How Semiconductors Are Shaping the Current Century

Semiconductors shape this century because almost every product and service that defines it, from smartphones and cloud computing to AI, electric and self-driving cars and connected factories, runs on chips. Each new generation of chips does more computing for less energy and cost, and that steady improvement is what makes new applications practical. The flip side is dependence: when chip supply stalls, whole industries stall with it.

Image Courtesy: Bing

Key takeaways

• Semiconductors are the base layer of smartphones, IoT, AI, vehicles and communications networks.
• The industry is large and growing: the Semiconductor Industry Association (SIA) reported global sales of about 574 billion US dollars for 2022.
• AI depends heavily on GPUs, FPGAs and custom accelerators, plus high-bandwidth memory.
• Supply chain disruption, rising costs and competition between countries are the main challenges.
• Continued investment in research, new business models and partnerships will decide who leads.

Where semiconductors are shaping the century

Semiconductors have played a vital role in shaping the current century, powering an ever-growing range of electronic devices and applications. Here are a few examples.

Smartphones

Semiconductors are the backbone of modern smartphones, enabling high-speed internet connectivity, camera technology and processing power. A single phone contains an application processor (SoC), memory and storage chips, a cellular modem and RF front end, power management chips, image sensors and many smaller parts. Billions of phones are sold each year, which makes mobile one of the largest markets for chips.

Internet of Things (IoT)

Semiconductors also power the spread of IoT devices, from smart home appliances to industrial sensors, which are changing the way we interact with the world around us. Here the needs are different from phones: low cost, very low power and long battery life matter more than peak performance, so microcontrollers, sensors and low-power radios do most of the work.

Artificial Intelligence (AI)

The development of AI systems and applications relies heavily on advanced semiconductors, particularly graphics processing units (GPUs) and field-programmable gate arrays (FPGAs), which provide the computing power needed for complex machine learning algorithms. Since this post was first written, custom AI accelerators from cloud providers and high-bandwidth memory (HBM) stacked next to the processor have become just as important, because training and running large models is often limited by memory bandwidth rather than raw arithmetic.

Autonomous and electric vehicles

Semiconductors are critical to autonomous vehicles, powering the sensors, processing units and communication systems that let these vehicles operate safely and efficiently. Electric vehicles add a second large demand: power semiconductors, such as silicon carbide and IGBT devices, that control the traction motor and battery charging.

Communications and cloud

Networks and data centres sit behind every one of the applications above. 5G base stations, optical transceivers, switches and server processors are all semiconductor products, and their efficiency sets how much the digital economy costs to run.

Why the effect keeps compounding

For decades the number of transistors that can be built economically on a chip has roughly doubled every two to three years, the trend known as Moore's Law. That scaling has slowed and become more expensive, but improvements in design, packaging (including chiplets and 3D stacking) and specialised architectures continue to deliver more performance per watt. A simple illustration, using illustrative numbers only: if the energy needed for a given computation halves every three years, then over 15 years it falls by a factor of 2 to the power 5, or 32. Applications that were too costly or too power-hungry in one decade become routine in the next.

Application	Key chip types	What the chips make possible
Smartphones	SoC, modem, memory, image sensors, power management	Computing, camera and connectivity in a pocket
IoT	Microcontrollers, sensors, low-power radios	Cheap, connected sensing with long battery life
AI	GPUs, FPGAs, custom accelerators, HBM	Training and running large models
Vehicles	Processors, sensors, power devices	Driver assistance, electrification, autonomy
Networks and cloud	Server CPUs, switches, optical and RF chips	The infrastructure behind digital services

Looking ahead

Demand for semiconductors is set to keep growing, driven by the ongoing digital transformation of society and the adoption of new technologies such as 5G networks and, in the longer term, quantum computing. SIA figures illustrate the scale: reported global sales were about 574 billion US dollars in 2022, and the association later reported about 628 billion dollars for 2024, with AI-related chips a major driver.

The industry also faces several challenges:

• Supply chain disruption: the shortages of 2020 to 2022 showed how a shortfall in inexpensive chips can halt car and electronics production.
• Rising costs: leading-edge fabs and design projects cost more with each generation.
• Competition: companies and countries are competing for talent, capacity and technology leadership, and governments are funding domestic manufacturing.

To address these challenges, semiconductor companies will need to keep investing in research and development to stay ahead, while also exploring new business models and partnerships to drive innovation and growth. Overall, the next decade is likely to see continued growth and innovation in the industry, with new applications and use cases emerging as the technology advances.

FAQ

Why are semiconductors called the foundation of modern technology?

Because nearly every electronic function, computing, storage, sensing, communication and power control, is carried out by semiconductor devices.

Which application drives semiconductor demand most today?

Smartphones and PCs have long been the largest markets, while data centre and AI chips have been the fastest growing in recent years.

What are the biggest risks to the industry?

Concentrated manufacturing locations, supply chain disruption, rising costs of advanced nodes, trade restrictions and a shortage of skilled engineers.