Intel is finally expected to make the shift to its 10nm node later this year in the volume market with Ice Lake-SP and Alder Lake-S after nearly six years of 14nm lineups. Although the naming convention dictates that the Intel 10nm node should be behind TSMC’s 7nm in terms of density, performance, and power efficiency, that’s not quite the same. At least when it comes to peak transistor density, Intel’s 10nm node is roughly on par (or slightly better) than TSMC’s 7nm node with a peak density of 101 million transistors per mm2. In comparison, TSMC’s 7nm node is limited to 91 million transistors per mm2. Keep in mind that this is the peak density and not that average, so the actual figures may vary a bit.
Similarly, TSMC’s 5nm EUV process has a density of 171M/mm2 while Intel’s 7nm node has a peak density of 200-250M/mm2. As you can conclude from these observations, Intel’s process nodes are much denser than the corresponding TSMC nodes, and it’d be fair to say that the chipmaker’s 10nm and 7nm processes are comparable to TSMC’s 7nm and 5nm, respectively.
Intel appears to have realized that pure technical prowess won’t be enough if it’s to retain the process leadership in the semiconductor market and is looking to rename its process nodes to better reflect their complexity and performance. According to a report from OregonLive, the company will be renaming its 7nm process to 5nm to get ahead in the foundry business.
Although this technically makes sense, it will make things really confusing and even mislead investors as it makes you think that Intel has essentially jumped a node without actually doing so. Regardless, it’s unclear whether the company will actually go ahead with this or stick with its original naming conventions.
Estimates put peak densities of Intel and TSMC nodes (million transistors per mm2) at: (Via)
- TSMC 10nm: 52.5
- TSMC N7: 91
- TSMC N5: 171
- TSMC N3: 290
- Intel 14nm 37.5
- Intel 10nm: 101
- Intel 7nn: 200-250
