IBM Unveils World's First Sub-1-Nanometer Chip Technology, Stock Surges Over 6% in Pre-Market Trading
Miles Bennett
IBM Research announced the world's first 0.7-nanometer node transistor architecture, sending shares up over 6% pre-market. If this technology reaches mass production, it redefines the density and efficiency ceiling for AI chips.
What does 0.7 nanometers actually mean?
Today's most advanced AI chips run on 3nm process technology. The mainstream sits at 5nm and 4nm. IBM jumped straight to 0.7nm — several generations smaller than the current cutting edge.
This means → the same chip area can pack far more transistors, delivering more computing power at lower energy cost.
In plain terms = if a chip is a plot of land, current technology builds 20-story towers on it. IBM just showed you can build 50 stories on the same plot — more capacity, less electricity.
What is the core innovation — "nanostacks"?
IBM's key breakthrough is called nanostacks. Today's nanosheets — thin, horizontal layers of transistors — lie flat. Nanostacks stack transistors vertically instead.
This means → without increasing chip area, density rises simply by building upward. It is the next-generation architecture direction after nanosheets.
IBM Research chief Jay Gambetta: "The question we asked is — after nanosheets, how do you keep going?"
Can this be manufactured? What does IBM itself say?
Gambetta was explicit: this is still at the research stage. IBM is not claiming manufacturability.
His exact words: "Only when manufacturing cost and performance gains match up well will this technology truly have an impact. From early data, that likelihood is very high."
This reflects IBM's playbook — prove it in the lab, then license it to foundries for mass production. The 2nm prototype IBM showed in 2021 was later adopted by every leading foundry. Same path.
IBM doesn't make chips — so who uses this?
IBM sold its semiconductor fabs to GlobalFoundries back in 2015, exiting physical manufacturing entirely.
In plain terms = IBM's role is "technology inventor," not "factory operator." It proves the road works in the lab; foundries like TSMC and GlobalFoundries then scale it up.
This means → the breakthrough matters for IBM's stock not because IBM sells chips, but because it reaffirms IBM's R&D leadership — the same story as the 2021 2nm milestone.
Is IBM also betting on quantum computing?
In May, the U.S. Department of Commerce and IBM jointly committed $1 billion to build an independent quantum-chip foundry called Anderon. In the same funding round, GlobalFoundries received $375 million for domestic quantum manufacturing capacity.
Gambetta: "We see quantum computing as the next core technology, but at the same time, we're working hard to make sure classical computing doesn't hit a dead end."
This means → IBM's strategy is a two-track bet: push traditional silicon chips to their physical limits with nanostacks on one side, and prepare the quantum era on the other.
Content is for reference only, not financial advice.