Why Semiconductors are at the Center of Technology and Geopolitics

Introduction

As you read this article, billions of semiconductors are working to power your device, store data, and keep you connected to the world. These tools have reshaped the digital age and are embedded in everything from satellites and smartphones to medical devices and electric vehicles. A global race has emerged as countries work to dominate the semiconductor market and manufacture the most advanced chip.

Invented in the United States 65 years ago, semiconductors sparked a technological revolution that continues to shape the global economy. Today, countries are working to grow their market through competitive government incentives and high R&D investments. There is growing pressure for countries to onshore chip production by strengthening their supply chain in order to domestically create a leading product. As the demand for faster and more energy-efficient technology grows, the need for advanced semiconductors has become crucial.

What are Semiconductors

Semiconductors are extremely intricate; researchers spend months using specialized equipment to manipulate pure elements like silicon or compounds like gallium arsenide to create this technology. The electrical properties and conductivity of these materials are altered, enabling semiconductors to control electrical currents and turn electricity on and off. Once they have the desired properties, semiconductors are packaged into chips that power our computers, phones, and other electronic devices.[1]

Graphics Processing Units (GPUs) are a specific type of chip that is at the heart of AI data center operations. Originally designed to accelerate graphics rendering in gaming and animation, GPUs are now the most commonly used hardware for training and deploying AI models.[2] Their architecture allows for parallel processing, enabling thousands of simultaneous computations that are critical for deep learning, pattern recognition, and large-scale data analysis.[3] However, GPUs do not operate in isolation. Central processing units (CPUs) remain essential to the functioning of AI servers. CPUs serve as the core computational unit that manages system operations, executes general-purpose tasks, and coordinates the execution of GPU workloads. While GPUs process volumes of simple instructions in parallel, CPUs handle the serial, control-heavy instructions necessary for the overall flow of data and computation. CPUs have fewer but more powerful cores, making them ideal for managing the operating system, preprocessing data, and scheduling tasks across the server. In AI data centers, CPUs and GPUs work in tandem—CPUs acting as the orchestrators, and GPUs as the accelerators—creating a computing environment optimized for flexibility and performance. The importance of these chips in AI advancement illustrate why countries are racing to domestically create semiconductors.

Semiconductors R&D

There are four main areas of semiconductor research and development:[4]

1. Design and development: Semiconductor companies, referred to as fabless companies, design chips and then outsource the manufacturing
2. Fabrication: Specialized factories with unique equipment manufacture the chips
3. Testing and Assembly: Separate companies test the chips to ensure they function properly before assembling them into products
4. Distribution: Chips are finally distributed to companies to sell to end-users or incorporate into their products

The current process to create a chip requires extensive time and resources as complex equipment is needed to design and build this technology. This has led to challenges in the semiconductor industry as multiple international partnerships are necessary to translate economic inputs into tangible outputs. Currently, no single country or company is able to internally manufacture the quantity or quality of chips needed to meet the needs of a modern economy.[5] Therefore, sellers must have thousands of suppliers that can fabricate chips, although this process is complicated, as some advanced chips are only manufactured by a single supplier.

The global dependence on semiconductor technologies has driven the United States to focus policies on achieving self-sufficiency by investing in domestic research and development ventures and continuing investment in knowledge exchanges and collaboration—an endeavor at the heart of the Deep Tech @ Duke initiative.

International Competition and Funding

Countries are working to increase semiconductor investments as rising international tensions make it more difficult to collaborate and build cutting-edge technology. For example, while most fabless companies are located in the United States, the majority of the fabrication is completed in Southeast Asia, with the largest manufacturer being the Taiwan Semiconductor Manufacturing Company (TSMC). The tools to fabricate chips are complex, and TSMC has honed this process, leading to dominance in this phase of R&D. This expertise provides the company with significant influence in global chip development and requires leading fabless labs like Apple and NVIDIA to strike international deals.

Riccardo Masucci, Director of Security and Technology Policy at Intel Corporation and Duke Technology Policy Visiting Scholar, has worked internationally on the EU Chips Act and Technological Sovereignty. The EU Chips Act articulates how “along the chips supply chain, Europe has a strong position in some segments (e.g., in the provision of core intellectual property (IP) blocks and fabrication tools) but lags behind in many other segments . . . [such as] in design and design automation tools.”[6] Accordingly, much of Masucci’s work with Intel aims to secure and rebalance global supply chains in the semiconductor industry to better facilitate worldwide trade. 

Innovation in the U.S. and the CHIPS Act

The Creating Helpful Incentives to Produce Semiconductors (CHIPS) and Science Act, passed in 2022, provided $52 billion to grow domestic research manufacturing and increased the budget for federal science agencies. U.S. semiconductor companies were also able to apply for government subsidies under the Advanced Manufacturing Investment Credit (AMIC) to help with onshore chip production. Projections show that as a result of this initiative, the U.S. semiconductor industry will triple in capacity from 2022 to 2032.[7] New legislation (e.g., Building Advanced Semiconductors Investment Credit (BASIC) Act) is working to continue incentivization for domestic growth. With bipartisan support, the CHIPS Act aims to create opportunities for U.S. growth and dominance in semiconductor development.

Aaron "Ronnie" Chatterji, founding director of Deep Tech and business professor at Duke University, was instrumental in pushing semiconductor legislation. He served as a Senior Economist at the White House Council of Economic Advisors (CEA) and CHIPS coordinator. During a talk at Duke, Chatterji emphasized how important it is for the U.S. to effectively implement the CHIPS Act: "But we have to get this right, both for national security and economic reasons. It is that imperative that kept me going when I served in the White House and what drives the team doing the work today." With the CHIPS Act, the goal is for the U.S. to become more self-reliant and be able to craft chips from their initial design to their final integration into a product.

Duke Deep Tech Initiative’s Goals 

To capitalize on the potential of semiconductor technology, the Duke Deep Tech Initiative is developing research teams to study emerging technical, policy, and ethical challenges. This will spur collaborations across Duke’s campus and the greater Research Triangle. Within Deep Tech, research will highlight ways semiconductors intersect with other domains like quantum, AI, climate finance, and cybersecurity. The voices of key stakeholders and significant developments will be highlighted in blog posts and events. 

To learn more about Deep Tech’s mission, upcoming events, and our work visit our webpage!

The Deep Tech Initiative at Duke University is working to leverage semiconductors and proactively consider governance challenges that will arise. Deep tech refers to technologies rooted in scientific breakthroughs with the power to transform industries and shape national priorities. At Duke, we bring expertise across AI, quantum computing, semiconductors, renewable energy technologies, and cybersecurity to drive cross-sector research and examine how these innovations are fueling reindustrialization, shifting supply chains, and redefining global strategy.