State of Moore's Law 2024

In December 2022, I wrote a blog post discussing the speculation around Moore’s Law being “dead.” At the time, the conversation revolved around the physical limitations of transistor scalability and how emerging technologies like AI and quantum computing might keep the spirit of Moore’s Law alive. Fast forward to 2024, and the landscape has evolved dramatically. AI has taken center stage, quantum computing is making strides, and the semiconductor industry is racing to push the boundaries of what’s possible.

So, where does Moore’s Law stand today? Is it still relevant, or has it finally met its end? Let’s dive in.

The AI Boom and Its Impact on Moore’s Law

2024 has been a watershed year for AI. Large Language Models (LLMs) like ChatGPT have dominated the tech landscape, driving an insatiable demand for computational power. Nvidia, with its industry-leading GPUs, has emerged as the undisputed champion of the AI market. Meanwhile, Intel has struggled to keep pace, failing to deliver the scaled CPUs it promised. AMD has made some gains but continues to grapple with its infamous driver and software support issues.

The AI boom has also led to significant investments—approximately $100 billion has been poured into AI research and development, with generative AI being a major focus. This influx of capital has not only accelerated innovation in AI but has also breathed new life into Moore’s Law.

Steve Jurvetson, a prominent venture capitalist, recently shared an eye-opening update on Moore’s Law. He highlighted a semi-log graph showcasing a staggering 1,000,000,000,000,000,000,000x improvement in computation per dollar over the years. His conclusion? Moore’s Law isn’t just alive; it’s thriving.

The Moore's Law Update
NOTE: this is a semi-log graph, so a straight line is an exponential; each y-axis tick is 100x. This graph covers a 1,000,000,000,000,000,000,000x improvement in computation/$. Pause to let that sink in.

Humanity’s capacity to compute has compounded for… pic.twitter.com/oo7dGgVUhh

— Steve Jurvetson (@FutureJurvetson) December 2, 2024

The Race to 1nm: Pushing the Limits of Lithography

While AI has been the star of the show, the semiconductor industry hasn’t been idle. Companies are working tirelessly to develop lithography machines capable of producing chips at 1nm resolution. This is no small feat—achieving such precision requires overcoming immense technical challenges.

One such company, Lace Lithography, is at the forefront of this effort. Their innovations could pave the way for smaller, more powerful, and energy-efficient chips, ensuring that Moore’s Law remains relevant for years to come.

The Current Scale of Transistor Gate Length

The current commercial scale for transistor gate length is around 2 to 3 nanometers, with companies like TSMC and Intel leading in this domain. However, research has pushed beyond this, with experimental gates reaching down to 0.34 nanometers using materials like graphene.1, 2

Progress of Key Companies

Intel: Intel has been vocal about extending Moore’s Law, introducing technologies like RibbonFET and PowerVia to make significant leaps in transistor density. They aim for a trillion transistors on a single device by the end of the decade, which would require breakthroughs in both process technology and packaging.

Intel CEO Pat Gelsinger says until the Periodic Table is exhausted, Moore's Law is alive and well, and there will be a trillion transistors in a single device by the end of the decade pic.twitter.com/ORVtQVGBEb

— Tsarathustra (@tsarnick) January 18, 2024

AMD: AMD has been focusing on chiplet designs and advanced packaging techniques to increase performance without solely relying on transistor scaling. Their approach, like Intel’s, involves multi-die architectures to distribute functions across multiple chips for efficiency and performance gains.

NVIDIA: Known for its GPU technology, NVIDIA has leaned into AI-specific hardware, which doesn’t strictly adhere to Moore’s Law for performance increases. Their advancements in AI and graphics processing are more about architectural improvements and specialized hardware rather than transistor count. However, they continue to benefit from semiconductor advancements in terms of power efficiency and density.

Others: Companies like TSMC are crucial in manufacturing, pushing the boundaries of what’s possible with current semiconductor technology. Samsung and other players are also involved in this race, focusing on both process node shrinkage and innovative chip designs to continue performance improvements.

Quantum Computing: A Game-Changer

Quantum computing is another area that’s making waves. Google’s Quantum Chip Willow, for instance, can perform calculations in approximately 5 minutes that would take the world’s largest supercomputer 10^25 years to complete. While quantum computing is still in its infancy, its potential to revolutionize computation is undeniable.

The pace of innovation in both AI and quantum research has been staggering, and these technologies are playing a crucial role in extending the life of Moore’s Law.

Open Source LLMs: A Glimpse into the Future

On the AI front, there’s exciting news for the open-source community. Deepseek, an open-source LLM, promises to deliver state-of-the-art performance with significantly less computational power than models like ChatGPT. This development is a game-changer, as it lowers the barrier to entry for AI research and development, further democratizing access to cutting-edge technology.

The Road Ahead

So, where does all this leave us? While the physical limitations of transistor scalability remain a challenge, the combined forces of AI, quantum computing, and advancements in lithography are ensuring that Moore’s Law continues to hold true—at least for the foreseeable future.

I’m hopeful that we’ll soon see chips with 1nm gate lengths, delivering unprecedented levels of computing power in smaller, more efficient packages. The journey ahead is filled with challenges, but the pace of innovation suggests that we’re in for some fascinating times.

A Final Thought

As we look to the future, it’s worth noting that the tools we use to navigate this rapidly evolving landscape are also changing. Google Search, once the undisputed king of information retrieval, is becoming obsolete. What are you using to stay informed? Let me know in the comments—I’d love to hear your thoughts!

In the meantime, one thing is clear: Moore’s Law may have faced its share of challenges, but it’s far from dead. If anything, it’s evolving, driven by the relentless pace of technological innovation. And that’s something to be excited about.

BTW, google search is obsolute. Do commnet what are you using or if you want to know what you should.

Reference: https://www.cnet.com/tech/computing/how-intel-will-keep-moores-law-cranking-for-years-to-come/