Where is New Semiconductor Fab Capacity Being Built?

There are a number of factors to take into account when analyzing the growth of the semiconductor industry. New electronics and improved digital functions also rely on increased processing power from much smaller components, resulting in more capable products in compact packages—the art of digitalisation at its finest. . 

Over the past few years, industries have seen their greatest feats of digitalisation and development stunted by the disruption to this market and, despite a necessary recovery, there are still major shifts taking place as sales increase by up to 15.2% year-on-year. 

Now in “severe surplus,” application is key for the sector, and those able to shift their efforts and deliver more artificial intelligence-compatible solutions will continue to reap their share of the market. For many years, the onus has been on the Asia-Pacific (APAC) region to deliver globally, accounting for a 60% share of the global market. 

With more components in the market, the real competition combines price with functionality, and firms like NVIDIA and Advanced Micro Devices (AMD) are reliant on more powerful chips to build increasingly intelligent hardware for some of the world’s leading technology companies. 

Trend: AI Drives Semiconductor Innovation

As AI becomes a staple component in digital devices and cloud services, operating such functions requires access to extensive pools of data as well as the capacity for more information to be incorporated as things unfold. The reality is that cloud operators require greater support from data center providers to enable significantly more computational power for their developments. 

This echoes up the chain to companies like NVIDIA, which manufactures next-generation graphics processing units (GPUs) that are capable of supporting AI functions, and further up the chain is chip production—in this case, Taiwan Semiconductor Manufacturing Company (TSMC) directly supports the global sales of the US-based hardware developer. 

Now that the semiconductor industry has entered a recovered state—a healthy one, in fact—the focus is on where new fabrication capacity will arise, and the race to leverage AI could be an important factor.

US Cloud Providers Share the Chip Demand

According to data by TrendForce, the increasingly high demand for popular AI servers will be influenced by North American cloud providers, such as Google, Microsoft, and Amazon Web Services (AWS)—key customers of NVIDIA and AMD. 

This will result in greater collaboration between the likes of NVIDIA and AMD as GPU providers and the technology giants leading the way in the digital environment. As a result, it is predicted that 60% of the global demand will come from the top players—Meta being another one—and Microsoft will acquire the most. 

So, what is the result of this? It’s likely we’ll see the US technology giants collaborating to build a high stake in the AI-driven world due in part to Asia’s leading semiconductor maker. 

In terms of GPU and other hardware fabrication, though, in order to deliver the AI functions that big businesses anticipate, the capacity for such components will take place in the US as a result of its internal partnerships between key hardware and software businesses. 

Odds Between the US and China for Global Semiconductor Fabrication

The reality that dawns on the technology industry is Asia’s overwhelming stake in the world’s technology innovations. An example of this can be found in the Newport Wafer Fabrication facility in the UK, which now shares majority ownership with Nexperia—a Netherlands-based yet Shanghai-owned business under Wingtech. 

The US-based company Vishay announced in November 2023 that it would buy Newport Wafer Fab for US$177 million, which somewhat represents the powers shifting from pillar to post—between the US and China. Nevertheless, it seems there can be no discussion in this sector without Asia’s participation—as we’ll discuss next. 

Europe to Regain Chip Fab Capacity

In September 2023, the European Union’s European Chips Act came into effect in a bid to regain control of local semiconductor sales and improve resilience in a fluctuating market, strengthening the region's technological leadership. This will allow greater investment in the EU semiconductor market of a staggering US$45 billion (circa €41 billion), with US$11.7 billion (€10.7 billion) injected into new research and development (R&D) and manufacturing capacity. 

It’s important to note that these funds are meant to attract investment from leading chip makers like Intel—the largest investor in its trade bloc—and TSMC, but largely to attract more development of European fab capacity. 

While Europe’s plans are highly ambitious, its extensive climate commitments also pose challenges as more emphasis is placed on reduced industrial emissions, which, in turn, poses constraints as energy efficiency is also a critical factor in its growing manufacturing efforts. 

Current Semiconductor Fab Challenges

Some of the challenges specific to the semiconductor industry are to be considered by the leading technology firms that wish to acquire more power in smaller packages. 

Lithography Constraints: Traditional optical lithography has served the industry well as the size of compounds reduced significantly over the years. Now, such printing technology reaches a ceiling, experiencing more complex issues with micro-components. The development of Extreme Ultraviolet Lithography (EUVL) hopes to overcome these limitations, but it can also be costly and defective. 

Shrinking Processing Nodes: In the true nature of digital, companies require more power in smaller and smaller packages. As semiconductors increase in density and performance, manufacturers face issues such as more current leakage, variability, and even more defects in production. 

Material Sourcing: Shrinking semiconductors requires new composition and more challenging materials to acquire. Increasingly compact units tend to incorporate novel metal compounds and materials with a higher capacity to store energy when compared to silicon dioxide. 

Cost: Countries increasing their capacity can expect a US$15 billion to US$20 billion cost to build a sufficient fabrication facility. In order to compete at high volumes, companies require more complicated systems to ensure production efficiency and quality components. For example, the EUVL solutions necessary to cater to this are expected to double compared to their current US$150 million individual cost. 

Skills Shortage: Despite the recovery of its material and component supply shortage and burgeoning state-of-the-art technologies, the industry still experiences a shortage of skilled workers to tackle a higher output.

The Environmental Aspects of Increase Fab Capacity

Markets must also pay attention to the ongoing trends in the energy sector. As energy sourcing and decarbonization factor into all industrial activities, increasing fab capacity comes as organizations are also expected to reduce their energy consumption—an energy-intensive process to build low-emission technologies of the future would only shift the industry’s impacts higher up the supply chain. 

Semiconductor Fab Capacity Hinges on Capital

All things considered, retaining the current growth trajectory requires significant investment in skills and innovation. Emerging countries in this space—as is the case with Europe—bank on the support of the leaders in the industry to provide the funds necessary to acquire larger shares of their local markets. 

It seems as though all paths lead back to the East, where well-established leaders continue to gain the attention of other regions to support their capacity increase in some way, shape, or form.