In short: At its latest Foundry Discussion board occasion, Samsung revealed a contemporary roadmap, exhibiting off bleeding-edge course of applied sciences and superior chip packaging options. With breakthroughs in 2nm and 1.4nm processes, the corporate appears poised to compete with rivals like TSMC amid the generative AI growth.
First up, Samsung took the wraps off two new course of nodes aimed toward pushing the boundaries of chip efficiency and effectivity. The headliner is the 2nm SF2Z course of, slated for mass manufacturing in 2027. This one packs a intelligent trick with a bottom energy supply community (BSPDN) that routes the ability rails on the flip aspect of the silicon wafer. This design eliminates pesky interference points and voltage drops that hamstring high-performance computing builds.
The SF2Z additionally brings basic PPA (energy, efficiency, space) enhancements over Samsung’s authentic 2nm SF2 node. Observe that SF2 is the corporate’s second-generation 3nm course of rebranded to 2nm, possible created to assist it compete higher with Intel Foundry. Mixed with Samsung’s different 2nm nodes, this addition takes the variant rely as much as 5 for this fabrication course of.
Whereas Samsung markets the SF2 and SF2P nodes for cellular units, its audience for SF2X and SF2Z is high-performance computing (HPC). In the meantime, the ultimate 2nm node will hit automotive functions in 2027. The opposite new node unveiled is the 4nm SF4U, which makes use of optical shrinking to spice up PPA when it enters mass manufacturing in 2025.
Samsung additionally famous that work on its SF1.4 (1.4nm) course of is progressing properly towards mass manufacturing in 2027. Nevertheless, the corporate admitted that cracking the sub-1.4nm barrier would require breakthrough materials and structural improvements.
One key benefit Samsung is leaning on is its growing maturity with gate-all-around (GAA) transistors. These 3D transistor buildings supply superior switching efficiency and decrease energy than conventional FinFET designs. Samsung has been producing GAA chips since 2022 and plans to combine the expertise into its upcoming 2nm course of.
Samsung additionally pulled again the curtain on plans to introduce co-packaged optics (CPO) beginning in 2027.
“Alongside our confirmed GAA course of optimized for AI chips, we plan to introduce built-in, co-packaged optics (CPO) expertise for high-speed, low-power information processing, offering our clients with the one-stop AI options they should thrive on this transformative period,” famous the corporate in a press launch.
Lastly, the corporate claimed its foundry enterprise has seen AI-related gross sales surge 80 % over the previous yr throughout numerous course of nodes. That does not come as a shock contemplating the continued AI growth that is benefited different chipmakers, together with Nvidia, which not too long ago reported an astonishing 262 % year-on-year soar in income.
