Samsung Foundry Raises Chip Prices Up to 15%

🔑 Enhanced Key Takeaways

•The price increases by Samsung Foundry also extend to its older 8nm node, which saw a nearly 10% hike in July 2026.
•The primary driver for these price adjustments is the surging global demand for AI chips, which has led to a significant tightening of capacity across the semiconductor industry, particularly for leading-edge process nodes.
•Chinese customers are reportedly accepting the steepest price increases, largely due to U.S. export restrictions on advanced chipmaking equipment, which limits their alternative foundry options and increases their reliance on available capacity.
•Samsung's foundry division has been operating at a loss since 2022, but these recent price hikes, combined with an influx of high-profile contracts, are projected to make the unit profitable as early as next year.
•Competitor TSMC, the market leader, has also announced price increases of 5% to 10% across all sub-5nm nodes starting in January 2026, with some services potentially rising by around 25% in 2027, indicating a broader industry trend of rising foundry costs.

📊 Competitor Analysis▸ Show

Foundry	Key Process Nodes (Advanced)	Market Share (Q1 2026)	Pricing (Avg. Wafer Cost)	Key Features/Notes
Samsung Foundry	SF4 (4nm), SF5 (5nm), SF3 (3nm GAA), SF2 (2nm GAA)	~7%	5nm: ~$13,000, 3nm (3GAE): ~$15,000	SF4/SF5 use FinFET & EUV; SF3/SF2 use Gate-All-Around (GAA) MBCFET; SF4X (HPC variant) offers +10% perf, -23% power vs. standard SF4
TSMC	N5 (5nm), N4 (4nm), N3 (3nm), N2 (2nm)	~70-73%	5nm: ~$18,500, 3nm (N3/N3E): ~$19,500, 2nm: ~$30,000 (est. H2 2025)	Dominant market leader; N5 offers 1.8x density of N7; capacity largely booked by AI orders; also raising prices (5-10% for sub-5nm in 2026, up to 25% in 2027)
Intel Foundry Services (IFS)	Intel 4 (internal), Intel 3 (foundry focus), 18A (1.8nm)	Not in top 10	N/A (struggling to compete on price/yield)	Intel 4 targeted for internal compute tiles; Intel 3 is the focus for IFS; 18A aims for 3nm-equivalent density with GAAFET (RibbonFET)

🛠️ Technical Deep Dive

SF4 (4nm): Entered mass production in 2021, primarily for mobile applications (SF4E). It utilizes FinFET transistor architecture and Extreme Ultraviolet (EUV) lithography.
SF4X (4nm HPC variant): This specialized version for High-Performance Computing (HPC) offers a 10% performance boost and 23% power reduction compared to the standard SF4. These improvements are achieved through advanced source/drain (SD) stress engineering, Transistor-level Design-Technology Co-Optimization (T-DTCO), and an optimized Middle-of-Line (MOL) scheme. SF4X also supports Ultra-Low-Vt (ULVT) devices, high-speed SRAM, and guarantees high Vdd (supply voltage) operation.
SF5 (5nm): Also entered mass production in 2021, designed to enable HPC systems with enhanced performance and reduced power consumption. Built on FinFET technology with EUV lithography, it provides up to a 25% increase in logic area efficiency, 20% lower power consumption, or 10% higher performance compared to its 7nm predecessor.
SF3 (3nm): Samsung's first 3nm process, SF3E, began mass production in 2022. It is notable for being the first to leverage Gate-All-Around (GAA) architecture with advanced nanosheet channels, which Samsung refers to as Multi-Bridge Channel FETs (MBCFETs). This transition from FinFET to GAAFET offers significant improvements in power, performance, and area (PPA) compared to 5nm, with a reported 1.23x performance increase or 45% power decrease, and a 16% area reduction.
SF2 (2nm): Samsung's second-generation MBCFET(GAA)-based node, SF2, began mass production in 2025, targeting next-generation mobile, HPC, AI, and automotive applications with enhanced stability and higher performance.

🔮 Future ImplicationsAI analysis grounded in cited sources

Samsung Foundry's profitability will significantly improve, potentially turning profitable as early as next year.

The combination of these price increases, surging AI chip demand filling capacity, and securing high-profile contracts is expected to reverse years of losses for Samsung's foundry unit.

The global semiconductor industry will continue to experience broad-based price increases across various process nodes.

Other major foundries like TSMC and UMC are also implementing or planning price hikes due to rising raw material, energy, and equipment costs, coupled with sustained AI-driven demand, indicating a systemic shift rather than an isolated event.

Samsung will likely gain market share in advanced process nodes as TSMC's capacity remains fully booked.

With TSMC's leading-edge capacity largely reserved by major customers for AI chips, other clients are increasingly turning to Samsung and Intel, providing Samsung with more leverage and opportunities to secure additional orders.

⏳ Timeline

2005-11

Samsung announces Qualcomm as its first major foundry customer.

2018-12

Samsung shares development progress of 3nm Gate-All-Around (GAA) MBCFET technology.

2019-04

Samsung completes development of its 5nm FinFET process technology (5LPE) and makes it ready for customer samples.

2021

Samsung begins mass production of its SF4 (4nm) and SF5 (5nm) process nodes.

2022

Samsung begins mass production of its first 3nm process (SF3E) using Gate-All-Around (GAA) architecture.

2025

Samsung begins mass production of SF2 (2nm), a second-generation MBCFET(GAA)-based leading-edge technology node.

2026-07

Samsung Foundry raises prices for SF4, SF5, and 8nm process nodes by up to 15% for new orders.

Samsung Foundry Raises Chip Prices Up to 15%

⚡ 30-Second TL;DR

Key Points

🧠 Deep Insight

🔑 Enhanced Key Takeaways

🛠️ Technical Deep Dive

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline

📎 Sources (35)

👉Related Updates

China Sets National Standards for Automotive Chips

Snapdragon 8 Elite Gen 6 Pro Leaks for $42