. “Geopolitical tensions and world events can cause supply chain interruptions in the semiconductor sector. Supply chain stability is critical given the increasing demand for semiconductors linked to artificial intelligence. To reduce possible interruptions and maintain resilience, companies need to diversify their suppliers, have access to necessary raw materials, and create backup plans.”
Furthermore, as companies struggle to recruit and maintain an adequate workforce, a lack of specialized people in creating cutting-edge semiconductor technologies for AI applications may impede innovation and product development.
But not all is bleak. Semiconductor companies have numerous strategic options to benefit from the growing market caused by the increasing demand for AI applications. First, focusing on the AI industry enables a larger market, which boosts earnings and market share.
“Working together and establishing alliances with AI firms may boost product offers, accelerate innovation, and more thoroughly meet market demands,” Rawat said. “Expanding product offerings to accommodate the diverse needs of the artificial intelligence industry fortifies resilience and draws in more customer base. Investing money into cutting-edge technologies like neuromorphic chips and quantum computing gives companies a competitive edge and puts them at the forefront of AI innovation. Maintaining a lead in AI developments and semiconductor design requires constant investment in R&D.”
Furthermore, long-term success in the ever-changing AI semiconductor market requires developing talent, broadening product offerings, strengthening supply chain resilience, forming strategic partnerships, keeping an eye on future technologies, and monitoring these developments.
Asian domination to continue, but US could rebound
Despite recent US and European efforts to enter the semiconductor manufacturing market, Asian countries are expected to lead. According to SIA, regionally, year-to-year sales were up 7.6 percent in China and 7.1 percent in the rest of Asia Pacific. Europe and the Americas saw 5.6 and 3.5 percent increases, respectively.
SEMI expects China, Taiwan, and Korea to remain the top three destinations for equipment spending through 2025. China is projected to maintain the leading position over the forecast period as the region’s equipment billings continue to soar.
However, Mampazhy pointed out that the recent efforts to increase manufacturing in US soil could have an impact in the near term.
“By revenue, the percentage of manufacturing on US soil is very likely to increase due to the efforts of Samsung, TSMC, and Intel, especially since these sources are likely to have much higher profit margins because they focus on advanced nodes,” Mampazhy said. “In that sense, the US is poised for a rebound in manufacturing. However, how much of this growth is driven purely by US companies remains to be seen.”
Copyright © 2024 IDG Communications, Inc.
This story originally appeared on Computerworld
