RAI Explainer: Strategic Importance of Continued U.S. Leadership in Chip Design

What is chip design and why is continued U.S. leadership in chip design important for the nation’s economic competitiveness and national security? What policies can support this goal? This RAI Explainer piece provides some answers.

Q1. What is chip design?

Integrated circuits (IC), widely known as chips, refer to electronic circuits in which a variety of electronic components, each serving different functions, are clustered together on a semiconductor substrate. Design companies create the layout of chips, optimizing variables including energy consumption, chip performance, and size to create a variety of chips that serve different purposes and functions. Overall, chip design—which encompasses both the design of physical chip and associated software—accounts for roughly half of the semiconductor industry's research and development (R&D) investments and value addition.

Q2. Why does chip design matter for U.S. competitiveness and national security?

Leadership in chip design gives U.S. companies a major competitive advantage in semiconductor technology. U.S. firms currently make up 46 percent of global chip design sales and a remarkable 72 percent of chip design software and license sales, supporting the nation’s technological competitiveness. Moreover, these revenues serve as the basis for continued investments in R&D, preserving the U.S. lead in design and creating new wealth. Indeed, many of the top-10 companies by market capitalization, including NVIDIA, Google, Apple, Facebook, Amazon, Microsoft, and Tesla are significantly involved in chip design. Furthermore, U.S. leadership in chip design ensures a reliable and secure supply of chips for government and other public and private software systems, bypassing security vulnerabilities associated with Chinese companies like Huawei.

Q3. What is the process of chip design research and development?

Summarizing and formulating chip design R&D is challenging given its complexity. Designers need to collaborate across technical domains and engage with professionals who carry out front-end and back-end processes—including both packaging experts and system architecture specialists. This collaboration further extends to end users of semiconductors, such as automotive and IT device companies. Driving this collaboration is the need to meet companies’ evolving specifications for higher performance and specialized types of semiconductors.

Traditionally, chip design R&D starts with chip designers considering factors such as current research advances, circuit dimensions, performance metrics (like operating speed and power consumption), development timelines, production costs, and market analyses. In general, chip design is a complex process that is far from linear, but involves three major steps: (1) selecting the basic design for the IC, (2) designing the algorithms and data flows necessary to  meet the system’s requirements and constructing logic circuits needed to implement the chip’s intended function, and (3) designing physical circuits, and testing simulation tools to ensure correct electrical and logical functionality and manufacturability.

Initially, the circuit designer selects a basic design for an IC. These designs are typically offered by an intellectual property (IP) vendor—which includes companies such as Arm or Imagination. In exchange for the chosen design, the circuit designer pays an upfront license fee for the IC design to its IP vendor, accompanied by royalties upon the shipment of products that employ the IP. This transaction creates a significant financial burden for the circuit designer. (The emergence of RISC-V, an open-source IC design originating from the University of California, Berkeley, may in some circumstances alleviate this financial burden on circuit designers.)

Next, this basic design is used to create a logic circuit (electronic circuit) using a special computer language to describe and model digital circuits. Circuit simulations are then performed to verify the functionality and performance of the design. A simulation that achieves the desired functional specifications is then finalized and used to build a physical circuit.

In the physical circuit development phase, intricate decisions are made regarding the precise placement and dimensions of transistors, components, and the interconnections that govern the flow of electricity on the silicon chip. Once the physical circuits undergo thorough verification, the chip design process is considered complete. Throughout both the electronic logic and physical circuit design processes, designers rely on a sophisticated type of software program known as electronic design automation (EDA), to streamline and expedite the design and verification procedures.

Q4. What are the current and emerging R&D concerns in chip design R&D?

Designing semiconductor circuits entails substantial R&D expenses, encompassing EDA tool usage fees, IP usage fees, and labor costs, all of which continue to escalate with advancements in semiconductor technology. For instance, the development of a 7-nanometer (nm) chip necessitates an investment of an estimated $223 million, whereas the creation of a smaller, next-generation 3nm chip demands an investment of $650 million, approximately three times the cost of developing a 7nm chip. These costs continue to increase as semiconductor technology advances. Crafting a state-of-the-art central processing unit (CPU) typical requires a skilled design company to spend several years’ worth of workhours, with an additional several years’ worth of workhours needed to integrate that CPU into a “system-on-chip” (an IC that fully integrates an entire electronic or computer system onto it), further driving up labor costs.

This rising cost of R&D in the semiconductor industry is placing a burden on both established and emerging semiconductor companies. Furthermore, U.S. design firms may confront declining revenues due to export restrictions that limit access to China’s market.

Q5. What are the current skills challenges facing the chip design industry?

