Unlocking Potential: How the CHIPS Act Fuels Semiconductor Expansion for Automation and Beyond

The CHIPS and Science Act, a bipartisan piece of legislation, appropriated nearly $53 billion with the goal of reviving US manufacturing, production, and research in semiconductors. With modern semiconductor factories, or fabs, requiring hundreds of millions of dollars to billions of dollars in total investment before they come online, these complex facilities can represent excellent market opportunities for companies operating in the robotics and automation spaces.

Beyond the direct opportunities for providing automation and robotics solutions within the fab, increased domestic chip production is poised to offer supply chain benefits and reduce the bill of materials for any product using chips. With semiconductors functioning as a major choke point for supply chains following COVID, particularly for the automotive and electronics industries, increased supply can offer a degree of safety to manufacturers throughout the US.

However, and perhaps unsurprisingly, the confluence of major government spending along with capital-intensive and technically challenging engineering projects has resulted in a slow start to this initiative. Only one significant fab has been granted funding, despite hundreds of applications. More are expected, but currently private funding is driving much of the immediate growth in this space.

The CHIPS Act targeted several areas of concern in semiconductor production, including declining domestic production capacity, the importance of domestic production as a bulwark against supply chain disruptions and ensuing economic issues, and to ensure a domestic source of semiconductors for military purposes. Understanding the multiple objectives of this act can help contextualize the various funding and policy measures included in it.

Will the CHIPS Act Reverse Shrinking US Chip Production Capacity?

The US share of modern domestic semiconductor manufacturing capacity declined precipitously, from 37% in 1990 to 12% today, partly owing to heavy subsidies, investments, and incentives offered by competing countries, such as China, South Korea, Singapore, Japan, and Taiwan. Estimates by the Semiconductor Industry Association put the level of grants and subsidies in those countries alone at over $70 billion dollars.

Modern fabs that can produce chips on the most recent nodes (a node is marketing terminology that gives context to a chip’s performance improvement) are expensive. Therefore, they require some degree of public support. For example, TSMC’s planned Arizona fabs, which will produce 4nm chips (4nm is near cutting edge, but is not the most complex node currently in volume production), was initially budgeted at $12 billion before jumping to $40 billion with new plans.

A similar pattern holds for other manufacturers, with Intel announcing plans for 2 new fabs for $20 billion, Micron planning a $20 billion set of fabs in New York, and other fabs from companies like Texas Instruments and Samsung all requiring over $10 billion in investment each. These fabs are incredibly complex manufacturing operations. They rely on a world-spanning supply chain, along with many complex technical solutions for manufacturing.

How Will the CHIPS Act Protect the Supply Chain?

As the last few years have shown, disruptions to today’s complex supply chains can cause massive economic impacts. Semiconductors are crucial components of the finished product and specific models are difficult to substitute for.

While the Global Supply Chain Pressure Index (GSCPI) has recovered from the dramatic heights of 2020 and 2021, the highly concentrated nature of semiconductor production leaves the industry vulnerable to further disruptions from nature or conflict.

This is particularly visible in the contract foundry space, where pure play companies like GlobalFoundries and TSMC operate the plants to manufacture designs from fabless companies like Apple, NVIDIA, and AMD. With TSMC commanding over 50% of the market share in this space, and the vast majority of their production capacity being concentrated in Taiwan, it’s clear that any disruption to production on the island would have far reaching impacts.

Beyond the immediate risks to actual chip production, the complexity of modern supply chains amplifies the risk of even minor disruptions. Beyond the supply chain shocks from COVID, smaller-scale disruptions like Texas’s severe winter weather and a fire at a plant in Japan were able to drastically impair automotive production.

The potential for disruption is made clear when the sheer number of chips in modern products is considered. The average number of chips per vehicle doubled between 2017 and 2021, according to the CSIS. This demand for chips in automative applications is continuing to grow thanks to the shift to electric vehicles, greater demand for chip-intensive vehicle features like driver assistance systems and automation, and even architectural shifts to the concept of a “software-defined vehicle”.

A shortage of any of these chips can lead to stalled production, the need for drastic retooling and redesigns, or feature adjustments of the vehicle – for example, BMW was forced to cut features like premium audio packages, wireless phone charging, passenger seat adjustments, digital keys, and even adaptive cruise and parking assistance.

The impact of a shortage isn’t just limited to vehicles, either. While delaying production of a $60,000 vehicle over a chip worth a few cents is the most visible example, these same commodity components are used across computers, appliances, consumer electronics, and other goods.

An important thing to understand is that not all chips are created via the same process, for both technical and economic reasons. Many of the notable shortages among less expensive chips can be traced to the falling domestic production capacity for older nodes. The Congressional Report Service cites falling cap-ex and a lack of available equipment as factors impairing the production of these chips on older nodes; notably only $2 billion of the CHIPS act was designated for these “mature” processes.

