Samsung Foundry announces new SF2Z node with backside power delivery

Earlier this year, Samsung confirmed its first 2 nm mobile SoC, widely believed to be the Exynos 2500, had taped out on its SF2 node. Even Arm picked SF2 for its next-gen Cortex smartphone CPU. Samsung has now announced two new nodes at the annual Samsung Foundry Forum in California. The first node is codenamed SF4U, an enhanced version of the current-gen SF4. Samsung claims it offers PPA (power, performance and area) improvements via a die shrink. It is expected to enter mass production in 2024. The company hasn't divulged much information about

Next up, there's Samsung's newest cutting-edge node- SF2Z. It is an enhanced version of the SF2 node mentioned above, which will be reserved for AI and HPC (High-performance computing applications). Notably, SFZ2 will be the first Samsung process to include backside power delivery. TSMC plans to add in its 2026-bound 16A node. Intel has successfully demonstrated the concept already and it will feature on the Intel Foundry's 18A node alongside next year's Panther Lake CPUs.

Lastly, Samsung confirms its 1.4 nm (SF1.4) is on track for a 2027 reveal. Similarly, SF2P, an enhanced version of next year's SF2 will start mass production in 2026 alongside SF2X, a node that focuses exclusive on HPC and AI products. So far, the company has managed to keep up with TSMC, although many OEMs still prefer the Taiwanese chipmaker for a variety of reasons. Whether or not this will chance in the coming years remains to be seen.