Don't Show Again Yes, I would!

Samsung preparing to launch AI-focused 2nm chips in 2025


Samsung‘s semiconductor foundry is on track to mass-produce 2nm chips in 2025 and 1.4nm chips in 2027. At the ongoing Samsung Foundry Forum (SFF) event in San Jose, California, the company revealed that its process technology development is progressing smoothly as planned. The Korean firm also announced two new process nodes and AI solutions for the next generation of electronics.

Samsung is preparing to introduce 2nm mobile chips next year

Samsung has long aimed to start 2nm mass production in 2025. Shortly after introducing its first 3nm chip in 2022 produced on its 3GAE process node, the company unveiled its semiconductor roadmap for the next five years. It planned to launch the second-gen 3nm process node (3GAP) in 2024. The Exynos 2500, which should power the Galaxy S25 series early next year, will probably be made on the 3GAP 3nm process.

The roadmap further revealed that Samsung will introduce 2nm chips in 2025, followed by a second-gen 2nm process node in 2026. Finally, in 2027, the company would enter the sub-2nm era, bringing a 1.4nm chip to the market. All this while, the Korean firm would be continuously increasing its production capacity for advanced chips. The total capacity is projected to grow over 3x between 2022 and 2027.

Samsung preparing to launch AI-focused 2nm chips in 2025

Samsung says its semiconductor technology is developing along these lines. At SFF, the company introduced an improved 2nm process node called SF2Z. Slated to enter mass production in 2027, it reduces voltage drop compared to the first-gen 2nm node (SF2), “enhancing the performance of HPC designs.” It also enhances power, performance, and area (PPA) thanks to backside power delivery network (BSPDN) technology.

See also  The Ethical Dilemma ─ Discussing Consent and Authenticity in Porn

The Korean tech titan also launched a high-value 4nm variant at SFF. Dubbed SF4U, the new 4nm process node “offers PPA improvements by incorporating optical shrink.” Samsung aims to start mass production on this node alongside its first-gen 2nm node in 2025. The company says it is also “actively shaping future process technologies below 1.4nm through material and structural innovations.”

Samsung Foundry 2nm chips 2025 1Samsung Foundry 2nm chips 2025 1

The GAA transistor architecture has become imperative in the AI era

Samsung uses the gate-all-around (GAA) transistor architecture in its 3nm chips. The new architecture brings PPA improvements over FinFET, the older architecture used up to 4nm chips (TSMC still uses FinFET and plans to upgrade to GAA with its 2nm chips next year). The Korean firm aims to benefit from an early adoption of advanced transistor technology.

According to Samsung, “structural advancements like gate-all-around (GAA) have become imperative to meet power and performance demands” of the AI era. The company says its GAA process has matured significantly over the past two years, both in yield and performance. Its GAA production will substantially expand in the coming years as it moves to 2nm and 1.4nm semiconductor process technologies.

Samsung’s ongoing foundry event in the US also brought Samsung AI Solutions, a turnkey AI platform integrating the strengths of the company’s Foundry, Memory, and AVP (Advanced Package) businesses. Customers get high-performance, low-power, and high-bandwidth solutions that can be tailored to suit their AI requirements. The Korean firm plans to introduce an all-in-one, CPO-integrated one-stop AI solution in 2027.

Samsung Foundry 2nm chips 2025 3Samsung Foundry 2nm chips 2025 3



Source Link Website

Share:

John Smith

My John Smith is a seasoned technology writer with a passion for unraveling the complexities of the digital world. With a background in computer science and a keen interest in emerging trends, John has become a sought-after voice in translating intricate technological concepts into accessible and engaging articles.

Leave a Reply

Your email address will not be published. Required fields are marked *

fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp fyp