Technology in Clinical Trials #7 – Taiwan's Silicon Powerhouse: How Tiny Chips are Fueling the AI Revolution

As companies worldwide are pushing for the adoption of artificial intelligence (AI)-based tools, it is often said that ”What steam was to the First Industrial Revolution, is what AI will be to the fourth”.

My previous newsletters have focused on the software bit of AI. In this issue, however, we explore the hardware behind AI – namely, microchips.

The progression of each industrial revolution follows a distinctive "S-curve" pattern. This curve begins with the learning phase, which is typically a lengthy period characterized by experimentation and trial-and-error as pioneers work to develop functional solutions. Once the core technology is established, the curve moves into the implementation phase, where organizations focus on integrating the technology into their production systems. Companies which are at the forefront of the Fourth Industrial Revolution are treading steadily along the implementation phase.

Finally, the S-curve reaches the optimization phase. During this stage, industries converge on best practices, new standards and protocols become widely adopted, and cost structures begin to stabilize. This S-curve model illustrates how revolutionary technologies evolve from initial experimentation to widespread adoption and eventual refinement across industries.

Personally, I must say that it is exciting to see how much development there is in the field of AI across various sectors. At the same time, there is a lot of talk about “new AI solutions and applications” without deeper understanding of what actually fuels AI. Hence, in this month’s issue of the newsletter we explore the hardware fueling the AI boom.

???? The Muscle Behind Machine Minds

Hardware serves as a vital foundation of AI innovation, offering the computational muscle required to swiftly process and analyze enormous data sets. One of the most crucial and commonly discussed pieces of hardware used to fuel AI are microchips.

A microchip (also called a chip, a computer chip, an integrated circuit or IC) is a set of electronic circuits on a small flat piece of silicon (a semiconductor material). Some of the main types of microchips include:

• A Central Processing Unit (CPU), which serves as the brain of the computer, executes instructions from software applications.
• Graphics Processing Units (GPUs) are adept at simultaneously handling multiple computations, making them ideal for the parallel processing requirements of AI algorithms.
• High-performance GPUs from companies such as 英伟达 , AMD , and 英特尔 are widely used in the AI community. Many systems are designed to integrate these powerful GPUs, offering high-performance platforms for AI workloads.
• Tensor Processing Units (TPUs) are designed to excel at tensor operations, the heart of many deep learning tasks. Hardware that supports or integrates TPUs provide a substantial performance boost, enabling faster and more efficient AI operations.
• Neural Processing Units (NPUs) simulate a human brain’s neural network, allowing to process large amounts of data in parallel, performing trillions of operations per second.
• Field-Programmable Gate Arrays (FPGAs) are customizable for specific AI functions and can be programmed to perform a variety of logical functions. Unlike CPUs or GPUs, which have a fixed function once manufactured, FPGAs can be reprogrammed to desired application or functionality requirements after manufacturing.
• Memory systems which are used for rapid data storage and access.

?????? Decoding Microchip Supply Chains

The microchip or semiconductor supply chain encompasses an interconnected network of organizations involved in the entire lifecycle of microchips - from conception to consumer.

This intricate system includes design firms, manufacturing facilities, testing centers, packaging companies, and distribution networks. The process is multifaceted, requiring seamless coordination across various stages. It begins with sourcing raw materials and components, proceeds through designing chip architecture and fabricating the semiconductors, continues with conducting quality assurance tests and packaging the finished chips, and concludes with distributing to manufacturers and integrating into end-user products.

This comprehensive chain ensures that semiconductors progress efficiently from initial concept through to final delivery, ultimately reaching consumers in the form of various electronic devices and technologies (including those which fuel AI tools). The microchip industry is also a major employer, with hundreds of thousands of people working in the design, manufacturing, and distribution of semiconductors worldwide.

The main operations involved in the microchip supply chain include:

• Design and development - Semiconductor design companies, often called "fabless" companies, specialize in creating and developing innovative chip designs, either in-house or through collaborations with partners. The term "fabless" indicates that these firms focus on design and marketing, outsourcing the actual fabrication of their chips to specialized manufacturing facilities.
• Production - Specialized manufacturing facilities known as "foundries" are responsible for the physical production of semiconductor chips. These foundries receive detailed designs from fabless companies and utilize advanced, custom-built equipment to fabricate and etch the intricate circuits onto silicon wafers.
• Testing and assembly - Following fabrication, semiconductor chips undergo rigorous testing procedures to verify their functionality, a process typically conducted by specialized testing and assembly firms. Once the chips successfully pass these quality checks, they are integrated into various electronic components and devices for consumer and industrial applications.
• Distribution - After production, semiconductors are subjected to comprehensive testing by specialized firms to ensure optimal performance. Chips that meet quality standards are then incorporated into a wide range of electronic products, serving both consumer and industrial markets.

?? Taiwan’s Silicon Powerhouse

Since I am based in Taiwan, I thought it would be a great opportunity to discuss Taiwan’s role in the intricate microchip supply chain.

Even though Taiwan’s role in the global microchip supply chain has been in the spotlight only during the past few years, the East Asian nation has in fact been a chip manufacturing giant for decades. The nation's long-standing experience and advanced production capabilities have positioned it as a key enabler of the Fourth Industrial Revolution for its global partners in the AI field. Taiwan leads the world in semiconductor foundry (production) services and excels in chip packaging and testing. In integrated circuit (IC) design, it ranks second globally, surpassed only by the United States. Taiwan has also achieved a significant milestone by initiating mass production of the world's first 3nm chip, a cutting-edge component set to power the next generation of AI technologies.

Taiwan's manufacturing capabilities extend beyond microchip fabrication, producing approximately two-thirds of AI servers worldwide. Through ongoing investments and international collaborations, the country aims to maintain its position at the forefront of AI innovation, applying the technology across various industries.

Besides advancing production capabilities of high-end chips and garnering media attention, there have also been a lot of local events and announcements in the AI space:

• For example, I recently attended an event organized by Nucleate and Nucleate Taiwan on Al in Biotech: Applications in Healthcare which focused on recent advancements in the use of AI solutions in life sciences and healthcare.

• The annual Computex Taipei event noted record attendance in 2024, with 1,500 companies from 36 countries participating in the proceedings, and attracting over 50,000 attendees. The event featured computing product announcements from tech giants in the AI space such as AMD , 华硕 , 英特尔 , and 英伟达 .

• Microsoft, Amazon and Google have set up local AI research and development centers, allowing Taiwan to extend its AI partnership beyond manufacturing and testing.
• TSMC, Taiwan’s largest semiconductor foundry has started constructing its first fabrication facility in Europe (Germany) and has announced it will build its second chip factory in Japan.

???? Bridging Borders Through Teamwork

Even though certain countries may specialize in the production of specific components (such as Taiwan's microchip fabrication capabilities), the Al supply chain is very complex and relies on cross-border cooperation.

International collaboration is essential for the robust development and sustainability of the AI supply chain, as no single nation possesses all the necessary resources, expertise, and infrastructure to meet growing global demands. By fostering international partnerships, the industry can leverage diverse strengths, accelerate innovation, ensure supply chain resilience, and address potential bottlenecks more effectively. This global teamwork not only promotes knowledge sharing and standardization but also ensures that the benefits of AI can be realized equitably and sustainably on a global scale.

I hope you (the reader) learned something new from this newsletter, and (just like me) feel excited about the future and what the Fourth Industrial Revolution will bring!

? Disclosure statement ?

Lastly, I would like to point out that I am NOT associated with and am NOT sponsored by any of the organizations I refer to in this edition of the newsletter.