Intel says it’s on track to produce 1nm-class chips by 2027

Nanometres are no longer small enough.

The Intel Inside logo from 2003, with the acronym 'AI' replacing Intel.

Intel has added the 1nm-class (10A) process to its roadmap while putting high hopes on 1.8nm-class (18A) manufacturing. The silicon giant is working on climbing back the technology leadership ladder, helped by massive investments in chip manufacturing.

Intel’s 1nm-class manufacturing process – or as it likes to call it 10A, with A denoting Angstroms – is expected to be in mass production some time in 2027. At the same time, the company is planning to create fully autonomous AI-powered fabs in the future.

Intel is focusing on node advancement to compete against TSMC and Samsung. It will use manufacturing prowess to build chips for its own product groups and also for other fab-less brands via the Intel Foundry (previously Intel Foundry Services) initiative. In the meantime, Team Blue aims to start production of its 14A (1.4nm) in 2026, with 18A rolling out in 2025 and 20A arriving later this year on Arrow Lake CPUs.

20A and 18A nodes have been in production since 2023, at least in some fashion. In fact,18A is so important to Intel that Pat Gelsinger said he bet the whole company on it. But that doesn’t mean the brand is putting all its eggs in the same basket. Intel is also pursuing its advanced packaging technologies, such as Foveros, which aim to combine multiple tiles to make ever more complicated chips without the limitations that stem for using a monolithic die.

18A is also the culmination of Intel’s five nodes in four years strategy. With it, the chip giant aims to take back market leadership. It offers many advancements such as PowerVia, which puts the power wires below the transistors instead of up top. According to Intel, this improves frequency and reduces power leakage.

While Intel didn’t disclose any performance figures, it told Tom’s Hardware that a node is classified as offering double-digit power/performance improvement. Pat Gelsinger further confirms this by saying that a new node brings around 14% to 15% improvement. This means that Intel’s 10A is scheduled to be 15% faster or more efficient than the 14A node, or some combination of either.

Lastly, Intel is investing over 100 billion dollars (USD) over the next five years to expand fab capacity. Though nominally a large amount, it’s simply the necessary outlay for Intel’s plans to become the number two foundry, by revenue, before the end of this decade.