Semiconductor 101: Understanding the jargons, what semiconductors actually do and the complex global supply chain

Authored By: Nesara S R & Abhishek Singh

Through this article, we aim to shed light on the intricacies of the semiconductor landscape and offer our take on where significant value-creation opportunities may arise in this space in the coming years.

To cater to the needs of all the different personas of the potential readers i.e. those that are completely new to the topic and would need a primer to those that have already spent some time reading about the category and would like to focus more on the nuances, we have structured the in a way that mirrors the journey of the research process we undertook:

1. First, we embark on understanding the jargon and background of the semiconductor industry — if we are part of the audience at the Olympics arena, we need to know the different games played by who the key players.
2. Second, we explore the tailwinds in this industry, focusing on two key perspectives: technological advancements and geopolitical developments (after all Olympics is about a bunch of nations fiercely vying for the top spots on the podium).
3. Lastly, we examine the implications of these tailwinds and see what the next big things in this space could look like.

Semiconductor 101: Understanding the jargons and what do semiconductors actually do?

A chip or an Integrated Circuit (IC), as often used interchangeably, can be compared to a plate that brings together a variety of electronic ingredients, including resistors, transistors, sensors, diodes, and more, these components on the chip come together to create a unique integrated offering, tailored for a specific application (camera sensing, processing, memory etc). This chip acts as the base, holding everything together and serving it up for practical use in electronic devices. Chips are created on the surface of a silicon (or any semiconductor material) wafer, that undergoes a series of processes:

1. oxidation to build an insulation layer that blocks current leakages,
2. photolithography i.e. designing the chip onto the wafer, etching to carve out the unnecessary materials,
3. deposition and ion implantation to add insulation and conducting characteristics layer by layer on the wafer,
4. metal wiring to add a conductive layer on the top of the wafer and finally the wafer is cut into small chips.
5. Finally, the chips are bonded to the PCB board, moulded into the desired shape and finally packaged to protect them and make them usable.

Packaging involves enclosing the IC in a protective casing, usually made of plastic or ceramic. This casing provides physical protection and helps to connect the IC to the external world through pins or other connectors.

Source: Knowledge, Hitachi

Now every device itself might be made of a buffet of chips, each serving its own purpose while communicating with each other to make the device work the way it’s supposed to. Take a PC for instance, it carries a chip for processing, memory, audio, network, power management etc.

While PCs typically have this architecture of one chip for each use-case, smaller devices like smartphones, tablets, wearables etc have a single chip called a System-on-Chip (SoC).

As the name suggests, SoCs are designed to provide a complete system solution within a single chip. It combines various electronic elements such as processors, memory, input/output interfaces, and sometimes additional components like graphics processing units (GPUs) or digital signal processors (DSPs).

Source: Researchgate

You might often hear these chips being described by their “node size” such as 3nm, 5nm, or 28nm. Understanding node sizes in semiconductors can be a bit tricky. It’s not about the thickness or dimensions, say 28 nanometers, but rather think of it like the size of the finest lines or details you can draw on a piece of paper.

Now, why does node size matter? It’s essentially seen as a proxy for advancements in chip tech. As technology advances, chip manufacturers strive to make their chips smaller, more powerful, and energy-efficient. Smaller node sizes enable more components to be packed onto the chip. With more components, the chip can handle complex tasks and calculations.

Chips of different node sizes are used in various applications based on their specific characteristics and requirements but a general rule of thumb is smaller nodes would typically find themselves in use cases that need high computing power, and high power efficiency. These use cases are shown in the figure below to give a general sense of where different node sizes are deployed.

So, when you hear about a new chip with a smaller node size, it means the manufacturer has achieved a breakthrough in making components even smaller and more efficient. This leads to improved performance, faster speed, and more advanced devices. Remember, a smaller node size indicates more advanced!

Global value chain: That is enabling almost every act of our daily digital life

The global semiconductor industry boasts a market value of $500–600B, with Asia-Pacific and America emerging as the largest consumers and exhibiting the highest projected YoY growth. Among the various broad categories of semiconductors, ICs dominate the market, accounting for approximately 80% of the total market share, surpassing other categories such as discrete semiconductors, optoelectronics, and sensors.

Before we deep-dive into the value chain of the semiconductor industry, let’s understand the broad categories of semiconductors:

1. ICs are essential in powering electronic devices and are widely used in applications such as computers, smartphones, automotive electronics, and consumer electronics.
2. Discrete semiconductors, another category, comprise individual semiconductor components such as diodes, transistors, and thyristors. These discrete devices are utilised for specific functions like rectifying, amplifying, and switching electrical signals in various electronic circuits.
3. Optoelectronic devices are those that convert electrical signals into light or vice versa. Optoelectronic components include light-emitting diodes (LEDs), photodiodes, and laser diodes. These components find extensive usage in telecommunications, lighting, optical fibre systems, and imaging technologies.
4. Sensors, the fourth category, are devices that detect and respond to physical stimuli such as light, temperature, pressure, or motion, and convert them into electrical signals. They are crucial in a wide range of applications including automotive systems, industrial automation, medical devices, and environmental monitoring.

