In 1965, Gordon Moore, co-founder of Intel, made an observation that would define the technological progress of the past decades: the number of transistors on a microchip doubles approximately every two years, while the cost per transistor remains constant. This prediction, known as Moore’s Law, has driven the explosive growth of computing power and the miniaturization of electronics. But what does this law mean today, as we approach the boundaries of physical limitations? And how could the concept of a “Hyper Moore’s Law” reshape our future?

The History of Moore’s Law

When Moore published his observation in 1965, microchips were in a relatively primitive state. The first commercial chip, the Intel 4004 from 1971, contained only 2,300 transistors. However, continuous refinement of manufacturing techniques, such as photolithography and chemical vapor deposition, led to exponential growth. Today, chips with billions of transistors are the norm, pushing the boundaries of matter with nanometer-scale technology.

Moore’s Law quickly became more than a technological prediction; it evolved into an economic principle and strategic guideline for the semiconductor industry. Each year brought more computing power at the same cost, enabling the rise of personal computers, smartphones, and artificial intelligence.

The Crisis: Is Moore’s Law Dead?

In recent years, Moore’s Law has come under pressure. In 2021, Nvidia CEO Jensen Huang even declared that Moore’s Law was “dead.” The challenges are significant: transistors are now so small that they are only a few atoms wide, making quantum mechanical effects such as tunneling and heat dissipation major hurdles. This renders further miniaturization extremely expensive and complex.

As a result, companies like Nvidia have shifted focus from pure hardware innovation to “co-design,” where software and hardware are optimized together. A prime example is the development of GPUs (graphics processing units) specifically for artificial intelligence, such as Nvidia’s tensor cores designed for matrix calculations. These innovations play a key role in training neural networks, the backbone of modern AI applications.

Hyper Moore’s Law: The New Promise

Recently, however, Huang revised his earlier statements. Instead of predicting declining growth, he now envisions a “Hyper Moore’s Law,” where performance doesn’t just double but potentially triples annually. This promise doesn’t rely on traditional transistor scaling but rather on system-level innovations.

For example, Nvidia’s new Blackwell platform claims to speed up the training of large language models by a factor of 30 and improve simulations by a factor of 20. This is achieved through technologies like NVLink, a system that enables more efficient communication between GPUs in large data centers. While independent tests are still pending, and overheating issues have been reported, this demonstrates the industry’s shift toward scalable solutions beyond individual chips.

The Future of Computing

What does this mean for the coming years? We are on the brink of an era where the classical limitations of Moore’s Law are overcome through a combination of co-design, alternative technologies, and new manufacturing methods. Candidates such as photonic chips, spintronics, and quantum dots offer hope but are not yet ready for mass production.

However, it’s important to note that the cost of research and development has grown exponentially. Since the 1960s, these costs have increased by a factor of 20, highlighting the need for strategic decisions and collaboration between hardware and software companies.

Moore’s Law, or a variant of it, remains crucial to technological progress. Whether it’s artificial intelligence, scientific simulations, or everyday applications like connecting your phone to a printer, the future of computing depends on our ability to address these challenges.

With the promise of a Hyper Moore’s Law, the constraints of classical transistor technology may become less relevant over time. It’s an exciting era where technology and innovation continue to converge, driven by a new interpretation of an old law.