What the Nanometre Figure on a Chip Actually Tells You

Where the Number Came From — and Why It Changed

Decades ago, the nanometre figure in a chip's spec referred to an actual physical dimension — specifically the gate length of a transistor, the tiny switch that forms the fundamental unit of computation. A 90nm chip meant transistor gates were roughly 90 nanometres wide. As manufacturing improved, those dimensions shrank predictably, and the number tracked reality closely.

That relationship broke down around the 28nm generation in the early 2010s. Chip engineers began using more complex three-dimensional transistor structures — such as FinFETs — that couldn't be described by a single linear measurement. Foundries continued using smaller numbers with each new generation, but the figures became generation labels rather than physical measurements. Today, a chip marketed as 3nm contains features far larger than three nanometres; the number signals process advancement, nothing more.

This context matters when reading spec sheets. The nm figure tells you where a chip sits in a manufacturer's generational roadmap, not a precise size you can compare across companies. For a broader breakdown of what processor specs mean in practice, see our processor spec guide.

What a Smaller Node Number Actually Signals

Even though the nm figure is no longer literal, it does convey real information when interpreted correctly. A more advanced (lower-number) node generally means:

• Higher transistor density: More transistors can be packed into the same die area, enabling more complex cores, larger caches, or additional specialized processors.
• Improved power efficiency: Smaller, more tightly packed transistors switch at lower voltages, reducing the energy each computation consumes. This is the most tangible consumer benefit.
• Potential for higher performance: Greater transistor budgets allow architects to add execution units, widen pipelines, or integrate hardware accelerators without inflating the chip's physical footprint.

Of these, power efficiency is typically what phone users notice most. A chip built on a more advanced node can sustain demanding workloads — gaming, video export, live translation — while generating less heat and drawing less battery. Architecture quality determines how effectively that efficiency budget is spent.

The Cross-Foundry Comparability Problem

Node naming becomes especially misleading when comparing chips from different foundries. TSMC, Samsung Foundry, and Intel Foundry each define their process nodes using their own internal criteria. A chip produced on TSMC's 4nm process and one on Samsung's 4nm process are not built to an identical specification — transistor density, leakage characteristics, and manufacturing techniques differ between them.

Independent semiconductor analysts have documented cases where a chip labeled with a larger node number from one foundry achieves competitive transistor density against a 'smaller' node from another. Accepting nm figures at face value across brands therefore risks drawing false conclusions about relative capability.

For everyday phone-buying decisions, this means the nm figure is a useful within-brand generational indicator — a newer node from the same manufacturer is almost always more efficient than the previous generation — but it should not be used to rank chips from different companies against each other without additional context.

What It Means for Real-World Phone Performance

Process node advancement enables capabilities beyond raw CPU speed. More advanced nodes create the physical space to integrate dedicated hardware blocks: image signal processors, security enclaves, and — increasingly relevant — neural processing units that handle on-device AI tasks. To understand how those AI cores actually work, see what neural processing units actually do in a modern phone chip.

Node generation also influences how a phone handles sustained load. Chips on older, less efficient nodes generate more heat under sustained tasks, triggering thermal throttling that reduces performance to protect hardware. More advanced nodes tend to sustain peak performance longer before the cooling system intervenes.

That said, the node number is only one variable in a complex equation. Software optimization, memory bandwidth, and display power consumption all shape the experience you'll actually notice. A phone with an older-node chip and excellent software tuning can outperform a newer-node device with poor thermal management. The nm figure is a useful input, not a definitive verdict. Where the flagship-vs-mid-range performance gap is real explores exactly these tradeoffs in practical terms.