Intel's 14A Process Expected to Perform Within 5% of TSMC's A14 Node

Intel's 14A Performance Target

Naga Chandrasekaran, Intel Foundry's chief technology and operations officer, announced that Intel's 14A process technology is expected to achieve performance 'within 5%' of TSMC's A14 node. Both are 1.4nm-class technologies. This projection suggests Intel's efforts to compete closely with TSMC in advanced chip manufacturing.

Ambiguity of 'Within 5%'

The phrase 'within 5%' is open to interpretation, meaning 14A could be either 5% faster or 5% slower than A14. Historically, Intel's process technologies have sometimes lagged TSMC's in transistor density while remaining competitive in performance or power, partly due to Intel's focus on CPU performance over raw density.

Current Competitive Landscape

Actual shipping processors indicate that Intel's 18A process is competitive with TSMC's N2 in maximum achievable CPU frequency. For example, Intel's Core Ultra X9 388H 'Panther Lake' reaches 5.10 GHz, while AMD's EPYC 9586F hits 5.0 GHz, and Apple's A20 Pro achieves 4.93 GHz. These real-world examples suggest that N2 processors do not show a substantial frequency advantage over 18A, with Intel's process sometimes showing higher observed CPU frequencies.

Expected Gains for 14A

Intel's internal estimates project that 14A will provide 15% to 20% higher performance at the same power, or 25% to 35% lower power at the same frequency and transistor count, compared to its 18A process. In contrast, TSMC expects its A14 to be 10% to 15% faster than N2 at the same power, or 25% to 30% lower power at the same clocks and transistor count.