How Will the CHIPS Act Address Production Issues?

The CHIPS Act targets these issues through several measures. The cornerstone is increased financial assistance for building and expanding domestic production facilities. $39 billion is targeted at fabrication, assembly, testing, advanced packaging, and R&D facilities.

This is good news for robotics manufacturers with cleanroom-capable robots and cobots. Wafer handling, as well as the transfer of chips, masks, or carriers are all great applications for robotic solutions.

Beyond the direct funding of manufacturing, the CHIPS ACT also provides a provision for a 25% investment tax credit for qualifying investments in semiconductor manufacturing and tooling. This credit is further enriched by being eligible for “direct pay” status, making it effectively refundable.

This diagram is best read left to right, as it shows the sources of funding, how they are pooled, then how they are allocated to various initiatives. The two current major commitments as of 3/20, to GlobalFoundries and Intel, are shown here. Future commitments are expected. Amounts are in millions of dollars. 

Additional funding is allocated to various programs and funds, including:

• $11 billion for R&D, via a variety of partnerships including the Manufacturing USA Semiconductor Institute, NIST research programs, and public-private partnerships
• $200 million for the development of a domestic semiconductor workforce
• $2 billion for the Department of Defense to develop defense-specific applications and pipelines
• $500 million for international supply chain activities connected to securing telecommunications (ITSI)

Overall, it’s clear the dramatic increase in domestic semiconductor investment will spur greater demand for semiconductor-specific manufacturing equipment. Direct funding, manufacturing investment credits, and other tax credits or incentives have all led to $256 billion in private investments also being made to increase manufacturing capacity.

Still, this legislation is not a panacea for chip production, nor is it a fast-track to purchases of robotics and automation solutions for the finished fabs. Notably, 460 firms have submitted statements of interest, but only a few small grants have actually been issued a preliminary memorandum of terms (a non-binding agreement between the company and the CHIPS administrative body). Among the roadblocks are onerous environmental reviews, with the National Environmental Policy Act averaging around 4.5 years to complete.

Things have recently picked up, with GlobalFoundries securing the first major grant of around $1.3 billion on 2/19, Intel securing $8.5 billion on 3/20, and the Biden administration  aiming to reduce  some of the permitting burden via an “interagency expert working group on permitting". Much of the potential of the CHIPS Act will hinge on how these issues are navigated, as well as how promptly the grants are disbursed – delays in fab construction are doubly expensive, with delays carrying both the very real construction costs and the harder-to-quantify cost of decreasing relevancy as node advancements are made.

Where Will the Funding Make an Impact?

While the much of the data points around CHIPS Act applications are not publicly available, per 15 U.S.C. § 4652(a)(6)(G), many companies have announced their intentions to seek funding. Furthermore, many of these major fab projects are expected to be located around or very near to existing fabs. This offers significant benefits for both the skilled and specialized semiconductor workforce, alignment with existing initiatives like educational partnerships, and because of technical reasons like access to fab-grade power, water, and seismological stability.

As a result, looking at current locations of major fabs and announcements from the key players can provide significant information about where these new fabs could be located:

• Chandler, Arizona, home to multiple fabs for Intel, running a variety of nodes including 22,14, and 10nm, along with planned expansion down to 20A
• Phoenix, Arizona, notable for being among TSMC’s first fabs outside Taiwan. Beyond TSMC, other semiconductor-supporting companies are developing facilities. Amkor just finalized a plan for a $2 billion packaging, design, and testing facility in the area.
• Texas, which will welcome new fabs from Samsung and Texas Instruments. Samsung is looking to build a fab for their 5nm process, while TI is focusing on a 28nm node.
• Columbus, Ohio, will see the start of construction on 2 fabs for Intel, slated for 2025 completion. These fabs, at a cost of $20 billion, are intended to produce 10nm chips and represent the first steps of an 8 fab complex totaling $100 billion in investment.
• Upstate New York will see multiple fabs, including what DRAM manufacturer Micron claims is the “largest semiconductor facility in the history of the United States”. Micron’s plan for Clay, NY is a ways off, with operations slated to begin in the “latter half” of this decade. Alongside Clay, the towns of Malta and Marcy will see fabs from GlobalFoundries and Wolfspeed respectively. Between DRAM and SiC from Wolfspeed, New York will be relatively unique in production types, compared to some of the other locations listed.
• Boise, Idaho will also see a planned fab from Micron, with a focus on producing DRAM, but an earlier production-start of around 2025

Common to all these construction initiatives is the need for automation to accelerate production. Whether it is wafer handling, wafer inspection, chemical and material handling, assembly and packaging, or metrology; modern semiconductor manufacturing provides a variety of applications for robotics, vision systems, and motion control products.