The difficulties facing the semiconductor industry have become so well-known recently that mainstream news stories about chip shortages, and the corresponding problems they’re causing, have become commonplace. Along the way, a much wider audience has begun to learn some of the idiosyncrasies and structural challenges that face the chip business.
Chip shortages are triggering important plans within big semiconductor firms. The most recent example is GlobalFoundries’ announcement of greatly expanded capacity at its existing chip fabrication plant in Singapore. The company announced a $4B investment in conjunction with Singapore government officials and committed customers to build a 250,000 square foot addition that’s expected to open for production in 2023.
The investment will enable the company to increase its overall output by 33%, growing its 300 mm wafer capacity from about 720,000/year up to just under 1.2 million/year, in addition to the nearly 700,000 200 mm wafers they also manufacture there each year. (Worthy of note, the news was relayed to me by an all-female contingency of executives from GlobalFoundries—a first in more than 20 years of covering the semiconductor business.)
The need for additional chip production capacity has become painfully obvious during the pandemic, so it’s clearly great to see the GlobalFoundries (GF) investment being made. The news of this expansion is important both to the company and to the semiconductor industry overall for several different reasons.
First, it places some much deserved attention on the part of the industry that, frankly, has been overlooked for much of the last few decades. For those who do watch chip industry developments, most of the press and attention is directed towards what are referred to as leading process nodes. In English, that means smaller and smaller chip sizes—for example 7nm (nanometer), 5nm, 3nm, etc. While these cutting-edge chips do power smartphones and a lot of other sexy, important devices, they only represent about 30% of the chip industry’s total revenues and about 10% of the units manufactured.
More importantly, they’re extremely difficult and expensive to develop. That’s why, several years back, GF made the—at the time, rather controversial—decision to no longer play the nanometer game but instead focus on adding new capabilities and features to chips that were built at larger transistor sizes (12 nm and up). Not only do these types of chips make up the bulk of both shipments and revenues for the semiconductor industry, but it turns out there are quite a few applications for which they are the best technology to use. Applications like RF (radio frequency), analog power, and non-volatile memory, for example, which are critical for things like 5G, automotive, IoT, and many other industries, are best suited to the chip sizes that GlobalFoundries chose to focus on.
Unfortunately, much of the industry had bought into the notion that everything related to chips was only better if it was smaller. As a result, appropriate amounts of investments in these less “sexy” types of chips were not made, and that led the entire industry to the critical shortage phase that it finds itself in. The GlobalFoundries investment is clearly a step in the right direction in starting to address the issue around fewer cutting-edge nodes. The fact that the company also referred to similar types of investments it plans to make in both the US and Europe highlights the increased emphasis these types of components are receiving.
Unfortunately, the GF news also highlights how expensive and slow changes to the chip industry are. While many major manufacturers, particularly automobile makers, are clamoring for as many chips as they can get as soon as they can possibly get them, the sobering truth is that there is no easy answer to this problem and that it will be several years before it can be resolved.
As long-time chip industry observers know, the semiconductor business has always been highly cyclical with many different phases of shortages and surplus because of the time it takes to fill in the gap. With some forecasts expecting semiconductor shipments and revenues to more than double throughout this decade, this time things may be different, and potential oversupply issues seem to be a long way off.
Another factor is that the environment around the chip industry is very different now than it was pre-pandemic. For one thing, the concentration of chipmaking facilities in very few regions around the world has become an enormous geopolitical issue. In fact, large governments around the world are now clamoring to put policies into place that encourage local production of chips.
Here in the US, the Senate recently passed the CHIPs for America Act, which sets aside $52 billion in incentives for US-based manufacturing and R&D efforts, while the newly introduced bipartisan Facilitating American-Built Semiconductors (FABS) Act is expected to provide additional tax incentives for US chip manufacturing. Government leaders are recognizing the criticality of a geographically diverse set of semiconductor suppliers, and even chipmakers themselves are starting to make more efforts at expanding their global reach—something that GlobalFoundries has been doing for more than 10 years.
The fact that the GF announcement was made in conjunction with the Singapore Economic Development Board is a clear manifestation of these principles, as many future chip manufacturing deals will also likely reflect these private-public partnerships. The roughly 1,000 highly skilled jobs expected to be created as part of the fab expansion is yet another reason government interest is so strong. The GF deal is also interesting because of the participation by what it termed “committed customers.” Clearly multi-billion-dollar manufacturing investments cannot be wisely made without firm, long-term financial commitments from key customers—it’s no longer a “build it and they will come” kind of environment. As a result, I expect to see other new fab facility news, both from GlobalFoundries and other industry players, to incorporate a customer element to them.
The pandemic has made clear our societal dependence on digital, chip-driven devices and services. As a result, providing the right political environment and appropriate financial incentives to ensure a steady, secure, and diverse supply of the critical chips that power them will unquestionably be a key priority for both businesses and governments for some time to come. As a US-based company, it’s great to see GlobalFoundries help set the tone for how these types of global semiconductor issues and deals will start to get done.
Bob O’Donnell is the founder and chief analyst of TECHnalysis Research, LLC a technology consulting firm that provides strategic consulting and market research services to the technology industry and professional financial community. You can follow him on Twitter @bobodtech.