Lesson 2 — Why Should You Care?
How Do Chips Actually Work?
Learning Material
1 pagesLesson 2 — Why Should You Care?
Understanding the Complex: How Do Chips Actually Work?
On a Tuesday in January 2023, the United States government announced that it was cutting off China's access to advanced semiconductor technology — not just chips, but the machines used to make them, the software used to design them, and any American who might help. The announcement sent shockwaves through financial markets, boardrooms, and government ministries from Beijing to Brussels.
The measure was framed as a national security action. Its scope was without precedent in peacetime: a sweeping attempt to deny a rival nation the tools needed to build the most powerful class of computing hardware.
Why would a government take such a dramatic step over manufacturing processes? Because those in power understand something that most people don't: whoever controls advanced chip production controls a disproportionate share of what happens next in the global economy, in military capability, and in the development of artificial intelligence.
This lesson is about why chips matter to you specifically — not as an abstraction, but in three concrete, practical ways.
Reason one: everything you own runs on chips, and your dependency is deepening.
Twenty years ago, a typical household contained perhaps a dozen chips. Today it contains hundreds. Your smartphone alone has chips for processing, graphics, wireless communication, GPS, motion sensing, power management, and camera image processing. Your car — if it's a recent model — has between 1,000 and 3,000 chips. Your thermostat, your washing machine, your television, your router: chips throughout.
This isn't merely a consumer convenience story. Medical devices — pacemakers, insulin pumps, hearing aids — depend on chips that must work without error, every second, for years. Aviation safety systems, railway control infrastructure, power grid management: all chip-dependent. When the 2021 shortage hit, it wasn't just car production that stalled. Medical equipment makers couldn't ship ventilators. Defense contractors ran short on guided munitions.
The integration of chips into physical infrastructure means that a chip shortage isn't just a tech-industry problem. It's a civilizational problem. The more digital our world becomes, the more vulnerable it is to disruptions in this single supply chain.
Reason two: chips have become the currency of geopolitical power.
The semiconductor industry is often called "the oil of the 21st century." It's a useful metaphor, but it gets something important wrong.
Oil, for all its geopolitical weight, is fungible. A barrel of Saudi crude can substitute for a barrel of Norwegian crude. Production is distributed across dozens of countries, and while cartels can cause disruption, no single nation controls the entire supply chain.
Advanced chips are different. The most sophisticated semiconductors — the kind needed to run AI models, train neural networks, and power high-performance computing — can currently be manufactured at only a handful of locations in the world, principally in Taiwan and South Korea, with a small portion in the United States. And the machines needed to make them are built almost exclusively in the Netherlands, by a company employing roughly 40,000 people.
This concentration makes chips a uniquely powerful lever in international relations. When the United States restricted China's access to advanced chip-making equipment in 2022, it wasn't applying the blunt force of tariffs or sanctions on finished goods. It was cutting off a technological prerequisite — the ability to manufacture at the frontier — that no amount of money alone can quickly replace.
China has invested hundreds of billions of dollars in attempting to build its own advanced chip industry. As of 2026, it remains, by most estimates, five to ten years behind the frontier. The gap is not for lack of effort or capital. It reflects the extraordinary accumulated expertise, trade secrets, and supply chain relationships that it takes decades to build.
This is why chips appear, increasingly, at the center of conversations about sovereignty, national security, and the balance of power between large economies.
Reason three: the AI race runs on silicon.
When people talk about the artificial intelligence boom, they usually focus on the software — the large language models, the image generators, the autonomous systems. But software runs on hardware, and the hardware that matters most for AI is a specific class of chips: graphics processing units (GPUs) and their newer cousins, purpose-built AI accelerators.
Training a large AI model requires an almost incomprehensible amount of computation. GPT-4, released in 2023, is estimated to have required the equivalent of tens of thousands of GPU-years of computation to train. Newer frontier models require more. This computation is not free — it has to happen on physical chips, in physical data centers, drawing real electricity.
The company that makes the chips most prized for this purpose — Nvidia — found its products in such demand that, at the peak of AI investment fever in 2023 and 2024, customers were willing to pay tens of thousands of dollars for individual graphics cards and wait months for delivery. Nvidia's market capitalization at one point exceeded $3 trillion, making it among the most valuable companies ever created.
The AI race is, in part, a chip race. Nations and companies that can access advanced AI chips have a significant advantage in developing AI applications. Nations that cannot are disadvantaged. This is precisely why export controls on advanced chips have become a tool of great-power competition.
So: chips determine what you can buy and whether your medical devices work. They determine who can project geopolitical power. And they determine who leads the development of the technology that is reshaping every industry. Understanding how they work — what they are, how they're made, and who makes them — is not a technical curiosity. It's a prerequisite for understanding the world.
Next lesson: The Background You Need — the three foundational concepts (silicon, transistors, Moore's Law) that make the rest of the course possible.
Reading time: approx. 9–10 minutes