Forbes Technology Council | Scott White 22 Jun 2022
In the face of growing geopolitical concerns and a global chip shortage, Western governments are considering generous packages to boost semiconductor manufacturing. In my last article, I discussed how such support packages have the potential to grow the economy and generate thousands of skilled jobs. This time, I look at another urgent question: How can the money be best spent?
In the late ’80s, the U.S. launched SEMATECH, a consortium that received $500 million in government funding to regain competitiveness in the U.S. semiconductor manufacturing industry. The consortium has a complicated legacy: Although SEMATECH became the standard for how industry and government could work together to restore manufacturing industries, it didn’t produce leaps in semiconductor manufacturing technology as originally envisioned.
A successful strategy will need to acknowledge the complexity of the semiconductor industry and address the parts in which domestic efforts can really move the needle. And although the legislations underway look promising, careful thinking needs to go into implementation to ensure long-term benefits.
A quick examination of the $52 billion U.S. CHIPS Act shows that the legislation places emphasis on developing foundries for three-nanometer transistor processes, aiming to compete with the semiconductor foundry giants in Taiwan and South Korea. These three-nanometer fabs are used to produce bleeding-edge integrated circuits (ICs) that power the latest smartphones and are extremely resource-intensive: A single three-nanometer fab can cost up to $25 billion to build and needs vast amounts of water and energy to operate, resulting in ICs costing hundreds of dollars or more.
However, recent supply chain issues have often been caused by the shortage of simpler run-of-the-mill ICs based on older generation transistor processes that cost just a few dollars or less, such as heated seat controllers for cars or drivers for LED/LCD displays. Domestic manufacturing of these types of chips is also vital and could be far more cost-effective and economically scalable.
Rather than focusing mainly on how to on-shore three-nanometer semiconductor technology, governments should encourage innovation across the entire electronics ecosystem. This includes supporting a wide range of process technologies in both silicon and alternative semiconductor materials that have distinct capabilities and where truly world-leading capabilities can be fostered.
With Moore’s Law nearing its physical limit (or at least its economic limit in continuing to reduce cost per transistor), future innovation will also require new approaches like chip-scale packaging, photonic integrated circuits and agile hardware customization. Another area in which the industry needs to urgently innovate is in the reduction of its environmental footprint, which is currently a major contributor to global carbon emissions.
In its missive to the U.S. Department of Commerce, the U.S. government-supported not-for-profit strategic investor In-Q-Tel recommended ensuring adequate support to startups and small businesses, which are integral to fostering innovation. It also advised setting up “fabrication sandboxes,” which give smaller companies access to commercial equipment and tools, thereby helping remove barriers to entry into the industry.
Across the pond, the U.K. Innovation Strategy positions the U.K. primarily as a pioneer in chip design, despite its homegrown semiconductor manufacturing clusters with strong capabilities in compound semiconductors and flexible chips. The U.K. would greatly benefit from having a trade association similar to the Semiconductor Industry Association in the U.S. to effectively convey the strengths and requirements of industry players. The government also needs to support industry plans for upskilling the domestic employee base to fill in the new jobs in the pipeline through training and apprenticeship programs at every education level.
Setting up a three-nanometer fab capable of manufacturing bleeding-edge AI chips would make for a nice headline, but it won’t help in building a sustainable semiconductor industry that covers the multifaceted needs of the wider markets for electronics. We need to collaborate with our politicians to help them understand the complex choreography of the industry and to guide governmental strategy to support the long-term vision of a secure, responsive and economically viable semiconductor supply chain.