Chip design companies demand the expertise of highly skilled professionals with advanced degrees in electrical, computer, or chemical engineering, and other related disciplines. In Taiwan, nearly 70 percent of employees in the chip design sector hold master’s degrees or higher.

However, design companies in many countries face a scarcity of highly skilled workers. For instance, more than 80 percent of UK companies with chip design capabilities are actively seeking new engineers, with over 1,000 vacancies currently unfilled.

The United States faces a similar skills challenge, but with a twist. The United States is a global leader in advanced science and engineering research and education. It attracts the brightest students from around the world.  Indeed, some 65 percent of students pursuing advanced degrees in electrical engineering and computer science in the United States are foreign nationals. However, these students often encounter labor restrictions that curtail their employment opportunities in the United States upon graduation. As a result, companies frequently transfer these skilled workers to their overseas subsidiaries, enabling them to continue contributing to U.S. enterprises. Qualcomm, for example, has setup recently a major design center in Chennai, India, drawing  on the skills of some 1,600 tech professionals. However, skilled workers may also establish spin-off companies locally, harnessing their U.S. training in the service of competing semiconductor industries outside the United States.

Q6. How are design firms integrated into the semiconductor ecosystem?

While the separation of chip design and manufacturing processes has become the industry norm, design and manufacturing, companies increasingly see the need to cultivate mutually beneficial relationships within particular regional innovation ecosystems. Taiwan Semiconductor Manufacturing Co. (TSMC) serves as an exemplar in this regard, maintaining an extensive ecosystem of semiconductor-related enterprises, encompassing EDA vendors, IP vendors, design houses, and manufacturing equipment suppliers. This ecosystem enables TSMC to respond to orders from a diverse clientele. Notably, TSMC has partnered with more than 20 chip design companies worldwide and is well-positioned to meet customer requirements through collaborative efforts with these design companies.

On the other hand, in the absence of an ecosystem for certain semiconductor design tools, foundries find it hard to accommodate orders from fabless IC companies, integrated device manufacturers, and other stakeholders. While measures in the 2022 CHIPS and Science Act encourage the resurgence of domestic semiconductor manufacturing, consideration of factors needed to establish a robust semiconductor manufacturing ecosystems is also needed.

Q7. How can the U.S. government foster the U.S. semiconductor design ecosystem?

The U.S. government has been actively contributing to the semiconductor design ecosystem through various efforts, including the following:

  • Intellectual property protections: The U.S. government has actively promoted robust intellectual property protections for innovators, including patent safeguards. These protections encourage private companies to invest in the costly R&D required for designing advanced semiconductors. This emphasis on IP protection helps the U.S. maintain its leadership in the chip design sector.
  • A variety of microelectronics initiatives: The DoD is working to utilize cutting-edge semiconductor technologies from the private sector in defense applications. As part of its microelectronics initiative, DoD is running a project called “Rapid Assured Microelectronics Prototypes using Advanced Commercial Capabilities” (RAMP), which provides funding to private design and manufacturing semiconductor companies to help them develop a secure design and prototyping capability.
  • The CHIPS and Science Act: The CHIPS and Science Act seeks to foster expansion of domestic wafer fabrication, testing and packaging facilities, including foundries that can manufacture chips designed by U.S. fabless firms.

In addition, the U.S. Department of Commerce (DoC) will address semiconductor design issues through the National Semiconductor Technology Center (NSTC), a new organization established by the CHIPS and Science Act that promotes semiconductor research and development (R&D).

  • Design Enablement Gateway: Under this planned DoC initiative, the NSTC is expected to make efforts to provide critical resources, including datasets and patents, to participating NSTC design firms, with the goal of accelerating semiconductor design R&D. While the concept holds promise, specific details of the initiative—including what framework will be used to promote collaboration among EDA companies, design firms, and fabs—are being worked out.
  • NSTC Investment Fund: The NSTC is also expected to establish an investment fund to help emerging semiconductor companies advance their technologies toward commercialization and detailed information on this initiative is pending.

Expanding subsidies and tax credits for semiconductor design R&D may be a further mechanism to support design leadership. Under the CHIPS and Science Act, subsidies and tax credits have been provided to help semiconductor companies manufacture more advanced chips domestically. Expanding these incentives to include semiconductor design R&D may alleviate some of the substantial R&D costs that companies face in designing more advanced semiconductors.

Hideki Tomoshige is a research associate with the Renewing American Innovation Project at the Center for Strategic and International Studies (CSIS) in Washington, D.C. Bailey Crane is a former research intern with the CSIS Renewing American Innovation Project.

 

Bailey Crane

Former Intern, Renewing American Innovation Project