๐งTom's HardwareโขFreshcollected in 23m
Samsung Delays 1.4nm to 2029
#semiconductors#process-nodes#euv-lithography#ai-chipssamsung-foundry-process-roadmapsamsung foundrysf2high-na euv
๐กSamsung's delayed 1.4nm roadmap could reshape timelines and sourcing for future AI accelerators.
โก 30-Second TL;DR
What Changed
Samsung Foundry's 1.4nm-class node is now targeted for 2029
Why It Matters
The revised schedule signals a slower transition to sub-2nm manufacturing and may affect future AI accelerator supply planning. High-NA EUV adoption could eventually improve the economics and scalability of 1nm-class chips.
What To Do Next
Update your AI hardware roadmap to model Samsung's 2029 1.4nm availability and evaluate alternative foundries for nearer-term accelerator designs.
Who should care:Enterprise & Security Teams
Key Points
- โขSamsung Foundry's 1.4nm-class node is now targeted for 2029
- โขSF2 is expected to remain in production for an unusually long period
- โขHigh-NA EUV is planned for 1nm-class nodes starting in 2030
๐ง Deep Insight
AI-generated analysis for this event.
๐ Enhanced Key Takeaways
- โขSamsung's strategic pivot reflects a broader industry trend of prioritizing yield optimization and cost-efficiency in advanced nodes over aggressive, high-risk scaling timelines.
- โขThe delay is reportedly linked to challenges in achieving competitive power, performance, and area (PPA) metrics for the 1.4nm node compared to existing SF2 iterations.
- โขSamsung is shifting focus toward 'SF2P' (2nm performance-enhanced) and 'SF2Z' (backside power delivery) variants to bridge the gap until 1nm-class production.
- โขThe adoption of High-NA EUV lithography is now explicitly tied to the 1nm-class node, marking a departure from earlier industry speculation that Samsung might integrate it sooner.
- โขThis roadmap adjustment aligns with Samsung's efforts to stabilize its foundry business unit, which has faced significant pressure from TSMC's dominance in the high-end AI chip market.
๐ Competitor Analysisโธ Show
| Feature | Samsung Foundry | TSMC | Intel Foundry |
|---|---|---|---|
| Advanced Node Strategy | Focus on SF2/SF2P/SF2Z longevity | Rapid transition to N2/N2P/A16 | Aggressive 18A and 14A adoption |
| High-NA EUV Status | Targeted for 1nm-class (2030+) | Pilot/Production integration (A16/A14) | Early adopter (High-NA tools installed) |
| Backside Power Delivery | SF2Z (Planned) | N2P (Planned) | PowerVia (Implemented) |
๐ ๏ธ Technical Deep Dive
- SF2 (2nm): Utilizes Multi-Bridge-Channel FET (MBCFET) architecture, Samsung's proprietary implementation of Gate-All-Around (GAA) transistors.
- SF2Z: Incorporates Backside Power Delivery Network (BSPDN) to reduce IR drop and improve power efficiency by separating power and signal routing.
- High-NA EUV: Refers to lithography machines with a numerical aperture of 0.55, enabling higher resolution patterning required for sub-2nm features without multi-patterning.
- 1nm-class nodes: Expected to leverage further refinements in nanosheet scaling and potentially transition to new channel materials or advanced 3D stacking techniques.
๐ฎ Future ImplicationsAI analysis grounded in cited sources
Samsung will lose significant market share in the AI accelerator segment through 2028.
The delay of the 1.4nm node leaves a performance gap that competitors utilizing 2nm and 1.6nm processes will likely exploit for high-performance computing contracts.
Capital expenditure on High-NA EUV infrastructure will be deferred until 2028-2029.
By pushing the 1nm-class node to 2030, Samsung has reduced the immediate urgency to deploy expensive High-NA EUV scanners in its current fab expansion plans.
โณ Timeline
2022-06
Samsung begins mass production of 3nm GAA process technology.
2023-06
Samsung Foundry Forum announces roadmap including 1.4nm target for 2027.
2024-07
Samsung updates foundry roadmap, emphasizing SF2 and variants like SF2P.
2025-11
Samsung reports yield improvements in 3nm and 2nm GAA development cycles.
๐ฐ
Weekly AI Recap
Read this week's curated digest of top AI events โ
๐Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: Tom's Hardware โ