Leading-edge foundry roadmaps for TSMC, Intel and Samsung — outlining the path to 1.4nm nodes and beyond

TSMC, Intel, and Samsung have confirmed the initiation of mass production of chips using 2nm-class process technology, marking a major milestone in semiconductor manufacturing. This development underscores the competitive race for advanced node technology, critical for high-performance computing, AI, and mobile applications.

In December 2025, TSMC began high-volume manufacturing with its N2 process at two Taiwan fabs, focusing on predictable scaling and specialized high-performance nodes. Samsung started production using its SF2 node, though some analysts suggest it may be a rebadged SF3P, with a focus on yield improvement rather than scaling breakthroughs. Intel announced its 18A node in November, with plans for high-volume production by 2025, emphasizing the integration of gate-all-around (GAA) transistors and PowerVia backside power delivery network (BSPDN).

Intel’s roadmap includes subsequent nodes such as 14A and 14A-E, slated for 2027-2028, which will introduce second-generation GAA transistors and High-NA EUV lithography. Intel’s aggressive timeline aims to reassert its leadership in process technology, with significant interest from external customers, including licensing deals like Musk’s Terafab project. Meanwhile, Samsung’s focus remains on yield improvement, with a more iterative roadmap that emphasizes defect reduction and process stability.

Why It Matters

This development is pivotal for the semiconductor industry, as the transition to 2nm technology promises substantial gains in performance, power efficiency, and transistor density. It will directly impact the capabilities of high-performance computing, AI, and mobile devices, shaping the competitive landscape among leading foundries. For consumers and enterprise users, this may translate into more powerful, energy-efficient chips in the coming years.

Yahboom K230 AI Development Board 1.6GHz High-performance chip/2.4-inch Display/Open Source Robot Maker Python, Supports AI Visual Recognition CanMV Sensor (with Adjustable Bracket)

【Flagship performance, extremely fast response】Equipped with a 1.6GHz main frequency chip, the KPU computing power is 13.7 times…

AmazonView Latest Price

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Background

Leading-edge process technology has historically driven the semiconductor industry’s growth, with TSMC, Samsung, and Intel competing to be first to market with smaller nodes. TSMC was the first to initiate high-volume production of 2nm chips, followed by Samsung’s earlier adoption of GAA transistors in 2022, though with yield challenges. Intel’s roadmap is notably more ambitious, aiming for multiple nodes with advanced GAA transistors and EUV lithography, despite recent delays and cancellations like the 20A node. These efforts reflect a broader industry push toward smaller, more efficient nodes amid rising demand for AI, 5G, and data center applications.

“Our N2 process is now in high-volume production, enabling us to deliver predictable scaling and tailored solutions for high-performance applications.”

— TSMC spokesperson

“Intel’s 18A node marks a significant leap in process technology, integrating GAA transistors and PowerVia to deliver industry-leading performance and efficiency.”

— Intel representative

“Our focus remains on yield improvement and process stability, ensuring that our 2nm nodes deliver reliable performance for our customers.”

— Samsung spokesperson

What Remains Unclear

It is not yet clear how quickly yields will improve for Samsung’s 2nm process, or whether Intel’s ambitious timeline for 14A and 14A-E will be met given recent delays. The full commercial impact of these nodes remains to be seen, especially regarding early customer adoption and supply chain readiness.

What’s Next

TSMC is expected to expand high-volume production of N2 chips in 2026, while Intel aims to ramp 18A production and initiate early 14A manufacturing. Samsung will continue refining its yield processes and may accelerate adoption of its 2nm technology once yield stability improves. Industry analysts will closely monitor yield data, customer interest, and the pace of EUV lithography adoption in the coming months.

Key Questions

What are the main differences between TSMC, Intel, and Samsung’s 2nm strategies?

TSMC focuses on predictable scaling and specialization, Intel emphasizes aggressive innovation with GAA transistors and EUV, and Samsung prioritizes yield improvement and iterative development, with less emphasis on scaling breakthroughs.

When will these new nodes be available for consumer products?

TSMC’s N2 is already in high-volume production, with broader adoption expected in 2026. Intel’s 18A is targeted for 2025-2028, while Samsung’s yield improvements may enable earlier deployment for specific applications, but widespread availability remains uncertain until yields stabilize.

How will these advancements impact chip performance and power efficiency?

Smaller nodes like 2nm are expected to significantly improve performance, reduce power consumption, and increase transistor density, enabling more powerful and energy-efficient devices across various sectors.

What are the main risks associated with these roadmap plans?

Yield challenges, delays in EUV lithography adoption, and execution risks could impact timelines. Samsung’s yield issues and Intel’s recent cancellations highlight potential hurdles in bringing these advanced nodes to mass production.