While ICs dominate the market due to their versatility and widespread application, the significance of discrete semiconductors, optoelectronics, and sensors should not be overlooked, as they enable very critical functions in various electronic systems.

Different types of ICs, their applications and cost

As per this report by McKinsey, the overall industry is growing with a CAGR of 6–8% projected to result in a $1T market by the end of this decade. A huge chunk of this growth is being driven by electronics for automotive, wireless communication and computing use cases. The recent AI frenzy has intensified this growth trajectory.

The manufacturing process of chips is highly complex and unique for every type of chip. In the broader landscape of the entire value chain of production of chips and electronic devices, you can have each player taking up one or more of the following major roles:

1. Chip Designers: These are like the architects who design the blueprints for the chips. They come up with the ideas and create the designs to manage what the chip does, how much power it consumes or how fast it runs.
2. Chip Manufacturers: They use their specialised skills, machines and tools to turn the designs into physical chips that can be used in electronic devices like smartphones, satellites and computers.
3. Component Suppliers: provide the necessary materials and parts for making the chips. They cater to everything manufacturers need, like raw materials and equipment, to build the chips.
4. Soft-IP Players: These players create and sell IP, the special recipe for chips. They don’t make or design the chips themselves but earn royalties by licensing their IP to chip designers or manufacturers.

Based on the roles they undertake, companies have historically adopted one of the following business models:

1. Fabless Companies: Fab-less refers to “lack of fabrication”. These companies focus on designing the chips but don’t actually fabricate them. Instead, they work with chip manufacturers who specialise in manufacturing and produce the chips for them. Qualcomm, Nvidia, AMD etc are some of the leading fabless companies.
2. Pure-Play Foundries: These companies specialise in manufacturing chips for other companies. They don’t design the chips themselves but provide the manufacturing facilities and expertise to make chips for chip designers or fabless companies. Leading foundries include TSMC, GlobalFoundries etc.
3. Integrated Device Manufacturers (IDMs): These companies do everything themselves, from designing the chips to manufacturing them. Intel, Samsung, are some well known IDMs.
4. Original Equipment Manufacturers (OEMs): OEMs work closely with fabless companies, IDMs, and foundries to integrate semiconductor chips into their end products. They leverage the capabilities of these semiconductor companies to source the necessary chips and collaborate on the design and production process. The likes of Apple, Dell, HP, Sony etc would all fall under this category.
5. Outsourced Semiconductor Assembly and Testing (OSAT): Refers to the outsourcing of the final stages of semiconductor manufacturing. An OSAT provider takes the individual chips and carefully assembles them onto packages or substrates using advanced techniques and carries out rigorous testing. This allows customers to focus on their core competencies, such as chip design and fabrication. ASE Technology Holdings, Amkor Tech, and Powertech Technology are some of the leading OSAT companies.
6. Contract Manufacturers: Contract manufacturers, also known as Electronics Manufacturing Services (EMS) providers, offer manufacturing and assembly services for Original Equipment Manufacturers (OEMs). Companies like Foxconn and Flex are prominent contract manufacturers.
7. Component and Equipment Suppliers: Produce and supply sophisticated tools and machinery used in chip manufacturing. These tools are essential for the complex processes involved in designing and fabricating semiconductor chips, which require hundreds of intricate steps and customised processes for different chip types. ASML, LAM Research, Tokyo Electronics, and Applied Materials are the players dominating this space.

Further, on examining the industry’s value chain, it became evident to us that the end-to-end dynamics vary for players based on their scale:

1. Large-scale manufacturers and enterprises, characterised by their significant market presence, have direct engagements with large fabs, design houses and contract manufacturers. They predominantly deal with channel distributors recommended by the fab houses. Recognizing the long lead times associated with sourcing, these enterprises strategically place orders based on future projections for a period of approximately six months. This proactive approach enables them to manage potential procurement imbalances effectively. Leveraging their scale and influence, these enterprises are advantageously positioned to swap orders with their distributors, ensuring the fulfilment of their requirements in cases of excess or deficit procurement. Overall, their scale grants them greater bargaining power, fostering direct relationships with suppliers, and enabling them to negotiate tailored solutions that meet their specific requirements.
2. MSMEs often navigate the value chain differently. They typically rely on distributors or resellers to fulfil their semiconductor procurement needs. These MSMEs may collaborate with local distributors to access a wide range of semiconductor components. However, due to their comparatively smaller scale, MSMEs often encounter limitations in their bargaining power. Moreover, during periods of heightened demand, securing a stable supply of semiconductors can pose significant challenges for these businesses.

Establishing oneself as a player in any of the above-described roles requires significant time and capital investment, ranging up to a whopping $8–10B for asset-heavy models like foundries. This investment is needed to acquire highly advanced equipment, have superior quality control, build a skilled labour force, establish a global supply chain, and develop strong R&D capabilities, among other factors.

