IBM Unveils World's First Sub-1nm Chip Technology, Promises Faster, More Efficient AI Computing

IBM has announced what it calls the world's first chip technology capable of producing semiconductors smaller than one nanometre, marking a significant milestone in the race to build more powerful processors for artificial intelligence.

The breakthrough introduces a new transistor architecture measuring 0.7 nanometres (7 angstroms), making it substantially smaller than today's most advanced commercial manufacturing processes. IBM says the technology could deliver up to 50% higher performance or 70% better energy efficiency compared to its 2nm chip technology unveiled in 2021.

The announcement comes as chipmakers around the world compete to build processors capable of handling increasingly demanding AI workloads while consuming less power.

Smaller Than 1nm. Why It Matters

For decades, the semiconductor industry has followed Moore's Law, the idea that more transistors can be packed onto chips over time, resulting in faster and more efficient processors. IBM's latest breakthrough pushes that trend even further.

The company says its new 0.7nm technology packs nearly 100 billion transistors onto a chip roughly the size of a fingernail, almost doubling the transistor density achieved by its 2nm prototype announced in 2021.

IBM's Secret Is 'Nanostack'

The biggest innovation is not just making transistors smaller. IBM has developed a new architecture called Nanostack, which arranges transistors vertically instead of laying them flat across the chip.

By stacking components in three dimensions, IBM can fit significantly more computing elements into the same physical space without relying solely on shrinking transistor dimensions. The company says this fundamentally changes how future chips could be designed.

According to IBM Research Director Jay Gambetta, the new architecture is "reinventing how chips are built" rather than simply making existing designs smaller.

A Challenge to TSMC and Intel

Although IBM no longer manufactures chips itself, its research has historically influenced the wider semiconductor industry. The company's earlier 2nm breakthrough eventually found its way into commercial manufacturing through partners. IBM has previously licensed semiconductor technologies to Samsung and Japan's Rapidus, though it has yet to announce who will manufacture the new sub-1nm chips.

The announcement also comes just days after Intel said its 18A process, equivalent to roughly 1.8nm, had entered risk production, highlighting how rapidly chipmakers are moving towards ever-smaller manufacturing nodes.

AI Is Driving the Race

The timing is no coincidence.

Artificial intelligence models continue to demand enormous computing power, pushing chipmakers to find new ways of improving performance without dramatically increasing electricity consumption.

IBM believes its new architecture could help power future generations of AI systems, cloud infrastructure, and high-performance computing platforms while using significantly less energy.

Don't Expect These Chips Anytime Soon

Despite the breakthrough, consumers should not expect sub-1nm processors to appear in laptops or smartphones anytime soon. IBM estimates commercial production is still up to five years away, and the company has yet to name a manufacturing partner capable of producing the technology at scale.

Even so, the announcement represents another reminder that the AI race is no longer being fought only through software. Increasingly, it is also about who can build the hardware capable of running tomorrow's AI models faster, more efficiently, and at lower cost.

If IBM's research ultimately reaches mass production, today's cutting-edge 2nm and 3nm chips may eventually look as dated as the processors that once powered the first smartphones.