Foundry Advanced Process Nodes: TSMC / Samsung / Intel (July 2026)

Advanced process nodes from the big three foundries (TSMC, Samsung, Intel): mass-production timing, transistor architecture and representative use, ordered by year. Q4 2025 is where all three converge on the 2nm generation.

FoundryNodeProductionTransistorRepresentative use
TSMCN5 (5nm)2020FinFET (first EUV)Apple A14
Samsung5LPE2020FinFETExynos series
IntelIntel 72021FinFETAlder Lake
Samsung3nm GAA (3GAE)2022MBCFET (GAA)HPC
TSMCN3 / N3E2022–23FinFET (last gen)Apple A17 Pro
IntelIntel 42023FinFET (EUV)Meteor Lake
IntelIntel 32023–24FinFETXeon 6
IntelIntel 18A2025RibbonFET + PowerViaPanther Lake
TSMCN2 (2nm)Q4 20251st-gen Nanosheet GAAApple A20
SamsungSF2 (2nm)Q4 2025MBCFET (GAA)Exynos 2600
TSMCA16 (1.6nm)Q4 2026 (planned)GAA + backside power
IntelIntel 14A2027 (planned)RibbonFET + PowerDirectTesla (confirmed)
TSMCA14 (1.4nm)H2 2028 (planned)2nd-gen GAA

Method & sources

Node names, production years and architectures are from each foundry''s official technology pages and press releases (tsmc.com, Samsung Semiconductor, newsroom.intel.com); customers/uses are mostly from tech media (AnandTech, Tom''s Hardware, TrendForce, DigiTimes) citing official info, as official pages rarely name customers. ⚠ Each foundry''s "nm" is a marketing label, not an actual gate dimension; the same "nm" across foundries is not the same physical size, so no equivalence conversion is made — official node names are listed as-is. "(planned)" marks announced/expected timelines, not yet in mass production (A16/A14/14A, etc.). Intel 20A was cancelled for external production (test vehicle only) and is not listed. PPA figures like SF2''s vary slightly by official venue. Timelines change; refer to the latest official announcements.

Source: https://www.tsmc.com/english/dedicatedFoundry/technology/logic/l_2nm

Retrieved: 2026-07-26