These high barriers to entry combined with a strategic focus on semiconductor capability building by certain governments and policy support (for ex: Taiwan in the early 1970s), have caused the global supply chain for chips to evolve in a very unique manner, with specific regions across the globe specialising in different parts of the production process. The complexity is further compounded by the fact that a single chip requires hundreds of different inputs. These inputs include essential components such as ultra-pure silicon wafers, gases, chemicals, and various types of high-tech manufacturing equipment. Altogether, the production of a chip involves over 1,000 intricate steps, crossing international borders approximately 70 times before it reaches the end customer!

This also explains the long lead times in this industry, which can be quite different from what we may be accustomed to in other sectors. When you place an order for chips, you should expect to wait an average of 24 to 32 weeks, which shot up to as long as 55–75 weeks during the pandemic. To support this intricate supply chain, large semiconductor companies often rely on a vast network of suppliers, numbering as many as 16,000 worldwide.

The figure below shows the geographic specialisation and as is evident, much of the capabilities that chip production demands is being fulfilled by the US, Taiwan, Japan, Korea, and China.

Source: Global Semiconductor Supply Chain, EPRS

Here’s a breakdown of the key areas of focus:

1. United States (US): US is known for its strength in chip design and software development. Many leading fabless and IDM companies, including Intel, NVIDIA, and Qualcomm, have their headquarters or major design centres in the US. The country has a strong ecosystem of skilled engineers and research institutions driving innovation in chip architecture and software algorithms.
2. Netherlands (Europe): The Netherlands, particularly represented by ASML, is a prominent player in the semiconductor equipment and components segment. ASML has practically monopolised the supply of advanced lithography systems used in chip manufacturing. Their cutting-edge equipment enables the production of high-performance chips with smaller node sizes, contributing to the advancement of the semiconductor industry globally.
3. Taiwan, South Korea: These countries are known for their expertise in semiconductor manufacturing and distribution. Taiwan is home to major semiconductor foundries, such as TSMC and UMC, which offer advanced manufacturing processes for chip production. South Korea hosts Samsung and SK Hynix, prominent memory chip manufacturers. TSMC and Samsung are currently the only players who can produce the leading-edge 2–7 nm chips.
4. Japan : The country has a long history of innovation and collaboration in the semiconductor industry, with strong R&D capabilities. Manufacturers like DISCO specialise in precision machinery for high performance chips while the likes of Toshiba Memory (now Kioxia), Renesas Electronics, and ROHM Semiconductor have made significant contributions to the development and manufacturing of memory technologies such as NAND flash, DRAM, and NOR flash. Shinko Electric Industries and Ibiden Co., Ltd, also headquartered in Japan, specialise in semiconductor packaging and assembly technologies.
5. China: China, with companies like SMIC and Unigroup, is rapidly developing its semiconductor industry, focusing on manufacturing and distribution. It also houses manufacturing houses with certain niche expertises such as:
6. Bitmain: specialising in mining hardware for cryptocurrencies, primarily known for its Bitcoin mining machines (ASICs).
7. Goodix: specialising in the design and manufacturing of fingerprint sensors and other biometric authentication solutions for smartphones and other devices.

Apart from this, China is regularly pumping billions of dollars to develop infrastructure to be able to produce leading-edge node-size chips as well.

When examining the global fabrication capacity across these key regions, several insights emerge:

1. An overwhelming 92% of the leading-edge capacity is concentrated in Taiwan, with the remaining 8% being located in South Korea, primarily held by TSMC and Samsung.
2. Chips can be broadly categorised into three main types: logic, memory, and DAO (digital, analog, and other chips). In the memory and DAO segments, South Korea and Japan collectively dominate.
3. Specific node sizes, which determine the size, power efficiency, and performance of chips, are concentrated in different regions. While extra-large node sizes are slowly becoming obsolete, all the rest of them remain equally crucial, requiring careful consideration of trade-offs for different applications. As a result, the entire world depends heavily on a fixed set of 6–7 regions for their entire chip supply, highlighting the significance of these key players in the industry.

This regionalised interdependency results in more than 50 critical points in the global supply chain where a single region holds over 65% of the global market share for a specific component or process. While this specialisation allows for efficient production, it also makes the supply chain highly vulnerable to disruptions caused by disasters, accidents, infrastructure failures, and geopolitical tensions.

In fact these vulnerabilities have already started to emerge, laying the groundwork for the second part of this article, where we will explore the key developments that are driving the need for diversification and resilience in the semiconductor supply chain.

Please follow the following link to access part 2 of this article, where we have talked about the disruption taking place in this multi-decade-old- global value chain and the major trends that we are excited about.

I think that’s it for now folks! I hope that this post was helpful to the readers. I would love to hear any comments and feedback on the above. And, if you’re building in semi-conductor space, or are a VC actively following this space, please don’t hesitate to get in touch. My email is abhisheksingh@riverwalkholdings.com. I would love to connect with you to brainstorm and identify the overlaps.

Additionally, we are maintaining a repository of all the semiconductor start-ups in India here. Please feel free to reach out if you would like to suggest a startup for inclusion in the list or propose modifications to any entry.